The Effect of Treaties on Foreign Direct Investment: Bilateral Investment Treaties, Double Taxation Treaties, and Investment Flows

the effect of treaties on foreign direct investment: bilateral investment treaties, double taxation treaties, and investment flows

This page intentionally left blank

the effect of treaties on foreign direct investment: bilateral investment treaties, double taxation treaties, and investment flows

edited by karl p. sauvant and lisa e. sachs

1

1 Oxford University Press, Inc., publishes works that further Oxford University’s objective of excellence in research, scholarship, and education. Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Mexico City Nairobi New Delhi Shanghai Taipei Toronto

Madrid

Melbourne

With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2009 by Oxford University Press, Inc. Published by Oxford University Press, Inc. 198 Madison Avenue, New York, New York 10016 Oxford is a registered trademark of Oxford University Press Oxford University Press is a registered trademark of Oxford University Press, Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press, Inc. _____________________________________________ Library of Congress Cataloging-in-Publication Data The effect of treaties on foreign direct investment: bilateral investment treaties, double taxation treaties, and investment flows / edited by Karl P. Sauvant and Lisa E. Sachs. p. cm. Includes bibliographical references and index. ISBN 978-0-19-538853-4 ((hardback): alk. paper) 1. Investments, Foreign—Taxation—Law and legislation. 2. Double taxation. 3. Treaties. I. Sauvant, Karl P. II. Sachs, Lisa E. K4528.E34 2009 332.67’3—dc22 200844052 _____________________________________________ 1 2 3 4 5 6 7 8 9 Printed in the United States of America on acid-free paper Note to Readers This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is based upon sources believed to be accurate and reliable and is intended to be current as of the time it was written. It is sold with the understanding that the publisher is not engaged in rendering legal, accounting, or other professional services. If legal advice or other expert assistance is required, the services of a competent professional person should be sought. Also, to confirm that the information has not been affected or changed by recent developments, traditional legal research techniques should be used, including checking primary sources where appropriate. (Based on the Declaration of Principles jointly adopted by a Committee of the American Bar Association and a Committee of Publishers and Associations.)

You may order this or any other Oxford University Press publication by visiting the Oxford University Press website at www.oup.com

To Alaion Jr., Ana Maria, Clarissa, Marcelo, Samira, Tanja, and Vitor, whom I wish every success in their lives Karl To my parents Jeffrey and Sonia—my best teachers and my greatest inspiration Lisa

This page intentionally left blank

contents Contributors ........................................................................................................ xi Foreword ............................................................................................................ xxi andreas f. lowenfeld

Preface ............................................................................................................... xxv john h. dunning

BITs, DTTs, and FDI flows: An Overview .................................................... xxvii lisa e. sachs and karl p. sauvant

part i: introduction 1. A Brief History of International Investment Agreements ........................... 3 kenneth j. vandevelde

2. The Framework of Investment Protection: The Content of BITs ............. 37 peter muchlinski

3. Explaining the Popularity of Bilateral Investment Treaties ....................... 73 andrew t. guzman

4. Double Tax Treaties: An Introduction ........................................................ 99 reuven s. avi-yonah

part ii: exploring the impact of bilateral investment treaties on foreign direct investment flows 5. Do BITs Really Work?: An Evaluation of Bilateral Investment Treaties and Their Grand Bargain ............................................................. 109 jeswald w. salacuse and nicholas p. sullivan

6. Bilateral Investment Treaties and Foreign Direct Investment: A Political Analysis ................................................................. 171 tim büthe and helen v. milner

7. Do Bilateral Investment Treaties Increase Foreign Direct Investment to Developing Countries? ....................................................... 225 eric neumayer and laura spess

8. The Impact of Bilateral Investment Treaties on Foreign Direct Investment .......................................................................... 253 peter egger and michael pfaffermayr

9. New Institutional Economics and FDI Location in Central and Eastern Europe ....................................................................... 273 robert grosse and len j. trevino

viii contents

10. Do Investment Agreements Attract Investment? Evidence from Latin America ................................................................... 295 kevin p. gallagher and melissa b.l. birch

11. The Global BITs Regime and the Domestic Environment for Investment ............................................................................................ 311 susan rose-ackerman

12. The Impact on Foreign Direct Investment of BITs ................................. 323 unctad

13. Do Bilateral Investment Treaties Attract FDI? Only a Bit . . . And They Could Bite ......................................................... 349 mary hallward-driemeier

14. Do BITs Really Work?: Revisiting the Empirical Link between Investment Treaties and Foreign Direct Investment .............................. 379 jason yackee

15. Bilateral Investment Treaties and Foreign Direct Investment: Correlation versus Causation .............................................. 395 emma aisbett

16. Why Do Developing Countries Sign BITs? ............................................. 437 deborah l. swenson

part iii: exploring the impact of double taxation treaties on foreign direct investment flows 17. Do Bilateral Tax Treaties Promote Foreign Direct Investment? ............ 461 bruce a. blonigen and ronald b. davies

18. The Effects of Bilateral Tax Treaties on U.S. FDI Activity ...................... 485 bruce a. blonigen and ronald b. davies

19. The Impact of Endogenous Tax Treaties on Foreign Direct Investment: Theory and Empirical Evidence ............................................ 513 peter egger, mario larch, michael pfaffermayr and hannes winner

20. Host-Country Governance, Tax Treaties, and U.S. Direct Investment Abroad ..................................................................................... 541 henry j. louie and donald j. rousslang

21. Tax Treaties for Investment and Aid to Sub-Saharan Africa: A Case Study .................................................................................. 563 allison christians

22. It’s All in the Timing: Assessing the Impact of Bilateral Tax Treaties on U.S. FDI Activity ............................................................. 635 daniel l. millimet and abdullah kumas

contents ix

23. Do Double Taxation Treaties Increase Foreign Direct Investment to Developing Countries? ...................................................... 659 eric neumayer

part iv: exploring the impact of tax and foreign investment treaties on foreign direct investment flows 24. The Effect of Tax and Investment Treaties on Bilateral FDI Flows to Transition Economies ........................................................ 687 tom coupé, irina orlova and alexandre skiba

Selected Bibliography on Bilateral Investment Treaties and Double Taxation Treaties .................................................................... 715 lisa e. sachs

Index ................................................................................................................. 725

This page intentionally left blank

contributors emma aisbett Emma Aisbett is a lecturer at the Crawford School of Economics and Government and Research Fellow at the Economics Program of the Research School of Social Sciences at The Australian National University. Recently graduated from the University of California at Berkeley, her research interests center on the question of how globalization can be harnessed to promote sustainable development. Aside from her contribution to this volume, recent work has analyzed the sources of debate concerning the impact of globalization on poverty and considered the efficiency of the regulatory takings doctrine that is implied by the language of most bilateral investment treaties (BITs). reuven s. avi-yonah Reuven S. Avi-Yonah is the Irwin I. Cohn Professor of Law and Director of the International Tax LL.M. Program at the University of Michigan Law School. He teaches courses on taxation, international taxation, corporate taxation, tax treaties and transnational law. He has published numerous articles on domestic and international tax issues, and is the author of International Tax as International Law: U.S. Tax Law and the International Tax Regime (2007) and U.S. International Taxation: Cases and Materials (with Brauner and Ring, 2005). He holds a Ph.D. in History and a J.D. from Harvard University. He has served as consultant to the U.S. Treasury and the Organisation for Economic Co-operation and Development (OECD) on tax competition issues and has been a member of the executive committee of the New York State Bar Association Tax Section and of the Advisory Board of Tax Management, Inc. He is currently a member of the Steering Group of the OECD International Network for Tax Research and Chair of the ABA Tax Section VAT Committee and an International Research Fellow of the Oxford University Centre for Business Taxation. melissa b.l. birch Melissa B.L. Birch is at the Fletcher School and the Global Development and Environment Institute at Tufts University. Her expertise lies in the areas of economic and social development, environmental sustainability, and democracy, particularly in Latin America and the Caribbean. In addition to research relating to sustainable development, she has worked in education in Monterrey, Mexico, and in natural resource management and education in the United States. She holds an M.A. in International Relations and Environmental Policy from Boston University and is currently a doctoral candidate at the Fletcher School.

xii contributors

bruce a. blonigen Bruce Blonigen is the Knight Professor of Social Science in the Economics Department at the University of Oregon and a Research Associate with the National Bureau of Economic Research. He has research interests in empirically examining international trade issues from a microeconomic and political economy perspective, especially with respect to multinational corporations and antidumping policies. His research has been funded by the National Science Foundation and published in such journals as the American Economic Review, the European Economic Review, the Review of Economics and Statistics, the Journal of International Economics, and the Canadian Journal of Economics. He also currently serves as coeditor of the Journal of International Economics. tim büthe Tim Büthe is an assistant professor of political science at Duke University, currently on leave as a Robert Wood Johnson Foundation Scholar at the University of California, Berkeley. His primary research interests are the evolution and persistence of institutions and the ways in which institutions enable and constrain actors. His work in international political economy has focused on standards and regulations in product and financial markets, the effects of BITs and trade agreements on foreign direct investment, the allocation of private-source development aid by nongovernmental organizations, and more broadly the influence of noncountry actors in world politics. He has also written about methodological issues, the politics of business confidence, and European integration. His work has been published in the American Political Science Review, World Politics, Law & Contemporary Problems, Governance, and other journals. He is online at http:// www.buthe.info. allison christians Allison Christians is an assistant professor of law at the University of Wisconsin Law School. She received her J.D. from Columbia University School of Law and her LL.M. in Taxation from New York University School of Law. Prior to joining the faculty of the University of Wisconsin Law School, she taught J.D. and LL.M. courses in federal and international income taxation at Northwestern University School of Law, and before that practiced tax law at the law firms of Wachtell, Lipton, Rosen & Katz and Debevoise & Plimpton in New York, where she focused on the taxation of domestic and cross-border mergers and acquisitions; spinoffs; restructurings and associated issues; and transactions involving private and public companies. Her scholarly interests include foreign policy, globalization, competition, and development aspects of taxation. tom coupé Tom Coupé obtained his Ph.D. from the Free University of Brussels in 2002 and has since been working in Ukraine, initially as assistant professor and currently as

contributors xiii

director of the Kyiv School of Economics. His research interests are labor economics, transition economics, economics of education, and political economy. ronald b. davies Ronald B. Davies received his Ph.D. from the Pennsylvania State University in 1999. After spending nine years at the University of Oregon, he accepted a professor position at University College, Dublin, in Ireland. His research focuses on foreign direct investment, with a particular focus on the use of tax policy to manipulate the behavior of multinationals. His work has been published in journals such as American Economic Review, European Economic Review, Journal of Public Economics, and Journal of International Economics. john h. dunning John Dunning is an emeritus professor of international business at the University of Reading, UK, and at Rutgers University. He has been researching into the economics of international direct investment and the multinational enterprise since the 1950s. He has authored, coauthored or edited forty-two books on this subject, and on industrial and regional economics. His latest publications are a book of essays, Globalisation at Bay, a two- volume compendium of his more influential contributions to international business during the past 30 years (Edward Elgar, 2002), and a newly edited volume on Making Globalization Good (Oxford University Press, 2003). The revised edition of his textbook Multinational Enterprises and the Global Economy (with Sarianna Lundan), first published in 1993, was published in July 2008 by Edward Elgar. peter egger Peter Egger is a professor of economics at the University of Munich and head of the department at the Ifo Institute of Economic Research in Munich. He is interested in applied theoretical and empirical work in international trade and foreign direct investment. More specifically, his research interests include the determinants of bilateral trade and foreign direct investment, the determinants of economic policy choice with regard to taxation, tariffs, and investment costs, the endogenous choice of firm organization, and the role of imperfect labor markets for international trade and foreign investment. He has published in journals such as the Journal of Econometrics, Journal of International Economics, European Economic Review, Journal of Applied Econometrics, and Journal of Urban Economics. kevin p. gallagher Kevin P. Gallagher is an assistant professor of international relations at Boston University, where he also serves as a Research Fellow at the Frederick S. Pardee Center for the Study of the Longer-Range Future. He is the author of The Enclave Economy: Foreign Investment and Sustainable Development in Mexico’s Silicon

xiv contributors

Valley (with Lyuba Zarsky), and Free Trade and the Environment: Mexico, NAFTA, and Beyond, in addition to numerous reports, articles, and opinion pieces on trade policy, development, and the environment. He has been the editor or coeditor for a number of books, including Putting Development First: the Importance of Policy Space in the WTO and IFIs, International Trade and Sustainable Development, and others. He is also a research associate at the Global Development and Environment Institute of the Fletcher School of Law and Diplomacy and Tufts University, an adjunct fellow at Research and Information System for Developing Countries in Delhi, India, and a member of the U.S.-Mexico Futures Forum. robert grosse Robert Grosse is director of Global Leadership Development and Learning at Standard Bank in Johannesburg, South Africa. He has taught international finance in the M.B.A. programs at Thunderbird, the University of Miami, the University of Michigan, and at the Instituto de Empresa (Madrid, Spain), as well as in many universities in Latin America. He is a leading author on international business in Latin America. His latest book is Can Latin American Firms Compete? (Oxford, 2007). He holds a B.A. degree from Princeton University and a doctorate from the University of North Carolina, both in international economics. He is a Fellow of the Academy of International Business. andrew t. guzman Andrew T. Guzman is a professor of law and director of the International Legal Studies Program at Boalt Hall School of Law, at the University of California, Berkeley. He holds a J.D. and a Ph.D. in economics from Harvard University. He has written extensively on international trade, international regulatory matters, foreign direct investment and public international law, and served as editor on the recently published Handbook of International Economic Law and authored How International Law Works. He is a member of the Institute for Transnational Arbitration’s Academic Council and is on the board of several academic journals. He has taught as a visiting professor at Harvard Law School, the University of Chicago Law School, the University of Virginia Law School, Vanderbilt Law School, the University of Hamburg, and the National University Law School in Bangalore, India. mary hallward-driemeier Mary Hallward-Driemeier is a senior economist in the Development Research Group of the World Bank. She has published articles on firm productivity, the impact of the investment climate on firm performance and determinants of foreign direct investment. She was the deputy director for the World Development Report 2005: A Better Investment Climate for Everyone. She helped establish the World Bank’s Enterprise Surveys Program, now covering more than 70,000 enterprises in 100 countries. She is also a founding member of the Microeconomics

contributors xv

of Growth Network. She received her M.Sc. in development economics from Oxford University as a Rhodes Scholar and received her Ph.D. in economics from M.I.T. abdullah kumas Abdullah Kumas received his B.S. in mathematics education from Bogazici University, Turkey, in 2000, his M.S. in applied mathematics from Oklahoma State University in 2004, and his M.A. in economics from Southern Methodist University (SMU) in 2006. He is now a fourth-year Ph.D. student at SMU. His research focuses mainly on international trade, applied econometrics, and labor economics. henry j. louie Henry Louie is an international economist in the U.S. Treasury Department’s Office of Tax Policy. He joined the Treasury Department in 1996. His principle work areas encompass economic research, analysis of current law and legislative proposals relating to international taxation, and the negotiation of bilateral income tax treaties on behalf of the United States. He represents the United States at the OECD’s working party that examines issues regarding the OECD Model Tax Convention. He received his M.A. in economics from Duke University in 1995 and his B.A. in 1990 from Georgetown University. andreas f. lowenfeld Andreas F. Lowenfeld is Rubin Professor of International Law at New York University Law School, where he has been on the faculty since 1967. He has taught, practiced, and written in nearly all aspects of international law for more than five decades and is frequently an arbitrator in international controversies, public and private. He is the author of a major treatise on International Economic Law, as well as casebooks and textbooks on conflict of laws and aviation law and a series of teaching books on international economic law. He was associate reporter of the American Law Institute’s Restatement (Third) of Foreign Relations Law, and coreporter of the Institute’s Project on International Jurisdiction and Judgments. He is an elected member of the Institut de Droit International and of the International Academy of Comparative Law. He is a graduate of Harvard College and Harvard Law School. mario larch Mario Larch is a researcher in the International Trade and Foreign Direct Investment Department at the Ifo Institute for Economic Research. He has research interests in the theory of multinational firms and trade, international economics, economic geography, and spatial econometrics. He has published in such journals as the European Economic Review, the Canadian Journal of Economics and the Journal of Comparative Economics.

xvi contributors

daniel l. milliment Daniel L. Millimet received his B.A. in economics from the University of Michigan in 1994, and his Ph.D. in economics from Brown University in 1999. He is currently an associate professor in the Department of Economics at Southern Methodist University in Dallas, Texas. He also serves as a member of the Editorial Council for the Journal of Environmental Economics and Management. His research focuses mainly on the empirical analysis of issues related to international trade, environmental quality, and schooling. helen v. milner Helen V. Milner is the B.C. Forbes Professor of Politics and International Affairs at Princeton University, the chair of the Department of Politics, and the director of the Center for Globalization and Governance at Princeton’s Woodrow Wilson School. She has written extensively on issues related to international trade, the connections between domestic politics and foreign policy, globalization and regionalism, and the relationship between democracy and trade policy. Some of her writings include The Political Economy of Economic Regionalism (coedited with Edward Mansfield, 1997), Internationalization and Domestic Politics (coedited with Robert Keohane, 1996), “Why the Move to Free Trade? Democracy and Trade Policy in the Developing Countries” (International Organization, 2005), “Why Democracies Cooperate More: Electoral Control and International Trade Agreements” (coauthored with Edward Mansfield and B. Peter Rosendorff, International Organization, 2002), and “The Optimal Design of International Institutions: Why Escape Clauses are Essential.” (coauthored with B. Peter Rosendorff, International Organization, 2001). peter muchlinski Peter Muchlinski is a professor in international commercial law at the School of Oriental and African Studies, University of London. He is the author of Multinational Enterprises and the Law (Second edition, Oxford University Press, 2007) and is coeditor (with Dr Federico Ortino and Professor Christoph Schreuer) of the Oxford Handbook of International Investment Law (Oxford University Press, 2008). He acts as an adviser to the United Nations Conference on Trade and Development (UNCTAD) on investment law issues. He is corapporteur to the International Law Association Committee on the International Law on Foreign Investment and occasionally advises in international investment arbitrations. eric neumayer Eric Neumayer is a professor in the department of geography and environment at the London School of Economics and Political Science since 1998. Before, he was an academic assistant at the Centre for Law and Economics at the University of Saarbrücken, Germany. An economist by training, he is the coeditor of the Handbook of Sustainable Development (with Giles Atkinson and Simon Dietz), the

contributors xvii

author of Weak versus Strong Sustainability: Exploring the Limits of Two Opposing Paradigms, Greening Trade and Investment: Environmental Protection Without Protectionism and The Pattern of Aid Giving: The Impact of Good Governance on Development Assistance, as well as numerous journal articles. He has broad research interests that all relate to evidence-based public policy. irina orlova Irina Orlova is a 2005 graduate of Economics Education and Research Consortium M.A. Program in Economics. She is currently working with CASE Ukraine. Her research interests lie in the areas of international trade, capital flows, and transition economies. michael pfaffermayr Michael Pfaffermayr is a professor of international economics in the Department of Economic Policy, Economic Theory and Economic History at the University of Innsbruck, Austria. He is managing editor of EMPIRCA, a CESifo Research Fellow and an International Research Fellow at the Oxford University Centre for Business Taxation. His research interests include international and industrial economics, applied econometrics, and especially foreign direct investment as it relates to multinational firms and trade. Currently, a major focus of his research is international tax competition. He has authored a number of publications. susan rose-ackerman Susan Rose-Ackerman is the Henry R. Luce Professor of Jurisprudence (Law and Political Science) and codirector of the Yale Law School’s Center for Law, Economics, and Public Policy. She has held fellowships from the Guggenheim Foundation and the Fulbright Commission and was a Visiting Research Fellow at the World Bank and a fellow at the Center for Advanced Study in the Behavioral Sciences, Stanford, California. She is the author of Corruption and Government: Causes, Consequences and Reform, 1999 (translated into thirteen languages); From Elections to Democracy: Building Accountable Government in Hungary and Poland (2005); Controlling Environmental Policy: The Limits of Public Law in Germany and the United States (1995); Rethinking the Progressive Agenda: The Reform of the American Regulatory State (1992); and Corruption: A Study in Political Economy (1978). She holds a B.A. from Wellesley College and a Ph.D. in economics from Yale University. donald j. rousslang Dr. Donald Rousslang received his Ph.D. in economics from the University of Oregon in 1974. He has served as a senior economist in the U.S. Department of Labor, as chief of the Research Division in the U.S. International Trade Commission, and as a senior economist with the Department of Treasury. He is currently a tax specialist with the Department of Taxation in Hawaii. He has also

xviii contributors

taught part-time at George Mason University, George Washington University and the University of Hawaii at Manoa. His main interests are applied microeconomics, international finance, and public finance. lisa e. sachs Lisa Sachs is the program coordinator at the Vale Columbia Center on Sustainable International Investment at Columbia University. She received a J.D. and a Master of International Affairs from Columbia in May 2008. Her academic research has focused on foreign investment, corporate responsibility, human rights, and economic development. jeswald w. salacuse Jeswald W. Salacuse is Henry J. Braker Professor of Law at the Fletcher School of Law and Diplomacy at Tufts University. He is the author of several books, including most recently The Global Negotiator (2003), Leading Leaders (2006) and Seven Secrets for Negotiating with Government (2008). He is a member of the Council on Foreign Relations and the American Law Institute and serves as president of an ICSID international investment arbitration tribunal. karl p. sauvant Karl P. Sauvant is the founding executive director of the Vale Columbia Center on Sustainable International Investment, Research Scholar and Lecturer in Law at Columbia Law School, codirector of the Millennium Cities Initiative, and guest professor at Nankai University, China. Before that, he was director of UNCTAD’s Investment Division. He is the author of, or responsible for, a substantial number of publications. In 2006, he was elected an Honorary Fellow of the European International Business Academy. He received his Ph.D. from the University of Pennsylvania in 1975. alexandre skiba Alexandre Skiba received his Ph.D. from Purdue University in 2003. Since then, he has taught at Purdue University, the University of Kansas, and the University of Wyoming. His research focuses on empirical investigations of international trade. laura spess Laura Spess is a Ph.D. candidate in the Department of Geography at Pennsylvania State University. Her current research focuses on intranational migration in the developing country context. She has coauthored papers on poverty and fertility in developing countries and urban-rural differences in mortality among older adults in China. nicholas p. sullivan Nicholas P. Sullivan is the author of You Can Hear Me Now: How Microloans and Cell Phones Are Connecting the World’s Poor to the Global Economy, and

contributors xix

the publisher of Innovations: Technology/Governance/Globalization. He was a United Nations-accredited business interlocutor to the International Financing for Development conference. He is a partner in the Global Frontier Fund, a private-equity fund of local funds in emerging markets, for which he compiles the annual Wealth of Nations Index, a ranking of 70 developing countries. A graduate of Harvard University and the Fletcher School of Law and Diplomacy, he is currently a visiting fellow at the Feinstein International Center at Tufts and a visiting scholar at the Legatum Center for Development & Entrepreneurship at Massachusetts Institute of Technology. deborah l. swenson Deborah L. Swenson is a professor of economics at the University of California, Davis, and a research associate with the National Bureau of Economic Research. Her research, which focuses on issues relating to the international location decisions of global firms, has appeared in academic journals, including the American Economic Review, Review of Economics and Statistics, Journal of International Economics, and Canadian Journal of Economics. Much of her work involves empirical analysis of foreign investment and outsourcing decisions that examine how the international differences in costs and policy environment affect the operating choices of multinational firms. She also investigates how multinational firms affect the global environment, by studying the impact of multinational firms on local economic performance and by studying the implications of multinational firm behavior for international tax systems. len j. trevino Len J. Trevino, Ph.D., holds the Gerald N. Gaston Eminent Scholar Chair in International Business in The Joseph A. Butt, S.J. College of Business at Loyola University New Orleans. His expertise lies in strategic management/policy, management of the multinational enterprise, global business strategy, foreign direct investment, and strategic management in emerging markets. His research centers on the theory of the multinational enterprise, foreign direct investment, and the intersection of strategic management and international business. He has published in many top academic journals, including Journal of International Business Studies, Management International Review, Journal of World Business, International Business Review, Journal of Labor Research, Business Horizons, and Transnational Corporations, among others. He has consulted with such organizations as Dow Brands, Eli Lilly, Monsanto, and the United Nations Conference on Trade and Development. kenneth j. vandevelde After graduating from Harvard Law School in 1979, Kenneth J. Vandevelde entered private practice in Washington, D.C. In 1982, he joined the Office of the Legal Adviser at the U.S. Department of State, where his responsibilities included arbitration of investment claims before the Iran-United States Claims Tribunal

xx contributors

and the negotiation of bilateral investment treaties. In 1988, he left the State Department to teach and to write his book, United States Investment Treaties: Policy and Practice. Since then, he has published numerous articles on bilateral investment treaties; has spoken about these agreements in Europe, Asia, Africa, and North and South America; and has served as a consultant on bilateral investment treaties to the U.S. Senate Foreign Relations Committee to foreign governments and to counsel for private investors. He currently is Professor of Law at Thomas Jefferson School of Law in San Diego, California, where he served as Dean from 1994 to 2005. hannes winner Hannes Winner is professor of economics and public finance at the University of Innsbruck. He is also adjunct professor at the Free University of BolzanoBozen (Italy). In 2003, he was visiting professor at the European University Institute in Florence. His fields of interest include public economics and economics of taxation, health economics, and econometrics. He has published in academic journals including International Tax and Public Finance, Regional Science and Urban Economics and the Canadian Journal of Economics. His current research is on the empirical impact of taxation on international production and location decisions of multinational firms. jason yackee Jason Yackee is an assistant professor of law at the University of Wisconsin. He holds a Ph.D. in political science from the University of North Carolina, Chapel Hill, and a J.D. from Duke University School of Law. His research centers on international investment law, international economic relations, foreign arbitration, and administrative law. His research on BITs and other topics has been published in a number of peer-reviewed journals and law reviews, including the Journal of Politics, International Politics and the Duke Law Journal.

foreword It would be tactless in a foreword to propose answers to the puzzle of bilateral investment treaties (BITs) and double taxation treaties (DTTs). Perhaps, though, it is not inappropriate to pose some questions that may whet the appetite of the reader. Although many questions about the diffusion and implications of these bilateral treaties are relevant to both BITs and DTTs, I focus on BITs, a topic on which many academic and policy discussions have centered. BITs (and investment dispute provisions in trade agreements, such as NAFTA) keep sprouting up—more than 2,500 as of the publication of this book. Nevertheless, all efforts to reach agreement on a multilateral investment treaty (in the United Nations, in the Organisation for Economic Co-operation and Development, in the World Trade Organization, and elsewhere) have failed, even though the substantive provisions in the proposals read just like BITs. How can this be explained? A number of observers, including this writer, have reached for the conclusion that the common provisions of BITs—non-discrimination, open access for investors, fair and equitable treatment, expropriation only for a public purpose and subject to full compensation, arbitration of investor-state disputes—now comprise or reflect the customary international law of foreign investment. Is this persuasive? Or is there no such law, as shown by the failure to achieve a multilateral agreement? Note that the answer to this question is not just for the scholar, but becomes critical in dispute settlement. A customary law of foreign investment, or even of interpretation of similar BITs, would mean that arbitrators hearing a dispute between a Xandian investor and the state of Patria could (and should?) rely on (or at least be guided by) decisions involving treaties of Tertia, Quarta, Quinta, etc. An opposite answer would make resolution of each dispute into a journey from square one. Why do developing countries in huge numbers enter into BITs and DTTs anyway? In some instances, the answer is clear. When, in the early 1990s, President Carlos Menem saw the way out of Argentina’s doldrums to be privatization of state-run utilities and other monopolies, he needed foreign capital. Foreign private capital, however, would come in only on the basis of a stable currency linked to the dollar and a bilateral investment treaty with each potential investor’s home state. When Mexico wanted to take part in the free trade arrangements between the United States and Canada, the anticipation of increased investment in Mexico to produce goods for the United States market depended on a secure investment protection regime as a necessary component of the North American Free Trade Agreement. But in other cases the answer is not so clear. Do BITs and DTTs actually attract foreign investment? Or is it true that absence

xxii foreword

of an applicable BIT or DTT discourages potential investors? And how can one really tell? More broadly, is the presence or absence of a BIT really an indicator of a good or bad investment climate, as stated, for instance, in the Convention Establishing MIGA, the Multilateral Investment Guarantee Agency sponsored by the World Bank? Many critics (not only in Latin America, but also in the United States and Canada) have pointed out that under a BIT and its analogues in free trade agreements, a foreign investor may enjoy greater legal protection than a domestic investor would. U.S. jurisprudence under the “just compensation” clause of the Fifth Amendment has a long and sometimes confusing tradition distinguishing between a “taking” (entitled to compensation) and an exercise of regulatory powers (generally not entitled to compensation). The Canadian constitution does not contain a property clause at all. Most Latin American constitutions contain variations on the concept of the “social function of property.” Typically BITs contain a stricter form of protection for investors in case of expropriation or deprivation of operating rights than is available under the law or in the courts of the host state. The answer of the proponents of BITs is that the criticism may be true, but the object of the exercise is precisely to give additional protection in order to encourage cross-border investment. If, then, BITs are not supposed to be neutral (as between domestic and foreign investment) can they nevertheless be fair? Finally, for this menu of appetizers, why do the industrial countries advocate for these treaties? To illustrate this trend, consider the fact that there are currently 103 treaties for France, 83 for Italy, 147 for the Germany, 102 for the United Kingdom, and 48 for the United States.1 Policies are rarely one-dimensional, and my impression is that BITs have rarely been subject to serious debate in the developed countries. Foreign investment was accepted as a natural feature of major corporations, nearly all of which became multinational well before globalization swept across the planet. If the business community wanted BITs and they did not cost much, why not employ them? But does not foreign direct investment encourage outsourcing, loss of domestic jobs, and a drain on the balance of payments? Can one justify BITs and DTTs on the basis that public sector foreign assistance (whether bilateral or through the World Bank and regional development banks) does not work? Private investment, by contrast, is designed to bring what the public sector cannot bring: technology, management skills, and access to global markets. Maybe foreign direct investment can act as a catalyst, contributing to a culture of incentives and innovation that will lift a country out of poverty. If so, by 1 Figures from ICSID. The U.S. total does not include NAFTA and Free Trade Agreements with Australia, Chile, Singapore, and others that contain chapters substantially replicating BITs.

foreword xxiii

encouraging their corporations to invest in developing countries, can the industrial countries justify their minimal commitment—0.47% of GDP for France and the United Kingdom, 0.36% for Germany, 0.28% for Japan, 0.22% for the United States—to public sector aid?2 My task was to raise a few preliminary questions. For answers—though not necessarily the answers—I invite the reader to proceed to the main text. Andreas F. Lowenfeld

2 Figures from The Economist, Pocket World in Figures, 2008 Edition.

This page intentionally left blank

preface The critical role of foreign direct investment (FDI) in global economic growth and development is by now widely recognized. In addition, the activities of multinational enterprises (MNEs) are acknowledged as some of the most dynamic components of the international division of labor. Over the past two decades, policy makers have increasingly come to appreciate that FDI is crucial to a country’s economic success. Past institutions and government strategies restrictive to FDI inflows have generally given way to those geared toward attracting and retaining such resource transfers. These have included several waves of investment liberalization, an increasing variety of investment incentives, and additional protections for foreign investors. This volume focuses on these institutions and, more specifically, on the extensive network of bilateral investment and tax treaties that have emerged over the past two decades. Inter alia, and in the light of global economic events, such measures have both upgraded the standards of treatment toward foreign investors and created new legal processes to protect their rights. The burgeoning number of bilateral investment treaties (BITs) and double taxation treaties (DTTs) also reflects the changing nature of inbound foreign investment flows, and the locational decisions of MNEs. Whereas, in the past, sitings of foreign direct investments were primarily determined by resource endowments and the size and structure of local and adjacent markets, today’s locational decisions are far more complex. Government policies and institutions have become increasingly important relative to the traditional economic fundamentals, especially for FDI from and to the advanced industrialized countries. For this reason, host economies seeking to attract inbound MNE activity are increasingly focusing their attention on improving the investment environment by adopting domestic institutions and policies that favor FDI, concluding international investment agreements, guaranteeing certain standards of protection, and providing favorable tax treatment for foreign investors. With the growing significance of these investment agreements—some 5,000 BITs and DTTs have been signed so far—it is time to evaluate whether, and in what ways, these treaties have had a positive impact on attracting the kind of FDI inflows sought by host countries. The present volume sets out a comprehensive and impressive collection of studies from the past decade on the impact of BITs and DTTs on FDI flows. As the editors note in the Introduction, the debate has been lively and varied, but is, as yet, inconclusive. The contributions in this volume illustrate some of the difficulties in identifying and evaluating the numerous factors that influence FDI flows, including the specific impact of bilateral treaties. Nevertheless, the debate

xxvi preface

about the effect that BITs and DTTs have on FDI remains relevant and important, especially as host countries incur sovereignty costs and risk costly arbitration in their attempts to attract and retain foreign direct investment. While this volume will not provide the reader with definitive answers on the roles of BITs and DTTs, it does bring together the most important studies on this issue. It is therefore an essential resource for those who wish to understand this particular debate. By placing their significance and contents in historical and economic perspective, the editors’ Introduction gives a helpful and concise introduction to BITs and DTTs. Rather than attempting to extract a simplified conclusion from the varied studies, they provide a useful overview of the key issues related to investment agreements and FDI flows. Importantly, the studies that follow include contributions from leading economists, legal scholars, and political scientists, who frame the debate from multiple angles and highlight its contextual complexity. In my opinion, this volume will be extremely useful for policy makers, practitioners, the investment community, political scientists, economists, international investment lawyers, and, indeed, all those interested in the field of foreign direct investment and the role of MNEs in international relations. John H. Dunning

bit s , dtt s , and fdi flows: an overview∗ lisa e. sachs and karl p. sauvant

In the past two decades, foreign direct investment (FDI) has been spurred by the widespread liberalization of the FDI regulatory framework, combined with advances in information and communication technologies and competition among firms. Most countries have opened themselves to foreign investment, improved the operational conditions for foreign affiliates and strengthened standards of treatment and protection. In fact, virtually all countries now actively encourage FDI, as it can bring capital, technology, skills, employment, and market access. Investment promotion strategies include the establishment of Investment Promotion Agencies (IPAs), the offering of incentives, the preparation and dissemination of investment guides, and, notably, the conclusion of international investment agreements, especially bilateral investment treaties (BITs) and double taxation treaties (DTTs). For countries, the basic purposes of concluding BITs and DTTs are, respectively, to assure investors that investments will be legally protected under international law and to mitigate the possibility of double taxation of foreign entities and, in this manner, to help increase FDI inflows. Whether BITs and DTTs do indeed affect the flow of FDI has been studied and debated for the past decade. This volume brings together published studies, updated articles and original pieces from that period dealing with that subject matter. Its focus is on BITs and DTTs because these (and especially the former) are the principal international investment agreements (IIAs), that is, instruments that, in a significant manner, address investment issues.1 While the focus of this ∗ The authors wish to thank José Alvarez, Reuven Avi-Yonah, Andrea Bjorklund, Dali Bouzoraa, Olivier De Schutter, Lorraine Eden, Susan Franck, Mark Kantor, Luke Peterson, Lauge N. Poulsen, Jeswald W. Salacuse, Jan Peter Sasse, Christoph Schreuer, Kenneth Vandevelde, and Christopher Wilkie for their very helpful comments and clarifications. Thanks are also due to Hamed El-Kady and Masataka Fujita for providing extensive data, to Matthew Beck for creating the graphs and tables, and to Matthew Quint, Chryse Bautista and especially Michael O’Sullivan and Kimmie Lathana for their instrumental assistance in preparing this volume. Finally, we wish to thank the publishers of the reprinted papers included in this volume for their permission and assistance. Any errors or omissions are of course those of the authors. 1. IIAs also include a number of bilateral free trade agreements with substantial investment chapters and regional and multilateral instruments on investment; studies of the effect of these agreements on FDI flows are not covered in this volume, although the number of

xxviii lisa e. sachs and karl p. sauvant

compendium is on the impact of BITs and DTTs on FDI flows, the chapters in the first section of this volume discuss the general nature of these agreements to facilitate the understanding of the subject matter; additional materials are listed in the bibliography. The present overview looks briefly at current FDI trends and their salient features, provides a general introduction to BITs and DTTs, and summarizes the findings of the studies contained in this compendium as to the impact of these agreements on FDI flows. A. FDI Trends and Characteristics The IMF and OECD define direct “foreign investment” as cross-border investment made by a resident entity in one economy (the “direct investor” or “multinational enterprise”) with the objective of establishing a lasting interest in an enterprise resident in an economy other than that of the direct investor (the “foreign affiliate”).2 “Lasting interest” implies the existence of a long-term relationship between a direct investor and a foreign affiliate, and a significant degree of influence on the management of the latter. A minimum stake of 10% of the ordinary shares of an enterprise is generally regarded as being the minimum threshold for a foreign investment to be classified as a direct investment for statistical purposes.3 FDI is typically measured in either inflows or stocks. “FDI inflows” refer to the capital provided by a foreign investor to a foreign affiliate (equity, loans, reinvestment earnings), while “FDI stocks” are the total value of foreign-owned assets at a given time. Over the past twenty years, FDI inflows have expanded substantially, from approximately $40 billion at the beginning of the 1980s, to $200 billion in 1990, to some $1.5 trillion in 2007 (Figure 1). Cross border mergers and acquisitions (M&As) are the principal drivers of this growth, as they are the main form of FDI in the developed world and an increasingly important one in emerging markets.4

countries covered by them rivals that of countries covered by BITs and DTTs. For a survey of studies of the impact of the NAFTA investment chapter, see José Alvarez, “The NAFTA’s Investment Chapter and Mexico,” in Rudolf Dolzer, Matthias Herdegen and Bernhard Vogel, eds. Foreign Investment: Its Significance in Relation to the Fight Against Poverty, Economic Growth and Legal Culture (Berlin: Konrad Adenauer Stiftung Foundation, 2006), p. 253. 2. OECD, Detailed Benchmark Definition of Foreign Direct Investment (Paris: OECD, 1996), third edition, and International Monetary Fund, Balance of Payments Manual (Washington: IMF, 1993), fifth edition. 3. It should be noted that the definition of “investment” in BITs is typically much broader and includes “everything of economic value, virtually without limitation,” in order to ensure flexibility in the BIT’s application. Calvin Hamilton and Paula Rochwerger, “Trade and investment: foreign direct investment through bilateral and multilateral treaties,” 18 N.Y. Int’l L. Rev. 1 (2005), p. 12. 4. In this overview, developed countries are members of the OECD, excluding the Republic of Korea and Mexico. Economies in transition are the former Socialist countries,

bit s , dtt s , and fdi flows: an overview xxix

figure 1. fdi inflows, global and by group of economies, 1980–2011* (trillion us dollars) 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

World

Emerging markets

Developed economies

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

Source: UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. 3, and Laza Kekic and Karl P. Sauvant, eds., World Investment Prospects to 2011: Foreign Direct Investment and the Challenge of Political Risk (London: The Economist Intelligence Unit, 2007). ∗ FDI inflow projections for 2007–2011 are derived from data in World Investment Prospects, whose regional definitions vary slightly from World Investment Report data.

The value of the global inward FDI stock is expected to climb to about $14 trillion by the end of 2007 (Figure 2). There are more than 80,000 multinational enterprises (MNEs) globally, with more than 800,000 foreign affiliates.5 The developed countries attract the lion’s share of world FDI flows (nearly two-thirds in 20076), with Asia being the most attractive region among emerging markets (Figure 3). Some two-thirds of world FDI inflows, and half of FDI inflows in developing countries, are in services.7

except if otherwise noted, and developing countries are all remaining countries. Emerging markets are economies in transition and developing countries combined. These categories vary slightly among different sources, so the data may reflect slight variations in country classifications. Similarly, the studies of the impact of BITs and DTTs included in this volume may employ slightly varied categorizations. 5. UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. 218. 6. Laza Kekic and Karl P. Sauvant, eds., World Investment Prospects to 2011: Foreign Direct Investment and the Challenge of Political Risk (London: The Economist Intelligence Unit, 2007), p. 6 (table 1). 7. As of 2005. UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. xvi.

xxx lisa e. sachs and karl p. sauvant

20 18 16 14 12 10 8 6 4 2 0

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

figure 2. world inward fdi stock, 1980–2011* (trillion us dollars)

Source: UNCTAD (http://stats.unctad.org/fdi/), and Kekic and Sauvant, op. cit. ∗Inward FDI stock projections for 2007–2011 are derived from data in World Investment Prospects which varies slightly from World Investment Report data.

Global FDI flows over the next few years will depend on the principal FDI determinants. The regulatory framework, on balance, will most likely remain favorable, with a further liberalization of FDI laws and regulations and the strengthening of the international investment law system; however, there are signs of a backlash against FDI that make the regulatory framework less welcoming in a number of countries.8 Investment promotion, too, will continue, although countries may shift toward a more targeted approach. Finally, in light of the turmoil in financial markets and the effect this may have on the real economy, economic growth—the principal FDI determinant—may decrease substantially or even turn negative in key economies, negatively affecting FDI flows. Current estimates are that this combination of factors will mean that such flows will remain at a plateau of about $1.4–$1.6 trillion in the next few years, with the caveat that the current economic turmoil does not turn into a widespread recession.9 Twelve of the top 20 FDI recipients are developed countries, with the top 20 accounting for three-quarters of world FDI inflows. The United States is the leading recipient and will likely retain its dominant position in 2007–2011 (Table 1). However, FDI into the EU as a whole is significantly higher than inflows into the United States, and the EU will continue to outstrip the United States as a host region for such investment.10 Among emerging markets, concentration also

8. Karl P. Sauvant, “Regulatory risk and the growth of FDI,” in Kekic and Sauvant, op cit., p. 71. 9. Kekic and Sauvant, op. cit., p. 6. 10. These data include intra-EU flows—that is, FDI flows from one EU country to another. If intra-EU flows were excluded from the calculation, FDI inflows to the United States would exceed inflows to the EU.

bit s , dtt s , and fdi flows: an overview xxxi

figure 3. fdi inflows into emerging markets, 2004–2006 (billion us dollars) 2004

2005

2006

$250

$200

$150

$100

$50

$0 Developing Asia

Latin America & Caribbean

Economies in transition

Middle East

North Africa

Sub-Saharan Africa

Source: Kekic and Sauvant, op. cit. table 1. fdi inflows, 2007–2011 average (billion us dollars and percent) Economy

US UK China France Belgium Germany Canada Hong Kong, China Spain Italy Netherlands Australia Russia Brazil Singapore Sweden Mexico India Ireland Turkey

Value

251 113 87 78 72 66 63 48 45 42 39 38 31 28 27 26 23 20 20 20

Source: Kekic and Sauvant, op. cit., p. 9.

Rank

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Share in world Total 17 8 6 5 5 4 4 3 3 3 3 3 2 2 2 2 2 1 1 1

xxxii lisa e. sachs and karl p. sauvant figure 4. the top ten emerging-market fdi recipients, 2006 (billion us dollars) 80

78.1

70 60 50

42.9

40 30

28.7

25.7 20.1

20

19.0

18.8

17.5

16.0

14.5

India

UAE

Poland

10 0 China Hong Kong Russia Singapore Turkey

Mexico

Brazil

Source: Laza Kekic, “Global foreign direct investment to 2011,” in Kekic and Sauvant, op. cit., p. 24.

remains relatively high, with the top ten recipients accounting for 55% of all inflows to emerging markets in 2006. China was by far the main FDI host among emerging markets in 2006 (Figure 4), with almost 6% of the global total, and is expected to rank behind only the United States and the United Kingdom in 2007–2011. While the vast majority of FDI flows emanate from developed countries, companies from emerging markets (mostly in Asia) are increasingly becoming important players in the world FDI market. An estimated 20,000 MNEs are now headquartered in emerging markets (Figure 5), and outward FDI flows from these economies rose to approximately $210 billion in 2006, or 17% of the global total (Table 2). The stock of this investment amounts to an estimated $1.8 trillion.11 Until relatively recently, most FDI flows from emerging markets took the form of South-South investment. But MNEs from these economies have also in more recent years undertaken some large, high-profile acquisitions in developed countries that have attracted considerable attention. The rise of multinationals from the South—especially those of state-controlled entities (including sovereign wealth funds)—is feeding rising protectionist and nationalist sentiment in parts of the developed world, which makes it all the more important to keep the development of emerging-market outward FDI in perspective. Despite the increase in these FDI flows in recent years (likely to be boosted further by

11. UNCTAD, World Investment Report 2007, op. cit., p. 255. Note that this value is based on a different dataset from the values for FDI flows, and the classification of “emerging economies” may vary slightly.

bit s , dtt s , and fdi flows: an overview xxxiii

figure 5. number of mne s from developed countries and emerging markets, 1992, 2000 and 2006 (thousands) 80,000 70,000

20,172

60,000

13,368

50,000 40,000

3,100

30,000

58,239

49,944 20,000

33,500

10,000 0 1992

2000 Developed countries

2006 Emerging markets

Source: UNCTAD, World Investment Reports 1993, 2003, and 2007. Annex tables.

investments by sovereign wealth funds), they are still dwarfed by investment originating in the developed world. B. International investment agreements As countries increasingly opened their doors to FDI in the 1980s and 1990s, they simultaneously entered into numerous international investment agreements, leading especially to an explosion in the number of bilateral investment treaties (BITs) and bilateral double taxation treaties (DTTs) (Figure 6). 1. Bilateral investment treaties The FDI surge during the past few decades has been accompanied by a similar growth of international investment agreements. Pride of place among these agreements belongs to BITs—treaties that seek to protect and promote foreign investment. To put the evolution of BITs

table 2. global fdi outflows (billion us dollars and percent) Country group Developed countries Share in total Emerging markets Share in total

1999

2000

2001

2002

2003

2004

2005 2006

1,004 93 73 7

1,104 87 163 13

687 89 81 11

477 88 64 12

505 93 41 8

723 88 94 12

731 82 160 18

Source: Kekic, “Global foreign direct investment to 2011,” in Kekic and Sauvant, op. cit., p. 28.

1,051 83 210 17

xxxiv lisa e. sachs and karl p. sauvant

Cumulative BITs & DTTs

figure 6. the growth in the number of bit s and dtt s , 1960–2006 (number) 3000 2500 2000 1500 1000 500 0

60

19

70

19

90

80

19

19

00

20

06

20

Years BITs

DTTs

Source: UNCTAD (http://www.unctad.org/iia).

into a historic context, Kenneth Vandevelde (Chapter 1) traces the history of such agreements. He distinguishes three eras of BIT development (Colonial Era, Post-Colonial Era, Global Era), and describes how investment agreements have been shaped by the political, economic and legal contexts of each period. He also discusses broadly the evolution of the content of BITs as well as the legal enforceability of their substantive provisions, and articulates several current developments that may herald a fourth era in their development. By the end of 2006, 2,573 BITs had been signed,12 most of them since 1990. In fact, from 1959 (when the first BIT was concluded between Germany and Pakistan) until the end of 1989, only 386 BITs had been signed; more than 2,000 BITs were entered into in the following 15 years. By the end of 2006, 177 countries had entered into one or more bilateral investment treaties. (UNCTAD has the best database of BITs, available on its website at http://www.unctad.org/ iia.) While BITs were originally signed overwhelmingly between developed and developing countries, developing countries now routinely sign investment treaties with other developing countries (and economies in transition). Indeed, 680 BITs had been signed between developing countries by the end of 2006, constituting 27% of the stock of BITs (Figure 7). The economies with the most BITs are led by Germany, China, and Switzerland (Figure 8).

12. J. Zhan, J. Karl and J. Weber, “International investment rule-making at the beginning of the 21st century: stocktaking and options for the way forward,” in José Alvarez and Karl P. Sauvant, with Kamil Gérard Ahmed, eds., The Future of International Investment Law and Policy (forthcoming).

bit s , dtt s , and fdi flows: an overview xxxv

figure 7. bit s concluded as of end 2006, by country group (percent) 27 3 13

7

10

40

Between developing countries Between developed and developing countries Between developing countries and South-East Europe & Commonwealth of Independent States Between developed countries Between developed countries and South-East Europe & Commonwealth of Independent States Between countries of South-East Europe and Commonwealth of Independent States

Source: UNCTAD, World Investment Report 2007, op. cit., p. 17.

figure 8. the ten countries with the highest number of bit s , june 2007 135

Germany China

119 114

Switzerland 103

United Kingdom Egypt

100 100

Italy

98

France 91

Netherlands Republic of Korea

86

Belgium and Luxembourg*

84 0

Source: www.unctad.org/iia.

20

40

60

80

100

120

140

160

xxxvi lisa e. sachs and karl p. sauvant

In addition to the conclusion of new BITs, countries are increasingly renegotiating existing treaties; there have been 121 renegotiated treaties by June 2008.13 Some countries are renegotiating treaties due to changed circumstances, such as to bring existing BITs in line with commitments under other investment agreements (for instance when the Central European countries acceded to the EU in 2004), or to add or update certain provisions, such as dispute settlement clauses. Although most BITs provide for tacit renewal upon their expiration, some countries are renegotiating expired BITs to amend host-country commitments or to clarify existing provisions. While many renegotiations are intended to strengthen investor protections, some renegotiated BITs narrow investor protections—at least in some respects.14 An increasing number of BIT renegotiations is expected in the coming years, as many BITs with a duration of ten to thirty years were signed in the 1990s. BITs only become legally binding instruments when they enter into force. Although the signing of a BIT may have some legal consequences for host countries under international law, this act does not “establish legally binding obligations of the latter vis-à-vis the foreign investors.”15 Some BITs stipulate that the agreement enters into force upon the signature of both parties. Most BITs, however, require each party to complete the domestic requirements necessary for the BITs’ entry into force, for instance the ratification by a national parliament and the notification of ratification to the treaty partner. By the end of 2005, 76% of all BITs signed until that point had entered into force; this share partly reflects the time lag due to relatively complicated domestic ratification processes. The share of BITs in force increases with respect to earlier BITs; for instance, more than 90% of the BITs signed in the first half of the 1990s have entered into force.16 BITs are agreements between two sovereign states. From the point of view of the capital-importing country, their basic purpose is to help to attract FDI.

13. UNCTAD Research Note, “Recent developments in international investment agreements,” UNCTAD/WEB/ITE/IIT/2005/1, available at: http://www.unctad.org/sections/ dite_dir/docs//webiteiit20051_en.pdf 14. Examples of the former are the extension of the prohibition of performance requirements and the strengthening of transparency provisions; examples of the latter are clarification concerning regulatory takings and fair and equitable treatment. 15. “The entry into force of bilateral investment treaties,” IIA Monitor, No. 3 (2006), UNCTAD/WEB/ITE/IIA/2006/9, p. 4. (“Before a treaty enters into force, contracting parties have a general obligation to refrain from acts that would defeat the object and purpose of the agreement. See Article 18 of the Vienna Convention on the Law of Treaties.”) In addition to the international law obligation not to defeat the object and purpose of a treaty that a country has signed but not yet ratified, a country’s domestic law (or policy) may well be to give that obligation effect by providing an investor a right to challenge a law that an investor argues violates the “object and purpose” of a signed but unratified BIT. 16. “The entry into force of bilateral investment treaties,” IIA Monitor, No. 3 (2006), UNCTAD/WEB/ITE/IIA/2006/9.

bit s , dtt s , and fdi flows: an overview xxxvii

From the point of view of the capital-exporting country, the basic purpose of BITs is to protect investors from political risks and instability and, more generally, safeguard the investments made by its nationals in the territory of the other state.17 This is why, originally, they were concluded primarily between developed and developing countries, as the former were virtually the only sources of FDI, and the latter were seen as often having risky and volatile business environments. Some of the more recent BITs, especially those with the United States and Canada, go further than protecting investors’ rights and require the liberalization of certain aspects of the FDI regime of a host country, for example by including provisions of national and most-favored-nation treatment at the establishment phase of an investment18 or by prohibiting host country governments from imposing certain performance requirements on foreign investments.19 Peter Muchlinski (Chapter 2) provides an overview of the principal substantive and procedural provisions of BITs. The substantive rights typically include a guarantee of prompt, adequate, and effective compensation for expropriation, freedom from unreasonable or discriminatory measures, a promise of “fair and equitable treatment” for foreign investments, guaranteed national and most-favored-nation treatment for investments, and assured full protection and security of investments. Together, these provisions are meant to boost investor confidence and the transparency of the policy environment. As mentioned, a number of more recently concluded BITs have expanded these rights somewhat to cover a wider range of host country activities in detailed and complex ways. A number of these provisions limit the regulatory flexibility of host countries to pursue not only economic development policies but other public policies as well. However, recent BITs also place somewhat greater emphasis on certain public concerns, including health, the environment, national security, labor rights, and transparency in information exchange and rulemaking. While BITs are largely similar in their substantive content and structure, recent innovations in their provisions have led to greater variation. In particular, three broad approaches seem to be emerging: the liberalization approach, used mostly by the United States, Canada, Japan and the Republic of Korea (and some other Western Hemisphere countries); the protection approach, mostly followed by European countries; and the more qualified protection approach, used mostly between developing countries. One notable difference is that the liberalization approach extends national treatment and most-favored-nation obligations to the

17. It should be noted that investments can also be protected through specific agreements between host country governments and foreign investors, or state contracts. These can be found especially in the natural resource sector. 18. Most of these treaties have, however, reservations that restrict the applicability of such clauses. 19. Tom Ginsburg, “International substitutes for domestic institutions: bilateral investment treaties and governance,” 25 Int’l Rev. of L. & Econ. 107, at 108 (2005).

xxxviii lisa e. sachs and karl p. sauvant

pre-establishment phrase of investment, while the two other approaches traditionally cover only the post-establishment phase. Additionally, the recent U.S. and Canadian model BITs clarify the meaning of, inter alia, the provisions on the minimum standard of treatment and regulatory takings, following lessons learned in recent NAFTA litigation, whereas the European BITs have not yet adopted these clarifications. BITs between developing countries are quite similar to the European BITs, but they often put more emphasis on exceptions and include clauses requiring the contracting parties to choose between litigation in the host country or in an international tribunal in case of a dispute.20 The procedural rights, one of the novel and noteworthy features of modern BITs, afford investors an adjudicatory mechanism to enforce substantive rights.21 Typically, investors can choose between arbitral panels at the World Bank’s International Centre for Settlement of Investment Disputes (ICSID), arbitration at another designated forum22 or ad hoc arbitration proceedings (especially UNCITRAL). This dispute settlement provision provides investors a remedy for unlawful or uncompensated actions by host states that affect their investments, usually without having to exhaust local remedies first before resorting to international arbitration. The designation of a third-party arbitration process frees investors from reliance on the political and judicial processes of host countries (which they often consider—rightly or wrongly—as being insufficient), and gives them direct access to protection under international law. If the proceeding is conducted under the ICSID Convention, the arbitration process is beneficial for the respondent state because it eliminates the possibility of diplomatic protection by the investor’s home country. While the ICSID Convention (Article 36.1) provides that both host country governments of contracting states and investors of contracting states can initiate investment-dispute settlement proceedings, BITs limit such initiation to investors.23 There were only a handful of internationally arbitrated investor-state disputes in the 1980s and early 1990s; however, by the end of 2007, 290 known international treaty-based arbitration cases had been initiated, involving at least 20. UNCTAD, “South-south cooperation in international investment arrangements,” UNCTAD/ITE/IIT/2005/3 (New York and Geneva: United Nations, 2005). Available at: http://www.unctad.org/en/docs/iteiit20053ch2_en.pdf 21. Most BITs have arbitration provisions, particularly the most recent ones; however, some do not, and some allow for very limited investor-state arbitration. 22. Other institutions available for arbitration include the ICC Court of Arbitration in Paris, the Stockholm Chamber of Commerce Arbitration, the London Court of International Arbitration, and various regional arbitration centers. 23. All treaty-based investment arbitration requires the consent of both parties. Host countries consent to treaty-based investment disputes in the dispute settlement clause of the BIT; since investors are not parties to BITs, BITs cannot constitute the investors’ consent. An investor’s consent is embodied in its claim, so no consent exists until the investor has filed a claim, at which point most BITs allow the host-country to file a counter-claim.

bit s , dtt s, and fdi flows: an overview xxxix

50

250 200

40 30

150 100

20 10

50 0

ICSID

Non-ICSID

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

0

Cumulative number of cases

350 300

60

1987

Annual number of cases

figure 9. known investment treaty arbitrations, cumulative and new cases, 1987 to end 2007

All cases cumulative

Source: UNCTAD, IIA Monitor, No. 1 (2008).

seventy-three countries—fifteen developed countries, forty-four developing countries, and fourteen economies in transition.24 Over three-quarters of these cases had arisen since the beginning of 2002, and close to two-thirds of them were filed with ICSID (or the ICSID Additional Facility) (Figure 9).25 Argentina has by far faced the highest share of claims filed at ICSID, most of which stem from its government’s emergency measures during the 2001 financial crisis. At least forty-six cases had been brought against Argentina for violating investment treaty protections, with twenty arbitrations brought in 2003 alone. Almost all treaty-based investment disputes are disputes brought by investors against host countries; the only known exception is a 2003 dispute between Chile and Peru, brought by Peru against Chile after a Chilean firm filed an investor-state claim against Peru.26 These proceedings have alleged treaty violations in response to a 24. UNCTAD, IIA Monitor, No. 1 (2008). There are also higher estimates. See the “Investment Treaty News: Year in Review 2006,” http://www.iisd.org/pdf/2007/itn_ year_review_2006.pdf. For a discussion of the reasons for this rise in investment disputes, see Jeswald W Salacuse, “Explanations for the increased recourse to treaty-based investment dispute settlement: resolving the struggle of life against form?”, in Karl P. Sauvant with Michael Chiswick-Patterson (eds.), Appeals Mechanism in International Investment Disputes (New York: Oxford University Press, 2008). 25. The total number of investor-state arbitration proceedings is not known because ICSID is the only institution that publicly provides a list of cases. UNCTAD, “Investorstate disputes and policy implications,” TD/B/COM.2/62 (Jan. 17, 2005), available at http://www.unctad.org/en/docs/c2d62_en.pdf. It appears that, in 2006, there were more non-ICSID cases than ICSID cases. See the “Investment Treaty News: Year in Review 2006,” http://www.iisd.org/pdf/2007/itn_year_review_2006.pdf. 26. Lucchetti S.A. and Lucchetti Peru S.A. v. Republic of Peru, ICSID Case No. ARB/03/4. There were a few other ICSID cases brought by states against investors; however in those cases, the basis of jurisdiction was not a treaty but a contract between the

xl lisa e. sachs and karl p. sauvant

range of state measures, including, for instance, emergency laws enacted during a financial crisis and the re-zoning of land for specific uses, and in a broad range of sectors such as construction, water and sewage services, telecommunications, financial services, mining, gas and oil production.27 Because of the confidentiality of proceedings, it is difficult empirically to evaluate data about parties to investment arbitrations, judgments and awards (if any).28 However, while imprecise and potentially subject to statistical biases, some trends can be inferred from data relating to known arbitrations. A study of 102 known investment treaty arbitration awards, deriving from eighty-two individual cases, showed that the tribunals resolved the treaty claims, and assessed whether damages were to be awarded, in fifty-two cases. In the majority of these (31 cases), the investor claimants were awarded nothing, while in twenty-one cases, the country respondent was ordered to pay damages to the investors.29 Furthermore, in the cases in which the investors did win, the awards were generally not large: only eight of the twenty-one cases that resulted in awards for the investors awarded them more than $5 million.30 There were four cases in which the award exceeded $10 million; the largest was a 2003 award to a Dutch-based firm in a dispute with the Czech Republic, in which the latter was ordered to pay some $270 million plus substantial interest for violating the terms of an investment treaty with The Netherlands.31 In 2006 and 2007, however, several awards in excess of $10 million indicate that the size of the awards may be increasing.32

investor and the state concerned: Tanz. Elec. Supply Co. Ltd. v. Indep. Power Tanz. Ltd., ICSID Case No. ARB/98/8 and Gabon v. Société Serte, ICSID Case No. ARB/79/1, Government of the Province of East Kalimantan v. PT Kaltim Prima Coal and others (Case No. ARB/07/3). 27. UNCTAD, “Investor-state disputes and policy implications,” TD/B/COM.2/62 (Jan. 17, 2005), available at http://www.unctad.org/en/docs/c2d62_en.pdf. A recent empirical study found that energy is the most heavily arbitrated sector, followed by the financial sector, food and beverage sector, transportation, and real estate transactions. Susan Franck, “Empirically evaluating claims about investment treaty arbitration,” 86 NC L. Rev. 1, at 38 (2007). 28. However, numerous awards and other decisions are available through a variety of sources. 29. Franck, supra note 27, at 52. 30. The investor claims are often very high. So in cases in which there have been awards, the awards have often been substantially lower than the initial claims. Franck, supra note 27, at 54. 31. CME Czech Republic B.V. v. Czech Republic, UNCITRAL, Final Award (Mar. 14, 2003), at http://ita.law.uvic.ca/documents/CME-2003-Final_001.pdf (awarding CME $269,814,000 in damages for breach of an investment treaty). 32. Note, however, that this trend has been dominated by the recent cases involving Argentina. It is also important to note that, in an unknown number of cases, investor claims are settled before the arbitration process is set in motion; the data on known awards do not account for such settlements between parties. In some cases, investors may be

bit s , dtt s, and fdi flows: an overview xli

It is unclear whether the rise in arbitrations will have an effect on the location of FDI or on countries’ perceived value of BITs—and, if there is an effect, what it will be. It is possible that this development may increase investors’ awareness of the protections and procedural rights that BITs afford investors and hence may cause more investors to consider the existence of investment treaties and the availability of arbitration when deciding on locations for FDI.33 On the other hand, the risk of investor success in these disputes, the financial burden of arbitration proceedings, and the sovereignty costs34 associated with the implementation of these treaties may lead some countries to conclude that the costs of BITs (and IIAs in general) outweigh their benefits. Ecuador’s announcement in February 2008 that it plans to withdraw from nine of its bilateral investment treaties and Bolivia’s withdrawal from the ICSID Convention in May 2007 may reflect a growing skepticism toward investment treaties and the international arbitration of investment disputes.35 As countries and investors adjust to the new provisions and scope of recent IIAs, as new patterns emerge in investment arbitration as regards, for example, the risk of being sued or the size of awards, and

using the threat of procedural rights of BITs to compel host country compensation or the recall of certain host country measures that adversely affected an investment. Furthermore, the possible “regulatory chill” resulting from the risk of treaty arbitration is impossible to measure. Therefore, the impact of BITs on the balance of power between investors and host countries is not measurable only from the awards effectively pronounced. 33. See Susan Franck, “Foreign direct investment, international treaty arbitration, and the rule of law,” 19 Pac. McGeorge Global Bus. & Dev. L.J. 337 (2007). 34. See especially Santiago Montt, “The BIT generation’s emergence as a collective action problem: prisoner’s dilemma or network effects?” Latin American and Caribbean Law and Economics Association Annual Papers, paper 043007’3, University of California, Berkeley, 2007. It should be noted in this context that all BIT guarantees can also be implemented unilaterally through domestic legislation, though domestic legislation cannot “internationalize” the commitment to certain forms of guaranteed investor treatment, which is a particularly valuable aspect of BITs in certain countries from the perspective of investors. 35. Moreover, on December 4, 2007, Ecuador notified ICSID that it would not consent to ICSID arbitration of disputes pertaining to investments in natural resources, such as oil, gas, and minerals. Pursuant to Article 25(4) of the ICSID Convention, a Contracting State may notify the Centre of the class or classes of disputes which the State would or would not consider submitting to the jurisdiction of the Centre. Such notification may be made at the time of ratification, acceptance or approval of the Convention or at any time thereafter. “Ecuador’s Notification under Article 25(4) of the ICSID Convention,” ICSID news release, December 5, 2007. There were also reports as of April 2008 that the Committee on Territorial Sovereignty of the Constituent Assembly of Ecuador has recommended to the floor a provision forbidding the government from submitting controversies to international arbitration. “Mesa de soberanía elimina el arbitraje internacional para el país,” Diario Hoy (April 15, 2008). Venezuela and Nicaragua have also suggested that they may withdraw from the ICSID Convention.

xlii lisa e. sachs and karl p. sauvant

as the size of the international treaty network grows, the effect of these treaties on FDI flows may evolve as well. Although both developing and developed countries have been concluding BITs at a rapid pace over the past few decades—as capital exporters seek to benefit from investor protections and capital importers hope to benefit through increased FDI flows—several scholars have been critical about BITs and their impact on developing countries. There is first of all the question whether, from a host country point of view, these treaties achieve their most basic objective— namely, to attract more FDI—precisely the focus of this volume. But there are broader considerations as well. Andrew Guzman (Chapter 3) argued that, whereas developing countries might have been better off negotiating a multilateral investment agreement as a group, individual developing countries defected in a prisoners’ dilemma situation as each tried to attract the largest possible share of foreign investment by concluding bilateral treaties with developed countries.36 He suggested that, by not taking collective action with other developing countries, countries have ratcheted up the investor protections that each country has committed to in bilateral agreements rather than working toward a multilateral agreement that could have made them all better off.37 Others are critical of BITs because they find that, by agreeing to international enforcement mechanisms and institutions, developing countries have neglected domestic legal institutions and mechanisms, such that “under some circumstances BITs may lead to lower institutional quality in subsequent years.”38 Finally, a number of authors have cautioned against the increased restrictions on developing countries’ regulatory flexibility imposed by BITs, even if new policies would be consistent with the country’s development objectives, or the furtherance of human rights goals, or would be necessary or desirable in response to a specific situation at a given time.39 Whether countries are responding to economic

36. Some of this reasoning seems to assume that, at any given time, there is a fixed pool of FDI for which countries compete. It is unclear to what extent this is accurate; the rise of FDI flows during the past decade indicates that this is not the case. 37. One could add, however, that, in the absence of BITs, host countries may be tempted to provide even more specific guarantees to foreign investors (especially efficiency-seeking foreign investors) which could lead to more “beggar-thy-neighbor” policies. Therefore, while BITs may leave developing countries worse off than a multilateral investment agreement would, they may do so less than to let investors “host-country shop” for special privileges. 38. Ginsburg, op. cit., p. 122. 39. See e.g., Luke Eric Peterson and Kevin R. Gray, “International human rights in bilateral investment treaties and in investment treaty arbitration,” International Institute for Sustainable Development (2003), p. 5. Peterson argued, for example, that “host states may wish to regulate the economy, including foreign investors embedded therein, in a manner which seeks to promote or protect certain human rights interests. . . . Where bilateral investment treaties are in place, foreign investors will often enjoy the ability to challenge

bit s , dtt s , and fdi flows: an overview xliii

crises (such as in Argentina following the 2001 financial crisis) or trying to protect nascent local industries, some BITs prohibit the country from imposing measures that could adversely affect foreign investors.40 While BITs are by far the most common agreements on foreign investment, beginning in the 1990s, investment issues have also been addressed increasingly and in a substantial manner in bilateral and regional free trade agreements, which, in the process, have become free trade and investment agreements (FTIAs).41 Typical investment provisions of FTIAs include most-favored-nation and national treatment. In addition to provisions that specifically address investment protection, they often also include liberalization clauses. In that manner, they might directly influence FDI flows by liberalizing investment conditions (by, most notably, opening sectors for investment). They might also indirectly influence FDI flows by enlarging the market, changing trade flows, setting and harmonizing standards, and improving a host country’s economic growth potential and overall investment climate. While the overall number of FTIAs and other treaties with investment provisions is still small (fewer than 10% of the number of BITs) their number is growing rapidly and, to the extent that they are regional, they cover more countries than there are agreements. As of June 2007, 251 FTIAs had been signed—nearly twice as many as five years earlier (Figure 10).42 These agreements are increasingly signed among developing countries; at the end of 2006, there were more than ninety IIAs other than BITs and DTTs concluded among developing countries.43 In contrast, the rise in the number of new BITs has slowed down in recent years, partly because most countries have already concluded BITs with their most important investment partners, partly because investment protection

these human-rights inspired measures through international arbitration.” See also, Ursula Kriebaum, “Privatizing human rights—the interface between international investment protection and human rights,” Transnational Dispute Management (2006), p. 14: “[An] investor may use BIT provisions to challenge human rights-inspired regulations that interfere with its investment.” 40. J.F. Perez-Lopez and M.F. Travieso-Diaz, “The contribution of BITs to Cuba’s foreign investment program,” 32 Law & Pol’y Int’l Bus. 529 (2001). 41. This represents a return to earlier patterns (e.g. the U.S. Friendship, Commerce and Navigation treaties) in which treaties addressed investment issues in the context of a range of issues. This approach allows trade-offs across issue areas (e.g. market access for increased investor protection). FTIAs also increasingly address host country responsibilities with respect to labor rights and environmental protection, topics ordinarily not addressed in BITs and never in DTTs. 42. UNCTAD Research Note, “Recent developments in international investment agreements,” UNCTAD/WEB/ITE/IIT/IIA/2007/6, available at: http://www.unctad. org/en/docs/webiteiia20076_en.pdf 43. UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. 17.

xliv lisa e. sachs and karl p. sauvant figure 10. number of other agreements∗ concluded, by period, 1957 june 2007 300 250 200 150 100 50 0 1957–1967 1968–1978 1979–1989 1990–2000 2001–2007 By period

Cumulative

Source: www.unctad.org/iia. ∗ International agreements, other than BITs and DTTs, that contain investment provisions.

is increasingly included in free trade agreements and other IIAs, and partly, perhaps, because arbitral awards in favor of investors may make some countries more wary about concluding additional BITs. Finally, there are also a number of inter-regional and multilateral investment agreements, most prominent among them the Energy Charter Treaty,44 the General Agreement on Trade in Services, and the Trade-related Investment Measures Agreement. While they are meant to establish a more welcoming enabling framework for FDI, their effect on FDI flows is difficult to ascertain. 2. Double taxation treaties The surge of FDI flows has also been accompanied by a surge of double taxation treaties. They serve a different but complementary purpose to BITs. While the primary purpose of BITs is to protect foreign investments, one of the main purposes of DTTs is to deal with issues arising out of the allocation of the revenues generated by these investments between host and home countries—for instance, how to allocate tax revenue from taxes imposed on income earned by multiple entities of a MNE system. The preferential tax and related arrangements in DTTs reduce the administrative complexity of foreign investments and facilitate the flow of goods and services between the treaty partners. As is the case with BITs, the propensity of countries to enter into

44. See e.g., Wälde, Thomas W. (ed.). The Energy Charter Treaty: An East-West Gateway for Investment and Trade (London, The Hague, Boston: Kluwer Law International Ltd., 1996).

bit s , dtt s , and fdi flows: an overview xlv

figure 11. dtt s concluded as of end 2006, by country group (percent) 16 3 12

38

24

7 Between developing countries Between developed and developing countries Between developing and South-East Europe and Commonwealth of Independent States Between developed countries Between developed countries and South-East Europe and Commonwealth of Independent States Between countries of South-East Europe and Commonwealth of Independent States

Source: UNCTAD, World Investment Report 2007, op. cit., p. 18.

DTTs reflects the growing role of FDI in the world economy and countries’ efforts to attract MNEs to their own territories and facilitate their operations. The steady rise of DTTs (Figure 6) began nearly four decades ago (while the sharp rise in BITs began in the early 1990s). By the end of 2006, 2,651 DTTs had been signed.45 Unlike BITs, which were initially concluded largely between developing and developed countries, DTTs were initially signed primarily between developed countries, as these were traditionally the main capital exporting and capital importing countries and thus faced the greatest double taxation challenges. Starting in the late 1960s, as emerging markets increasingly became host countries for FDI, and gradually also became home countries, the number of those countries that signed DTTs with both developed countries and other developing countries also began to rise rapidly (Figure 11). Developed countries still top the list of the countries with the largest number of DTTs, led by the United States, the United Kingdom and France (Figure 12). Most double taxation treaties are based on one of two international models, which in turn are based on models developed by the League of Nations. 45. Zhan, Karl and Weber, op. cit.

xlvi lisa e. sachs and karl p. sauvant figure 12. the ten countries with the highest number of dtt s , end 2006 United States

148

United Kingdom

145

France

129

Netherlands

121

Switzerland

110

Canada

108

Sweden

106

Denmark

104

Norway

100

Germany

98 0

20

40

60

80

100

120

140

160

Source: www.unctad.org/iia.

Developed countries coordinate their efforts on international tax matters in the Fiscal Committee of the OECD. In 1963, these countries issued a draft model DTT,46 which is updated periodically and has become the model for most DTTs concluded since its conception. In addition, a Committee of Experts on International Cooperation in Tax Matters, convened by the Secretary-General of the United Nations, issued a “Manual for the Negotiation of Bilateral Tax Treaties between Developed and Developing Countries,” as well as the “United Nations Model Double Taxation Convention between Developed and Developing Countries” in 1979, which are also updated periodically.47 The UN model was drafted specifically because of concerns that the OECD model was not appropriate for tax treaties between developed and developing countries, which involve non-reciprocal cross border activity. Nevertheless, the OECD and UN models share many common features. Most DTTs are based on one of these two models, with exceptions and variations depending on the specific relationship between the two contracting states. One other model DTT is that of the United States, first published

46. The “OECD Model Convention on Income and on Capital” (MTC) is available at: http://www.oecd.org/document/17/0,3343,en_2649_33747_35035793_1_1_1_1,00.html The electronic version of the MTC is based on the text as it was updated in January, 2003, but includes the 1963 and 1977 texts. 47. The “Manual for the Negotiation of Bilateral Tax Treaties between Developed and Developing Countries” is available at: http://unpan1.un.org/intradoc/groups/public/documents/un/unpan008579.pdf; the “United Nations Model Double Taxation Convention between Developed and Developing Countries” is available at: http://unpan1.un.org/ intradoc/groups/public/documents/UN/UNPAN002084.pdf.

bit s , dtt s , and fdi flows: an overview xlvii

in 1977, which incorporates much of the OECD model but also includes provisions to reflect the specific policies of the United States.48 Reuven Avi-Yonah (Chapter 4) provides an overview of the reasons for the rise of DTTs and their salient features. The need to address the issue of revenue allocation between host and home countries arose as increased international investment created a potential conflict of tax jurisdictions, that is, two or more jurisdictions had the right to levy tax on a single event or a single taxpayer (e.g., a company operating in several countries). Double taxation can also occur in other situations, for instance if jurisdictions have different tax definitions, residency requirements or income classifications. When the national tax laws of the two countries differ significantly, the jurisdictional conflict can also lead to improper conduct by taxpayers. Jurisdictional conflicts can be relieved unilaterally (under national tax laws) or at times multilaterally49; on the other hand, DTTs provide the most important and most common international measures to relieve double taxation problems. In order to help eliminate double taxation and to relieve jurisdictional conflicts, double taxation treaties standardize tax definitions in the countries party to a treaty, and they detail specific allocation rules for different categories of income, reducing uncertainty about the tax environment in both countries. DTTs also can limit transfer pricing, help to combat tax evasion (notably through the exchange of information), reduce the risk of treaty shopping50, provide non-discrimination rules, and outline ways in which tax disputes can be resolved by prescribing specific conflict resolution mechanisms and arbitration procedures.51 Furthermore, while unilateral measures can often eliminate double taxation on their own, thus obviating the need for DTTs to prevent double taxation, the treaties can still be useful in “borderline” situations, such as cases in which the source of income is disputed.52 Importantly, DTTs provide greater legal certainty to foreign investors with respect to the tax treatment of their cross-border

48. “U.S. Model Income Tax Convention of Sept. 20, 1996,” available at: http://www. treas.gov/offices/tax-policy/library/model1996.pdf. 49. The various attempts at multilateral agreements thus far have had little success. Those that have been somewhat successful have been supplemented by bilateral treaties among the parties to the multilateral agreement. Karl P. Sauvant and Jorg Weber, eds., International Investment Agreements: Key Issues (New York and Geneva: United Nations, 2004), volume II, p. 204. Available at: http://www.unctad.org/en/docs/iteiit200410v2_ en.pdf. 50. Treaty shopping is the routing of an investment and the associated income through a particular country in order to take advantage of treaty benefits intended for the residents of that country and its treaty partners. 51. Sauvant and Weber, op. cit., p. 203. 52. Tsilly Dagan, “The tax treaties myth,” 32 Journal of International Law and Politics 939, at 979 (2000).

xlviii lisa e. sachs and karl p. sauvant

activities in both the host and the home country.53 In the absence of DTTs, there could be more source taxation and less residence taxation, and there would be more administrative complexity, especially where investment flows are reciprocal. While the stated purpose of DTTs is to address these tax issues—the very limited preambles of most DTTs do not mention FDI—most countries (including the United States), as well as much literature and commentaries, have claimed that, by eliminating excessive taxation, tax treaties can help increase trade and investment between the two treaty signatories.54 Capital-exporting countries, foreign investors and capital-importing countries stand to gain from DTTs. For capital-exporting countries, tax treaties facilitate the foreign expansion of their own companies by relieving potential double taxation problems in foreign territories without risking improper tax evasion or fraud. Foreign investors benefit from taxation treaties because, even in cases in which there is no double taxation, tax treaties generally include greater and more comprehensive tax protections for investors than are available under the domestic tax rules of either host or home countries, which moreover can change at any time.55 Furthermore, DTTs determine the maximum rates of taxation (particularly the withholding tax rates) that can be imposed by a host country, and a number of countries give preferential rates to firms in countries with whom they have DTTs. Capital-importing countries benefit because the extra tax protection afforded to foreign investors can be an added incentive for foreign investment in their territories, if other locational determinants for FDI are satisfactory. Furthermore, when DTTs contain tax-sparing provisions (whereby residence countries would grant double tax relief for the tax that would have been due in the host country were it not for a tax incentive offered to the investor), these can make some of the incentives offered by host countries more effective. And many DTTs include an exchange-of-information provision that allows the developing country to obtain information exchanged from capital-exporting countries, which can help developing countries tax capital invested by their rich residents overseas. While DTTs may help developing countries attract more foreign investment, they also, prima facie, reduce the tax revenue of these countries. DTTs typically reduce source-based taxation (of the host contry), thereby shifting tax revenues from the source country to the residence country, and most developing countries are net capital importers. As one author notes, though “the contraction of taxing jurisdiction is technically reciprocal in the treaty document, the one-sided flow of capital toward LDC [less-developed country] as source-country ensures that only that country experiences a true contraction of its taxing jurisdiction.”56 53. Sauvant and Weber, op. cit., p. 204. 54. Allison Christians, “Tax treaties for investment and aid to Sub-Saharan Africa,” 71 Brook. L. Rev. 639 (2005), p. 658. 55. Sauvant and Weber, op. cit., p. 204. 56. Christians, op. cit., p. 658.

bit s , dtt s , and fdi flows: an overview xlix

The reduced tax revenues of developing countries can only pay off if DTTs do in fact lead to higher economic growth, for which more FDI flows may be integral.57 Solely from a FDI attraction perspective, it appears that developing countries may therefore need to decide whether it is better for them to preserve their tax jurisdiction over foreign investors in order to maximize their tax revenue, or to agree to relieve source-country taxation in order hopefully to attract more FDI. In a recent article, Tsilly Dagan illustrated the conundrum and presented a game theory rationale that explains why many developing countries have opted for the latter.58 C. The determinants of FDI Flows The issue examined in this volume is to what extent the expansion of the BIT and DTT networks has directly led to the rapid growth of FDI flows that occurred during the past decade or so. The principal difficulty for assessing the impact is that the existence of an investment or double taxation treaty is but one determinant that may affect decisions to invest abroad. Three sets of FDI determinants can be distinguished: the regulatory framework; investment promotion; and economic factors. THE REGULATORY FRAMEWORK . The number of BITs and DTTs concluded over the past couple of decades, and their pro-investor content, is reflective of the general movement of countries toward efforts to attract FDI by liberalizing their FDI regimes and creating national regulatory frameworks that are favorable for foreign investors. The regulatory framework of a host country is a key determinant for the location of foreign investment in so far as foreign investors simply cannot enter into, and operate in, a country if national laws prohibit or impede foreign investment. For over two decades, virtually all countries have been improving their investment climate by adopting national laws and regulations, including those that open more sectors to foreign investment, that facilitate inward FDI. Between 1991 and 2006, out of 2,533 national legal and regulatory changes relevant to foreign investment, 91% were in the direction of making the host country environment more favorable for FDI (Table 3). However, one should also note that, while the share of regulatory changes that are favorable to FDI remains high, the number of favorable changes has decreased significantly since 2004; in fact, the number of regulatory changes that are less favorable to FDI tripled between 2002 and 2006, perhaps signaling an increased skepticism in some countries of the benefits of FDI and a new tendency toward FDI protectionism. International investment agreements are part of the regulatory framework of a host country and can affect aspects of it directly. Most importantly, BITs establish certain standards of treatment that become parameters for national 57. One should note in this context that, while FDI inflows contribute to economic growth, economic growth alone does not necessarily guarantee human development, the ultimate goal of countries. 58. Dagan, op. cit.

Year Number of countries that introduced changes Number of regulatory changes More favorable to FDI Less favorable to FDI

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Total 35

43

56

49

63

66

76

60

65

70

71

72

82

103

93

93

–

82 80 2

77 77 0

100 99 1

110 108 2

112 106 6

114 98 16

150 134 16

145 136 9

139 130 9

150 147 3

207 193 14

246 234 12

242 218 24

270 234 36

205 164 41

184 147 37

2,533 2,305 228

Source: UNCTAD, World Investment Report 1996: Investment, Trade and International Policy Arrangements (New York and Geneva: UN, 1996), p. 132; UNCTAD, World Investment Report 2007. p. 14.

l lisa e. sachs and karl p. sauvant

table 3. national regulatory changes, 1991–2006

bit s , dtt s , and fdi flows: an overview li

regulations in the investment area, and DTTs establish or clarify tax treatment for foreign investors. Investment agreements may also provide for the opening of certain sectors of the host country economy. Finally, if a country has concluded a BIT with the home country of an investor that grants investors from that home country certain enforceable rights, the investor typically has access to international arbitration.59 INVESTMENT PROMOTION. In addition to liberalizing national policies concerning foreign investment, many countries facilitate foreign investment pro-actively. For this purpose, virtually every country has established an investment promotion agency to attract FDI and ease its operation. Among other things, these agencies offer incentives and various investment services, upgrade amenities for foreign investors, issue investors’ guides, generally seek to improve the ease of doing business in its country, and may engage in policy advocacy. ECONOMIC FACTORS . While a country’s regulatory framework must be enabling, and investment promotions can help, the locational decisions of MNEs ultimately depend on economic factors in prospective host countries. The economic determinants of FDI can be divided into three categories. Locational resources and assets include the quality of labor; the nature of the physical and institutional infrastructure; the availability of natural resources; and the technology system and the domestic enterprise base. Market variables include economic growth and per capita income, the size of the host country market for goods and services, access to regional and global markets, country-specific consumer preferences, and the structure of the host country market. Efficiency considerations include the cost of resources and other inputs such as transport and communication costs and membership in a regional integration agreement. It is on the basis of these economic factors that the business case for an investment is made. Moreover, these economic variables also have to be seen in the broader framework of a number of macro factors, such as the performance of the world economy, and a number of social and political factors, including the policy approach to private business in general. It is clear, then, that no individual factor, such as an investment treaty, could move FDI flows by itself, and it is equally clear that it is very difficult to isolate the importance of any particular factor. To put it differently: if BITs and DTTs affect

59. Some investment insurance programs require (technical) investment agreements between home and host countries in order for investors to obtain political risk insurance; such insurance programs normally require that the host country consents to arbitration in case of investor disputes. World Bank, World Development Report 2005: A Better Investment Climate for Everyone (Washington: World Bank, 2005), at 177. Available at: http://siteresources. worldbank.org/INTWDR2005/Resources/complete_report.pdf; see also Jason Yackee, “Conceptual difficulties in the empirical study of bilateral investment treaties,” University of Wisconsin Law School Legal Studies Research Paper Series, Paper No. 1053 (October 2007), p. 48. Available at: http://ssrn.com/abstract=1015088

lii lisa e. sachs and karl p. sauvant

FDI flows, they do so in the context of a host of other determinants, with a number of them considerably more important than individual aspects of the regulatory framework. In general, the regulatory framework of a host country is at best enabling; once it is permissive, the economic determinants become key, especially market size and growth, skills, resources, and costs. While the economic determinants are not everything, everything is nothing when it comes to attracting FDI. In this context, then, BITs and DTTs can help improve the regulatory framework by complementing host-country policies related to FDI, guaranteeing certain investor rights, making the legal and tax frameworks more transparent and stable for investors, and mitigating the potential impacts of political or economic instability by establishing certain enforcement procedures. If BITs and DTTs help improve the regulatory determinants for FDI, they then allow the key economic determinants—if present—to prevail. Despite difficulties with respect to identifying the specific impact of treaties on investment flows (given the wide range of other variables that need to be considered), as well as cause and effect relationships between the existence of BITs and DTTs and FDI flows, a number of scholars have attempted to assess the impact of such agreements on FDI flows. The following sections highlight the findings of the principal studies undertaken during the ten-year period from 1998 to 2007 and reproduced in this volume. D. BITs and FDI flows Given the principal purpose of BITs—to protect investment and hence encourage investment flows—it is only natural that the question has been raised whether they do, in fact, lead to higher investment flows. Jeswald Salacuse and Nicholas Sullivan (Chapter 5) and Tim Buthe and Helen Milner (Chapter 6) both determined that concluding BITs does have a positive effect on FDI inflows and that the effect is larger when developing countries conclude these agreements with economically more important countries. Analyzing the impact of BITs with OECD countries on aggregate FDI inflows to 100 developing countries, Salacuse and Sullivan found that, when developing countries concluded BITs with OECD countries, FDI inflows were likely to increase. Furthermore, they determined that a U.S. BIT was likely to have more of an impact than other OECD BITs in promoting overall FDI, and that a U.S. BIT was likely to promote U.S. FDI as well. Similarly, Eric Neumayer and Laura Spess (Chapter 7), looking at 119 developing countries between 1970 and 2001, found that developing countries that signed more BITs with developed countries that were major source countries of FDI received a higher share of FDI flowing to developing countries. Most authors agree that the strength of the impact of BITs on FDI inflows depends on several political, regulatory and economic factors, both within the host country and globally. For instance, Neumayer and Spess found that countries with faster-growing economies and larger populations receive more FDI. Moreover, they suggested that BITs may in fact function as substitutes for poor host country institutional quality. Precisely because political risk and volatility

bit s , dtt s , and fdi flows: an overview liii

are constraints on FDI inflows, Neumayer and Spess suggested that countries “with particularly poor domestic institutional quality possibly stand the most to gain from BITs,” and that the positive effect of BITs on FDI decreases as governments become more stable. In fact, it is possible that merely signing a BIT, before implementation actually occurs, has a positive signaling effect, as Peter Egger and Michael Pfaffermayr’s analysis of outward FDI stock from OECD countries (Chapter 8) suggested, though they did find that BITs that have entered into force have a stronger positive effect on outward FDI stock than those that have merely been signed. Buthe and Milner developed a theoretical argument that explains an increase in overall inward FDI flows as a function of the success of BITs as a political commitment by developing countries to economically liberal policies, which foreign direct investors generally seek, at least in developing countries. Investors consider these commitments to be more credible because BITs signal such commitments and governments’ compliance with them and, in addition, make breaking such commitments more costly. Several analysts have also found that the sheer number of BITs signed by a country influences FDI inflows. Neumayer and Spess found that countries with a higher cumulative number of BITs receive more FDI inflows. Robert Grosse and Len Trevino (Chapter 9) and Kevin Gallagher and Melissa Birch (Chapter 10) also found a strong positive relationship between the total numbers of BITs concluded by a country and FDI inflows to that country in studies that focused on Central and Eastern Europe and Latin America, respectively.60 Susan RoseAckerman (Chapter 11) took a global view, finding that, as the total worldwide coverage of BITs goes up, overall FDI flows to developing countries may increase, though the marginal benefit to any one country of signing BITs will decrease. Unlike Neumayer and Spess, Rose-Ackerman found that “the marginal impact of BITs is greater in countries that already have relatively effective legal regimes and favorable economic environments.” Not all empirical studies shed such a favorable light on BITs’ impact on FDI inflows. Several scholars determined that BIT protections cannot substitute or compensate for the economic and regulatory risks of a host country; to the extent that such treaties affect FDI flows, they do so as one of a host of other regulatory 60. In a 2008 study, Len Trevino, Douglas Thomas and John Cullen also find BITs to be “significant indicators of inward FDI in Latin America.” They argue that institutionalization is a process that can legitimize a host FDI market for foreign investors through three (cognitive, normative and regulative) pillars. They found that while BITs work through the regulative framework, the “signals BITs send through the cognitive and normative pillars are more dominant than those sent through the regulative pillar,” and that in general, “institutional processes that legitimize [a host country FDI market] through the cognitive and normative pillars . . . are better indicators of inward FDI than those that legitimize primarily through the regulative pillar [such as trade or tax reform].” Trevino, Len J., Thomas, Douglas E., and Cullen, J., “The three pillars of institutional theory and FDI in Latin America: An Institutionalization Process,” 17 International Business Review 118 (2008).

liv lisa e. sachs and karl p. sauvant

and economic determinants that impact FDI. A 1998 UNCTAD study (Chapter 12), one of the first to evaluate the impact of BITs on FDI flows, concluded that, on balance, BITs did not play a primary role in increasing FDI, and that a larger number of BITs ratified by a host country would not necessarily lead to higher FDI inflows. In another early study, Mary Hallward-Driemeier (Chapter 13) analyzed the bilateral flow of FDI from 20 OECD countries to 31 developing countries from 1980 to 2000 and noted that BITs had an insignificant effect on FDI flows. However, she also found that, rather than encouraging more FDI flows in riskier environments, BITs only have a positive effect on FDI flows in countries with an already stable business environment and reasonably strong domestic institutions. If a country signs a BIT while undertaking domestic regulatory reforms that facilitate FDI, it would be the institutional reforms and liberalization that may affect investors’ locational decisions rather than simply the BIT itself. Hallward-Dreimer’s results suggested that the size of a host country’s market is a more conclusive determinant of FDI flows than the conclusion of a BIT. Jason Yackee (Chapter 14) also found little evidence that BITs have any effect on FDI flows. 61 Emma Aisbett’s study (Chapter 15) demonstrated the importance of accounting for the endogeneity of BIT adoption when assessing the impact on FDI flows. She suggested that the relevance of BITs may vary by sector or that there may be reverse causality, where a higher growth rate of FDI leads to an increased probability of a BIT being negotiated. Deborah Swenson (Chapter 16) also found that countries were more likely to sign BITs if they already had high levels of FDI, suggesting that “the interest of existing foreign investors drove the signing of BITs, at least in part”; but she maintained that signing these BITs may have helped these countries retain existing levels of FDI. She also emphasized the importance of controlling for timing, intrinsic country attractiveness and investor identity in analyses of BIT effectiveness. She found that, controlling for these variables, data from the late 1990s suggest that BIT signing did help developing countries attract more FDI. Taken together, these analyses suggest that it is difficult to establish firmly the effect of BITs on FDI flows. Intuitively, one would expect that such treaties, by providing a sort of good housekeeping seal of approval, have a positive effect

61. In a recent paper, Peter Buckley et al. tested the effect of various supranational institutional factors, including BITs, on the decision-making of Chinese MNEs from 1991 to 2003, and found no significant relationship between Chinese outward FDI patterns and the conclusion of a BIT with China; they also did not find a signaling effect of the total number of BITs concluded by a country. (It should be noted in this context that the protection offered by Chinese BITs is not as strong as that of treaties of other countries.) See Peter Buckley et al., “Explaining China’s outward FDI: an institutional perspective,” in Karl P. Sauvant, with Kristin Mendoza and Irmak Ince (eds.), The Rise of Transnational Corporations from Emerging Markets: Threat or Opportunity? (Cheltenham: Edward Elgar, 2008).

bit s , dtt s , and fdi flows: an overview lv

on FDI flows as they signal that a country is interested in attracting such investment and that it provides certain guarantees under international law to protect it (thereby reducing the risk premium of an investment); and this signal is not only sent to a particular treaty partner but to the international investment community as a whole. The incidence of treaty shopping—whereby a firm invests in another country not from its home country but via a country that has a BIT with the prospective host country—also suggests that at least some firms deliberately seek the protection of a treaty. The rise in international arbitral cases shows, furthermore, that investors pursue their rights if they feel aggrieved. Moreover, if BITs not only protect investments but also liberalize entry and operations, one would expect a rise of inflows, assuming attractive investment opportunities. So why the different findings in the chapters included in this volume (methodological issues aside)? To begin with, most of the bilateral FDI stock and flow data are poor. Where they exist, moreover, the nature of FDI may play a role: the effect of BITs on investors’ locational decisions is likely weaker for natural resource and market-seeking investors for whom the economic determinants of FDI are clear, whereas such treaties might more likely influence the decisionmaking of efficiency-seeking investors for whom several investment locations may be otherwise equally attractive. But FDI data mostly do not allow one to distinguish clearly between these various types of FDI. Difficulties exist also in disentangling the causal effects from BITs on FDI flows from the causal effects of a simultaneous and autonomous liberalization of the national FDI regulatory framework—a trend, as shown, that is strong and pervasive. The level of development of the BITs partners—for instance whether BITs are signed between developed and developing countries or between developing countries, or whether the developing country is more or less developed—may also play a role. More generally, BITs may be relatively more influential in certain countries or contexts than in others, depending on the type of investments common to a country or the mix of other—more crucial—FDI determinants. The magnitude of the correlation between BITs and FDI, then, may vary for various countries and regions for reasons that are not captured or explored in the studies. Furthermore, the effect of BITs may change over time, for instance as the worldwide coverage of BITs continues to grow and as more or less all important countries conclude BITs with each other, the ability of these treaties to influence locational choices may even out. The diverse findings in the literature may also reflect variations in the provisions of BITs. For example, most regression analyses look at whether or not BITs were in place, without factoring in the varying degrees of investor protections and benefits in these treaties62, for example, as regards the breadth of arbitration

62. Franck, supra note 33; Deborah L. Swenson, “Why do developing countries sign BITs?,” 12 U.C. Davis J. Int’l L. & Pol’y 131, at 153 (2005).

lvi lisa e. sachs and karl p. sauvant

rights or the primacy of BIT rights over national law. Another variation that could account for disparities in the studies is that BITs that include liberalizing provisions in addition to investor protection provisions (especially BITs with the United States, Canada, and Japan) can influence FDI flows by opening sectors previously closed to foreign investment; assuming the economic determinants are right, it would not be surprising for “liberalizing” BITs to lead to more FDI. This could perhaps explain the different findings for countries that have concluded BITs with the United States as opposed to other OECD countries. The specific BIT effect can be further complicated if a BIT country enters, more or less simultaneously, bilateral or regional free trade and investment agreements: these latter agreements could have a similar “opening” effect for FDI and/or they could lead (via trade liberalization) to a larger market,63 with both effects potentially leading to an increase in FDI flows. Moreover, when the effect on FDI flows of BIT countries is compared with that of non-BIT countries, the comparison is complicated if the latter are covered by bilateral, regional or multilateral agreements with substantial investment provisions, blurring the distinction between these two groups of countries.64 Put simply, countries have multiple tools for protecting foreign investments and the interests of foreign investors in addition to BITs, so a more comprehensive study would need to account for alternative investment promotion and protection measures in addition to BITs.65 Crucial, however, is the importance of the economic factors—and BITs do not directly influence them. Unless, as pointed out earlier, they are favorable, FDI typically does not take place; and when they are favorable, and especially when they are strongly favorable, FDI can also take place in the absence of BITs. Since the economic factors trump virtually all other factors (assuming FDI is permitted), any study that seeks to isolate the specific BITs effect on FDI flows needs to include economic variables fully in its calculation. Considering the complex relationship between investment treaties and the various variables of the three sets of FDI determinants, it is not surprising that it is difficult to establish firmly the effect of BITs on FDI flows. It fits into this picture that, in a June 2007 survey of 602 senior executives of MNEs around the 63. For example, a World Bank study found that regional agreements that create larger markets positively affect FDI inflows when other institutional variables affecting the investment climate are satisfactory (though agreements that do not result in larger markets do not positively affect FDI flows). Richard Newfarmer, “Beyond merchandise trade: services, investment, intellectual property, and labor mobility,” in Global Economic Prospects 97, at 109 (2005), available at http://siteresources.worldbank.org/INTGEP2005/ Resources/gep2005.pdf 64. For example, many developing countries are covered by the GATS—and FDI in services to developing countries accounts for more than half of all FDI flows to these countries. 65. Swenson, op. cit., p. 153.

bit s , dtt s , and fdi flows: an overview lvii

figure 13. the influence of iia s on the locational decisions of corporate executives, 2007* (percent of respondents) To a very great extent 19

To a limited extent 48

Not at all 23

Don't know 9

∗ The question in the EIU survey was: “To what extent does the existence of an international investment agreement (for example, a bilateral investment treaty) influence your company’s decision on which markets to invest in?” Source: Matthew Shrinkman, “The investors’ view: economic opportunities versus political risks in 2007–11,” in Kekic and Sauvant, op. cit., p. 96.

world, roughly one-fifth of the recipients indicated that the existence of international investment agreements influenced their locational decisions “to a very great extent”66 while an equal share said that such agreements influenced their decisions “not at all.” At the same time, roughly half of the respondents indicated that IIAs influenced locational decisions “to a limited extent,” suggesting that other factors needed to be present (Figure 13). A World Bank report also noted that there is evidence that many investors may not be aware of existing BITs when they make locational decisions, and may in fact “remain oblivious until some issue arises when its provisions may be relevant.”67 E. DTTs and FDI Flows There are considerably fewer studies of the impact of DTTs on FDI flows than there are of the effect of BITs. On the one hand, this may not be surprising since one of the declared principal objectives of BITs is to promote FDI flows, and this objective invites a test whether it is indeed achieved; in addition, BITs are politically more sensitive because they directly influence the regulatory space of host

66. One could question, however, whether some of the senior executives who answered that IIAs influenced their locational decision making “to a great extent” may have strategically over-stated the importance of IIAs in their decision making in order to encourage the granting of such further protections IIAs may offer them. 67. World Bank (2005) World Development Report 2005, op. cit., p. 177.

lviii lisa e. sachs and karl p. sauvant

countries across a range of important policy areas. On the other hand, DTTs (highly technical treaties) should be important for the locational decision of firms as they can directly affect the “bottom line” of a company’s performance. Moreover, the incidence of DTT treaty shopping suggests that they influence the routing of investment flows. A well-known example is the fact that a good part of FDI into India is routed through Mauritius, as the latter has a more favorable DTT with India than those concluded between India and the original residence countries of investors.68 In any event, most of the difficulties that afflict the analysis of the impact of BITs on FDI flows also are relevant to the impact of DTTs on FDI flows and hence need not be rehearsed. Indeed, most empirical studies have generally found that DTTs entered into since the early 1980s have not had a demonstrable impact on investment flows. The analysis by Bruce Blonigen and Ronald Davies (Chapter 17) of the impact of bilateral tax treaties on FDI activity in OECD countries from 1982 to 1992 found that DTTs are associated with larger FDI stocks and flows. However, when older DTTs concluded many years before the period of their study are distinguished from newer DTTs entered into during their observed time period, they found that the newer treaties had no positive effect on FDI activity. In a subsequent study, Blonigen and Davies (Chapter 18) investigated U.S. FDI flows from 1980 to 1999 and found that DTTs concluded by the United States during this period had no significant effect on inward and outbound FDI.69 In fact, Blonigen and Davies concluded that recent DTTs have had negative effects on OECD outbound FDI flows, which they suggested could be a consequence of the elimination of tax avoidance opportunities in DTTs. Peter Egger et al. (Chapter 19) also found a negative effect of newly implemented DTTs on outward FDI stock from OECD countries when analyzing FDI data two years prior and two years after DTT conclusions from 1985 to 2000. A couple of the studies focus on the relative importance of tax treaties to the locational decision-making of MNEs. Henry Louie and Donald Rousslang (Chapter 20) investigated how both the quality of host-country governance and 68. In Union of India and Anr vs. Azadi Bachao Andolan and Anr (Oct. 7, 2003), the Supreme Court of India held that tax treaty shopping was valid under the Indo-Mauritius tax treaty. In part of its decision, the Supreme Court emphasized that tax treaty shopping could positively attract more FDI to India (“In recent years, India has been the beneficiary of significant funds through the ‘Mauritius conduit.’”) Eduardo Baistrocchi, “The structure of the asymmetric tax treaty network: theory and implications,” Bepress Legal Series, Working Paper 1991 (2007). 69. The aforementioned paper by Buckley et al. on factors affecting the decisionmaking of Chinese MNEs from 1991 to 2003 similarly found no significant relationship between Chinese outward FDI patterns and the conclusion of a DTT with China; they also did not find a signaling effect of the total number of DTTs concluded by a country. Buckley et al., “Explaining China’s outward FDI: an institutional perspective,” in Sauvant, with Mendoza and Ince (eds.), op. cit.

bit s , dtt s, and fdi flows: an overview lix

having a bilateral income tax treaty with the United States affect the rates of return that U.S. companies expect from their foreign investment. They found that poor governance causes U.S. companies to require significantly higher rates of return, thereby discouraging inward FDI, and that, after accounting for the quality of host-country governance, a tax treaty does not have an effect on the required rates of return. Similarly, Allison Christians (Chapter 21) used a case study of a hypothetical tax treaty between Ghana and the United States to demonstrate that a typical tax treaty with a low-income developing country does “not provide major tax benefits to the private sector [so] even if concluded, these treaties would not have a significant impact on cross-border investment and trade.” Some of the studies in this volume arrive at more guarded conclusions about the impact of DTTs on FDI flows. Daniel Millimet and Abdullah Kumas (Chapter 22) determined that assumptions in previous studies concerning the timing of the effect of tax treaties are important, finding that allowing for anticipatory and lagged effects of treaty formation indicates a more substantial, positive effect on FDI activity. Eric Neumayer (Chapter 23) found that developing countries that have signed more DTTs with major capital exporting developed countries are, in fact, likely to have received more FDI in return. However, his results showed that DTTs are only effective in middle-income developing countries, not in low-income developing countries, a qualification supported by Christians’ study. Apart from the difficulties already reviewed in the context of BITs, there are several possible specific explanations for why double taxation treaties may not lead to higher FDI flows. For instance, as with many provisions in BITs, several of the tax alleviation provisions in tax treaties that are expected to encourage FDI flows can also be implemented unilaterally through the domestic policies of host governments,70 so countries with DTTs may not provide a significantly different tax framework than countries without such treaties. Another possibility is that, while DTTs largely address the problem of double taxation, they also reduce opportunities for tax evasion by foreign investors, which may even act as a disincentive for FDI. Another rationale advanced by some of the literature is that DTTs limit the tax revenue of host countries, thereby reducing the governmental resources to construct the infrastructure necessary to attract and support FDI in the first place.71 More broadly, the existence and widespread use of tax havens, the possibility of allocating various charges across the affiliates of a corporate

70. For instance, most countries implement the foreign tax credit or exemption of foreign source income unilaterally. For a discussion of unilateral policies to relieve double taxation and how they compare to tax treaty provisions see Dagan, op. cit. 71. See, e.g., Buckley et al., “Explaining China’s outward FDI: an institutional perspective,” in Sauvant, with Mendoza and Ince (eds.), op. cit.

lx lisa e. sachs and karl p. sauvant

system and the use of transfer pricing may provide alternatives to DTTs for firms seeking to minimize their tax burden. F. BITs, DTTs and FDI Flows A potential omitted variable in all of the above studies, including the studies of the effect of BITs on FDI, is that each study focuses either on BITs or on DTTs but not on both, and none of the studies include a variable for the presence of the other type of treaty. Tom Coupé, Irina Orlova, and Alexandre Skiba’s study (Chapter 24) showed that the correlation between the two types of treaties is statistically significant. They estimated the effect of both BITs and DTTs on FDI flows from seventeen OECD countries into nine economies in transition, focusing on such economies to increase the homogeneity of their sample and because available data provide them with a good proxy for a wide range of home policies that may influence their results. Their results showed that transition economies that have BITs with developed countries receive more FDI inflows from those countries, but that tax treaties do not have a significant effect on FDI inflows. Their results also suggested that BITs are substitutes for institutional quality as the net effect of a BIT is smaller (but still positive) if the quality of a host country’s institutions is higher.

conclusions One uncontroversial truth is that virtually all countries value FDI as a means to advance their economic development. Therefore, not surprisingly, they compete with each other to attract investment. This competition for FDI has spawned literally thousands of international investment agreements intended to protect— and hence attract—investors, and, more generally, create a favorable investment climate. BITs and DTTs are a central part of this process as they are seen to enhance the locational advantages of countries by enshrining certain approaches to the treatment of foreign investors in international treaties, thereby improving the regulatory environment for investment. Even in the absence of conclusive evidence as to the effect of BITs and DTTs on FDI flows, countries continue to conclude these agreements, and the number of such treaties continues to grow. Governments could be signing these treaties because, as more countries conclude more and more of these agreements, they could be afraid that investors may avoid investing in countries that have not signed such treaties—so countries (especially developing countries) may feel they need to sign these agreements to stay competitive, or at least “to appear enlightened or receptive to modern international law trends.”72 UNCTAD has suggested that, in some cases, foreign

72. Ginsburg, op. cit., p. 117.

bit s , dtt s, and fdi flows: an overview lxi

investors with existing investments have encouraged their home country governments—or the host country governments—to conclude BITs to protect existing investments; this means that studies that find that BITs did not stimulate FDI flows might overlook that BITs positively affect FDI flows by helping host countries to retain existing levels of FDI.73 It is also possible that governments, even if they are not entirely sure whether BITs and DTTs lead to higher FDI flows, think that these treaties do not hurt such flows and, in any event, can serve other purposes—although there are trade-offs in terms of accepting international disciplines, with the corresponding reduction of national policy space. For example, some governments may want to use the commitments they have entered into in these treaties to advance domestic policy reforms. Conversely, governments could also be signing these agreements to signal to investors that they are prepared to bind their improved national policy frameworks and the regulatory changes that favor FDI in international agreements that cannot be changed unilaterally. This may be particularly important for countries that are politically or economically instable, or countries with high levels of corruption, as “investors may be especially concerned about the permanence or strength of domestic reforms implemented in [such] countries.”74 In that case, BITs and DTTs “may be the result of policy changes rather than the embodiment of them,” which is supported by the fact that, simultaneously with the adoption of these bilateral treaties, countries “were also adopting internal regulatory changes that made foreign investment more liberal.”75 Finally, governments that would want to strengthen the positive effects of especially BITs on FDI flows could go beyond relying on the indirect effects that are thought to be associated with better protection. They could do this by stipulating in BITs various measures that home countries could take to increase FDI flows to developing countries. Such measures could include, for example, various fiscal and financial incentives that home countries could grant to their firms if they invest in developing countries (and especially the least developed among them); technical assistance to build investment promotion capacities; information about investment opportunities; and improved market access. Such commitments, in fact, could also extend to efforts to enhance the benefits of FDI to host countries and their economic growth and development, for example, through the promotion of technology transfer and the creation of more linkages between foreign affiliates and domestic firms.76 The negotiation of new BITs and the renegotiation of BITs underway may offer opportunities to do so.

73. UNCTAD, Bilateral Investment Treaties in the Mid-1990s (New York: United Nations, 1998), p. 142. 74. Swenson, op. cit., p. 133. 75. Ginsburg, op. cit., p. 117. 76. See UNCTAD, Investment Promotion Provisions in International Investment Agreements (Geneva: United Nations, forthcoming).

lxii lisa e. sachs and karl p. sauvant

Aside from the specific motivation for or impact of these investment agreements, there is another effect of the proliferation of BITs: they strengthen the rule of law in the sphere of international investment and hence contribute to the emergence of international investment law. This is not to suggest that the network of BITs constitutes, in and of itself, a coherent international investment law system. But the fact that the great majority of countries subscribe to a range of standards that are similar in nature and that these standards are being clarified and refined through practice, may indicate that a number of the building blocks for such a system are being put in place. As international investment rule-making involves the great majority of countries,77 is a dynamic process and proceeds at a rapid pace, all countries have the opportunity to participate actively in designing the international investment law system and to seek to influence it in a manner that ensures that their interests are taken into account.

77. Important in this context is that, in contrast to earlier periods, emerging markets participate actively in this process. By the end of 2006, developing countries alone were signatories to 77% of all BITs, 61% of all DTTs, and 81% of all other international investment agreements, and a number of these involve only developing countries. See UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. 17.

part one introduction

This page intentionally left blank

1. a brief history of international investment agreements∗ kenneth j. vandevelde introduction One of the most remarkable phenomena in international law during the past two decades has been the extraordinary increase in the number of agreements concluded relating to the protection or liberalization of foreign investment. More than 2,800 such agreements now exist, with the great majority having been concluded since 1990.1 This number includes almost 2,600 bilateral investment treaties (BITs) as well as nearly 250 trade agreements that contain investment provisions.2 Although the number of agreements increased remarkably after 1990, international agreements relating to investment have a long history. Provisions relating to the protection of property abroad may be found in international agreements dating back to the late eighteenth century.3 This chapter traces the history of these international investment agreements. The history thus far comprises three separate eras. The first, the colonial era, began in the late eighteenth century and continued until the end of the Second World War.4 The second, the postcolonial era, began with the end of the war and continued until approximately 1990, with the collapse of the Soviet Union.5 The third, the global era, began in approximately 1990 and continues to the present.6

∗ This chapter is an update of “A Brief History of International Investment Agreements,” 12 U.C. Davis J. Int’l L. & Pol’y 157 (2005), copyright 2005 by The Regents of the University of California. All rights reserved. It is reprinted here with permission from the publisher. 1. United Nations Conference on Trade and Development [UNCTAD], Recent Developments in International Investment Agreements (2006–June 2007), UNCTAD/ WEB/ITE/IIT/IIA/2007/6 (Jan. 21, 2008), available at http://www.unctad.org/en/docs/ webiteiia20076_en.pdf. 2. According to UNCTAD, at least 2,573 BITs and 241 trade agreements with investment provisions had been concluded by the end of 2006. Id. at 2,6. 3. See infra note 7. 4. See infra Part I. 5. See infra Part II. 6. See infra Part III.

4 kenneth j. vandevelde

A. The Colonial Era Prior to the Second World War, the protection of foreign direct investment was not often a concern in international agreements. Most international economic agreements concerned themselves with establishing trade relations, though these agreements sometimes included provisions on the protection of property of nationals of one country in the territory of another country. The United States, for example, as early as the eighteenth century began to conclude bilateral treaties of “Friendship, Commerce and Navigation” (FCN), the purpose of which was to establish trade relations with its treaty partners.7 These treaties included provisions guaranteeing “special protection”8 or “full and perfect protection”9 to the property of nationals of one party in the territory of another party. They also required payment of compensation for expropriation10 and guaranteed to nationals of one party most favored nation (MFN) and national treatment with respect to the right to engage in certain business activities in the territory of the other party.11 Occasionally, they even provided limited protection for currency transfers.12 The focus was on protecting property, as opposed to investment. The principal source of norms for the protection of international investment stocks in the colonial era was customary international law, which obligated host

7. Not all of these treaties were formally titled treaties of “friendship, commerce and navigation.” The name is thus a generic one. The first such agreement was the Treaty of Amity and Commerce, U.S.-Fr., July 16, 1782, 8 Stat. 12, negotiated with France in 1778 by Benjamin Franklin, Arthur Lee, and Silas Dean. Other eighteenth century agreements include Treaty of Amity and Commerce, U.S.-Neth., Oct. 8, 1782, 8 Stat. 32; Treaty of Amity and Commerce, U.S.-Swed., Apr. 3, 1783, 8 Stat. 60; Treaty of Amity and Commerce, U.S.-F.R.G. (Prussia), July 9–Sept. 10, 1785, 8 Stat. 84; Treaty of Peace and Friendship, U.S.-Morocco, June 23–July 6, 1786, 8 Stat. 100; Treaty of Amity, Commerce and Navigation, U.S.-G.B., Nov. 19, 1794, 8 Stat. 116; and Treaty of Friendship, Limits and Navigation, U.S.-Spain, Oct. 27, 1795, 8 Stat. 138. 8. See, e.g., General Convention of Peace, Amity, Navigation and Commerce, U.S.Colom., art. Tenth, Oct. 3, 1824, 8 Stat. 306; Treaty of Peace, Friendship, Commerce, and Navigation, U.S.-Bol., art. 13, May 13, 1858, 12 Stat. 1003; General Treaty of Amity, Commerce, and Consular Privilege, U.S.-El Sal., art. 13th, Dec. 6, 1870, 18 Stat. 725. 9. See, e.g., Treaty of Friendship, Commerce, and Navigation, U.S.-Para., art. IX, Feb. 4, 1859, 12 Stat. 1091; Treaty of Friendship, Commerce, and Navigation, U.S.-Arg., art. VII, July 27, 1853, 10 Stat. 1005; Treaty of Friendship, Commerce, and Navigation, U.S.Costa Rica, art. VII, July 10, 1851, 10 Stat. 916. 10. See, e.g., Treaty of Amity, Commerce, and Navigation, U.S.-Congo, art. III, Jan. 24, 1891, 27 Stat. 926; General Treaty of Amity, Commerce, and Consular Privileges, U.S.-El Sal., art. 29th, Dec. 6, 1870, 18 Stat. 725; Treaty of Friendship, Commerce, and Navigation, U.S.- Nicar., art. IX, June 21, 1867, 15 Stat. 549. 11. See, e.g., General Treaty of Amity, Commerce, and Navigation, supra note 8, art. 3rd; Treaty of Commerce, U.S.-Yugo., art. I, Oct. 14, 1881, 22 Stat. 963. 12. Treaty of Commerce, supra note 11, art. II, 22 Stat. at 964.

a brief history of international investment agreements 5

countries to treat investment in accordance with an international minimum standard.13 Customary international law, however, offered an inadequate mechanism for the protection of foreign investment. First, some countries disputed that customary international law imposed an international minimum standard on the treatment of foreign investment.14 Most notably, the Latin American countries adhered to the Calvo doctrine, under which foreign investors were entitled only to the treatment that the host country afforded to its own investors.15 Second, even where it was agreed that an international minimum standard existed, the content of the standard was vague and arguably not particularly demanding.16 Third, in the absence of an agreement by the host country to submit the dispute to arbitration,17 the only mechanism offered by customary law for enforcement of customary norms was espousal. Espousal is a mechanism whereby an injured national’s country assumes the national’s claim as its own and presents the claim against the country that has injured the national.18 Espousal is often an unsatisfactory remedy for a number of reasons. First, the national’s country is under no obligation to espouse a claim19 and, in fact, a home country is often reluctant to espouse because espousal can disrupt the home country’s relations with the host country.20 Further, a home country may espouse a claim only after the national has exhausted his or her remedies under the law of the host country,21 a process that may require a substantial expenditure of time and money without satisfactory

13. See Ian Brownlie, Principles of Public International Law 527–28 (5th ed. 1998). As Brownlie notes, some states disputed the existence of such a standard. Id. 14. Id. at 526–27. 15. See Donald R. Shea, The Calvo Clause: A Problem of Inter-American and International Law and Diplomacy 17–20 (1955). 16. The classic formulation of the standard was that articulated in the Neer Claim, 4 R. Int’l Arb. Awards 60 (1926), in which the commission required that the treatment “amount to an outrage, to bad faith, to willful neglect of duty, or to an insufficiency of governmental action so far short of international standards, that every reasonable and impartial man would readily recognize its insufficiency.” The international minimum standard, however, was also said to include some more rigorous requirements, including an obligation to pay ‘prompt, adequate and effective’compensation for the expropriation of foreign owned property. See infra text accompanying note 96. 17. A state is not subject to the jurisdiction of an international tribunal without its consent. Reparation for Injuries Suffered in the Service of the United Nations, Advisory Opinion, 1949 I.C.J. 174, 177–78 (April 11). 18. See 8 Marjorie Millace Whiteman, Digest of International Law 1216–19 (1967). 19. Id. 20. Kenneth J. Vandevelde, United States Investment Treaties: Policy and Practice 10, 23 (1992). 21. See, e.g., Interhandel (Switz. v. U.S.), 1959 I.CJ. 6, 27 (Mar. 21); Ambatielos Claim (Greece v. U.K.) 23 I.L.R. 306, 344 (1956); 8 Whiteman, supra note 18, at 769–807.

6 kenneth j. vandevelde

resolution.22 Finally, once local remedies have been exhausted and the home country has been persuaded to espouse the claim, the investor loses control over the claim. The home country is entitled to settle the claim on any terms it wishes.23 And, in fact, because espousal is essentially a diplomatic process, there is no guarantee that the host country will agree to resolve the claim on any terms. Diplomacy was effective on occasion. During the nineteenth century the United States, for example, was able to persuade Latin American countries to agree periodically to the submission of claims of injuries to nationals to arbitration.24 As an alternative to diplomacy, nations sometimes utilized military force to protect foreign investments.25 The Roosevelt Corollary to the Monroe Doctrine, for example, explicitly authorized the use of force by U.S. troops in the western hemisphere to collect debts owed to U.S. citizens.26 And, in fact, the United States intervened in Latin America on repeated occasions during the first third of the twentieth century, until the Good Neighbor Policy of the Roosevelt administration ended the practice.27 In summary, several features characterize the international investment regime in the colonial era. First, trade and property protection provisions appeared in the same agreement. In the colonial era, countries generally did not negotiate separate agreements on property or investment. Second, the emphasis of the treaties was on establishing commercial relations. Property protection provisions were present in the agreements, but were clearly secondary in importance to the creation of commercial relations.28 Third, the network of treaties was limited in scope and the protection afforded was weak, particularly insofar as the treaties provided no means for enforcement.29 Thus, the nonlegal mechanisms 22. The requirement of exhaustion, however, may be excused if it appears that such remedies would not be effective. See generally Restatement (Third) Foreign Relations Law § 713 cmt. f (1987); Panevezys-Saldutiskis Ry. (Est. v. Lith) 1939 P.C.I.J. (ser. A/B) No. 76, at 18 (Feb. 28). 23. 8 Whiteman, supra note 18, at 1216–19. 24. One estimate is that between 1829 and 1910, the United States entered into approximately 40 arbitrations with Latin American countries. See Lionel M. Summers, “Arbitration and Latin America,” 3 Cal. W. Int’l L.J. 1, 7 (1972). 25. See, e.g., Edwin M. Borchard, “Limitations on Coercive Protection,” 21 Am. J. Int’l L. 303 (1927); Luis M. Drago, “State Loans in Their Relation to International Policy,” 1 Am. J. Int’l L. 692 (1907). 26. See Armin Rappaport, A History of American Diplomacy 223–29 (1975). 27. Id. at 228–29, 324. 28. See Samuel Bemis, A Diplomatic History of the United States 25–29, 65–84, 101–10, 200–02 (1965) (describing early history of FCN agreements, including their purpose). 29. International investment law as late as the period immediately following the Second World War has been characterized as “an ephemeral structure consisting largely of scattered treaty provisions, a few questionable customs, and contested general principles of law.” Jeswald W. Salacuse & Nicholas P. Sullivan, “Do BITs Really Work? An Evaluation of Bilateral Investment Treaties and Their Grand Bargain,” 46 Harv. Int’l L.J. 67, 68 (2005).

a brief history of international investment agreements 7

of military force and diplomacy were left to provide the principal means for protecting foreign investment. B. The Postcolonial Era The postcolonial era in the history of international investment agreements began with the end of the Second World War and continued until the collapse of the Soviet Union. Three events in particular shaped the structure and content of international investment agreements during that period. First, as a reaction to the severe economic depression that had preceded the war and that many believed had been exacerbated by the protectionist policies of the 1920s,30 the victorious allies forged a consensus in favor of liberalizing trade.31 In 1947 that consensus led to the conclusion of the General Agreement on Tariffs and Trade (GATT),32 which shifted the primary legal framework for international trade relations from bilateral to multilateral agreements and set in motion successive rounds of negotiations aimed at worldwide trade liberalization.33 A separate treaty, the Havana Charter, that was intended to create a liberal investment regime for both trade and investment never entered into force.34 Thus, entry into force of the GATT created a major multilateral organization with competence over trade but not investment. Investment would need to be treated outside the GATT framework, which to a large extent meant separately from trade. As the GATT became the principal forum within which international trade negotiations were conducted, bilateral trade agreements began to diminish in importance. The United States launched a new series of FCN agreements starting in 1946 and continuing for the next twenty years, a period within which twenty-one agreements were concluded.35 After 1966, however, the United States never again concluded an FCN agreement.

30. Bernard Hoekman & Michael Kostecki, The Political Economy of the World Trading System 2–3 (1995). 31. Rondo Cameron, A Concise Economic History of the World 370–71 (3d ed. 1997). 32. General Agreement on Tariffs and Trade, Oct. 20, 1947, 61 Stat. A-11, 55 U.N.T.S. 188. 33. The most recent to conclude was the Uruguay Round, which ended in December 1993. See John Croome, Reshaping the World Trading System (1995). The World Trade Organization is now engaged in the Doha Round of multilateral trade negotiations. 34. Hoekman & Kostecki, supra note 30, at 12–13. 35. For a discussion of the modern FCNs, See Henry C. Hawkins, Commercial Treaties and Agreements: Principles and Practice (1951); Robert Reubert Wilson, The International Law Standards in Treaties of the United States (1953); Robert Reubert Wilson, United States Commercial Treaties and International Law (1960); Herman Walker, Jr., “Modern Treaties of Friendship, Commerce and Navigation,” 42 Minn. L. Rev. 805 (1958); Herman Walker, Jr., “Treaties for the Encouragement and Protection of Foreign Investment: Present United States Practice,” 5 Am. J. Comp. L. 229 (1956); Robert R. Wilson, “A Decade of New Commercial Treaties,” 50 Am. J. Int’l L. 927 (1956); Robert R. Wilson, “Postwar Commercial

8 kenneth j. vandevelde

The postwar U.S. FCNs included the principal property protection provisions that had appeared in the FCNs of the colonial era. The postwar FCNs guaranteed “equitable treatment”36 and the “most constant protection and security”37 to property of foreign nationals and companies. Such property could not be taken without payment of just compensation.38 These FCNs also guaranteed to nationals and companies of one party national and MFN treatment with respect to the right to engage in various types of commercial activities, meaning that foreign investors in effect were entitled to national and MFN treatment with respect to the right to establish investment.39 Treaties of the United States,” 43 Am. J. Int’l L. 262 (1949); and Robert R. Wilson, “PropertyProtection Provisions in United States Commercial Treaties,” 45 Am. J. Int’l L. 83 (1951). 36. For example, the FCN with Greece provided that: “Each Party shall at all times accord equitable treatment to the persons, property, enterprises and other interests of nationals and companies of the other Party.” Treaty of Friendship, Commerce and Navigation, U.S.-Greece, art. I, Aug. 3, 1951, 5 U.S.T. 1829 [hereinafter FCN Greece]. 37. For example, FCN Greece, supra note 36, at art. VII(1), provided that: “Property of nationals and companies of either Party shall receive the most constant protection and security within the territories of the other Party.” 38. For example, FCN Greece provided that: Property of nationals and companies of either Party shall not be taken within the territories of the other Party except for public benefit, nor shall it be taken without the prompt payment of just compensation. Such compensation shall be in an effectively realizable form and shall represent the full equivalent of the property taken; and adequate provision shall have been made at or prior to the time of taking for the determination and payment thereof. It is understood that withdrawal of such compensation shall be in accordance with applicable laws and regulations consistent with the provisions of Article XV [relating to exchange controls] of the present Treaty. The provisions of the present paragraph shall extend to interests held directly or indirectly by nationals and companies of either Party in property which is taken within the territories of the other Party. FCN Greece, supra note 36, at art. VII(3) 39. For example, the Treaty of Friendship, Commerce and Navigation with Japan provided that: 1. Nationals and companies of either Party shall be accorded national treatment with respect to engaging in all types of commercial, industrial, financial and other business activities within the territories of the other Party, whether directly or by agent or through the medium of any form of lawful juridical entity. Accordingly, such nationals and companies shall be permitted within such territories: (a) to establish and maintain branches, agencies, offices, factories and other establishments appropriate to the conduct of their business; (b) to organize companies under the general company laws of such other Party, and to acquire majority interests in companies of such other Party; and (c) to control and manage enterprises which they have established or acquired. Moreover, enterprises which they control, whether in the form of individual proprietorships, companies or otherwise, shall, in all that relates to the conduct of the activities thereof, be accorded treatment no less favorable than that accorded like enterprises controlled by nationals and companies of such other Party.

a brief history of international investment agreements 9

The postwar FCNs included some innovations. First, they extended treaty protections to corporate entities.40 Earlier agreements had protected individuals. The postwar agreements for the first time also regularly included protection against exchange controls.41 Further, these agreements included a dispute

2. Each Party reserves the right to limit the extent to which aliens may within its territories establish, acquire interests in, or carry on public utilities enterprises or enterprises engaged in shipbuilding, air or water transport, banking involving depository or fiduciary functions, or the exploitation of land or other natural resources. However, new limitations imposed by either Party upon the extent to which aliens are accorded national treatment, with respect to carrying on such activities within its territories, shall not be applied as against enterprises which are engaged in such activities therein at the time such new limitations are adopted and which are owned or controlled by nationals and companies of the other Party. Moreover, neither Party shall deny to transportation, communications and banking companies of the other Party the right to maintain branches and agencies to perform functions necessary for essentially international operations in which they are permitted to engage. 3. The provisions of paragraph 1 of the present Article shall not prevent either Party from prescribing special formalities in connection with the establishment of alien-controlled enterprises within its territories; but such formalities may not impair the substance of the rights set forth in said paragraph. 4. Nationals and companies of either Party, as well as enterprises controlled by such nationals and companies, shall in any event be accorded most-favored-nation treatment with reference to the matters treated in the present Article. Treaty of Friendship, Commerce and Navigation, U.S.-Japan, art. VII, Apr. 2, 1953, 4 U.S.T. 2063 [hereinafter FCN Japan]. 40. See Herman Walker, Jr., “Provisions on Companies in United States Commercial Treaties,” 50 Am. J. Int’l L. 373 (1956). 41. For example, FCN Japan, supra note 39, at art. XII, provided that: 1. Nationals and companies of either Party shall be accorded by the other Party national treatment and most-favored-nation treatment with respect to payments, remittances and transfers of funds or financial instruments between the territories of the two Parties as well as between the territories of such other Party and of any third country. 2. Neither Party shall impose exchange restrictions as defined in paragraph 5 of the present Article except to the extent necessary to prevent its monetary reserves from falling to a very low level or to effect a moderate increase in very low monetary reserves. It is understood that the provisions of the present Article do not alter the obligations either Party may have to the International Monetary Fund or preclude imposition of particular restrictions whenever the Fund specifically authorizes or requests a Party to impose such particular restrictions. 3. If either Party imposes exchange restrictions in accordance with paragraph 2 above, it shall, after making whatever provision may be necessary to assure the availability of foreign exchange for goods and services essential to the health and welfare of its people, make reasonable provision for the withdrawal, in foreign exchange in the currency of the other Party, of: (a) the compensation referred to

10 kenneth j. vandevelde

resolution provision consenting to the jurisdiction of the International Court of Justice over disputes involving the interpretation or application of the agreement.42 The inclusion of a dispute resolution provision solved the problem that a host country could not be subject to the jurisdiction of an international tribunal without its consent,43 though it did not relieve investors of the need to exhaust local remedies and to persuade their home country to espouse their claim before pursuing a remedy under international law.44 The conclusion of the postwar FCNs with important investment provisions reflected the fact that investment protection for the first time had become a primary goal of the FCN agreements.45 The United States recognized that a bilateral treaty providing for investment protection was necessary. At the same time, however, the FCN agreements came to be seen as less than ideal vehicles because they were primarily trade agreements and trade relations now were being negotiated principally through the GATT.46 The second major event shaping the international investment regime of the postcolonial era was the process of decolonialization that began after the war and led to the creation of scores of newly independent but economically

in Article VI, paragraph 3, of the present Treaty [relating to expropriation], (b) earnings, whether in the form of salaries, interest, dividends, commissions, royalties, payments for technical services, or otherwise, and (c) amounts for amortization of loans, depreciation of direct investments, and capital transfers, giving consideration to special needs for other transactions. If more than one rate of exchange is in force, the rate applicable to such withdrawals shall be a rate which is specifically approved by the International Monetary Fund for such transactions or, in the absence of a rate so approved, an effective rate which, inclusive of any taxes or surcharges on exchange transfers, is just and reasonable. 4. Exchange restrictions shall not be imposed by either Party in a manner unnecessarily detrimental or arbitrarily discriminatory to the claims, investments, transport, trade, and other interests of the nationals and companies of the other Party, nor to the competitive position thereof. 5. The term “exchange restrictions” as used in the present Article includes all restrictions, regulations, charges, taxes, or other requirements imposed by either Party which burden or interfere with payments, remittances, or transfers of funds or of financial instruments between the territories of the two Parties. 42. For example, FCN Japan, supra note 39, at art. XXIV(2), provided that: “Any dispute between the Parties as to the interpretation or application of the present Treaty, not satisfactorily adjusted by diplomacy, shall be submitted to the International Court of Justice, unless the Parties agree to settlement by some other pacific means.” 43. See supra text accompanying note 17. 44. See supra text accompanying notes 18–23. 45. Vandevelde, supra note 20, at 17. 46. See supra text accompanying notes 32–33.

a brief history of international investment agreements 11

undeveloped countries.47 These newly independent countries were fiercely protective of their independence48 and came to regard foreign investment as a form of neocolonialism because it involved foreign control over the means of production.49 Another concern was that foreign investors would interfere in the domestic affairs of the host country.50 Although foreign direct investment created the greatest concerns, because it involved a foreign presence in the territory of the developing country, even trade with developed countries was viewed with suspicion. The fear was that trade between developed and developing countries would result in the exploitation of the latter.51 Many developing countries closed their economies to new foreign investment and began to expropriate existing investment.52 They also adopted import substitution policies, under which they would seek to produce needed goods and services locally rather than importing them from developed countries.53 Where economic relations with other countries were desired, developing countries would seek to form them with other developing countries.54 The third event was the emergence of the socialist bloc led by the Soviet Union. Immediately following the war, the socialist countries undertook massive expropriations of the private sector, including foreign-held assets.55 They also encouraged developing countries in the view that economic relations with the developed countries of Western Europe and North America would be inherently exploitative and that the best path to economic development lay in extensive state regulation of the economy rather than through the free market.56

47. The number of countries more than tripled as a result of decolonialization after the war. David S. Landes, The Wealth and Poverty of Nations 431 (W.W. Norton & Co. Ltd. 1999) (1998). 48. M. Sornarajah, The International Law on Foreign Investment 12 (1994). 49. Dean Hanink, The International Economy: A Geographic Perspective 234 (1994). 50. Robert Gilpin with assistance of Jean M. Gilpin, The Political Economy of International Relations 247–48 (1987); Michael P. Todaro, Economic Development 534 (5th ed. 1994). 51. Barry W. Poulson, Economic Development: Private and Public Choice 39 (1994). 52. These expropriations included notably the seizure of petroleum assets in Iran in 1951 and in Libya in 1955, and Castro’s expropriation of the private sector in Cuba starting in 1959. These waves of expropriations continued in the 1970s. One study by the United Nations has identified 875 expropriations occurring in sixty-two countries between 1960 and 1974. Salacuse & Sullivan, supra note 29, at 75 n.54. 53. John Rapley, Understanding Development: Theory and Practice in the Third World 22–25 (1996). 54. Starting in the 1960s, developing countries made numerous attempts to create preferential trading arrangements among themselves, most of which were unsuccessful. Dominick Salvatore, International Economics 314–15, 323–25 (5th ed. 1995). 55. Michael Barrett Brown, Models in Political Economy 193–267 (Penguin Books 2d ed. 1995); Rapley, supra note 53, at 44. 56. See generally E. Wayne Nafziger, The Economics of Developing Countries 106–08 (3d ed. 1997); Rapley, supra note 53, at 18–20.

12 kenneth j. vandevelde

By the early 1970s, the developing and socialist countries had mounted an effort in the UN General Assembly, where they held a numerical majority, to establish recognition of their right to expropriate foreign investment without payment of fair market value for the expropriated assets. On May 1, 1974, the General Assembly adopted the Declaration of a New International Economic Order (NIEO),57 which declared that states have “[f]ull permanent sovereignty” over their natural resources and other economic activities.58 State sovereignty includes “the right of nationalization or transfer of ownership to its nationals.”59 The declaration did not specify any obligation to pay compensation. On December 12, by a vote of 120-6 with ten abstentions,60 the General Assembly adopted the Charter of Economic Rights and Duties of States (CERDS).61 Article 2.2(c), which was adopted by a separate vote of 104-16, with six abstentions,62 declared that each state has the right “[t]o nationalize, expropriate or transfer ownership of foreign property, in which case appropriate compensation should be paid by the State adopting such measures, taking into account its relevant laws and regulations and all circumstances that the State considers pertinent.”63 The charter thus stated that compensation should be paid, not that it must be paid, and that the amount of compensation would be based on national law, which might not provide for any compensation, rather than international law.64

57. Declaration on the Establishment of a New Economic Order, G.A. Res. 3201(S-VI), UN GAOR, 6th Special Sess., 2229th plen. mtg., UN Doc. A/RES/3201(S-VI) (May 1, 1974), reprinted in 13 I.L.M. 715 (1974). 58. Id. p. 4(e). 59. Id. 60. The six states that voted in opposition were Belgium, Denmark, the Federal Republic of Germany, Luxembourg, the United Kingdom and the United States. The ten states that abstained were Austria, Canada, France, Ireland, Israel, Italy, Japan, the Netherlands, Norway, and Spain. Charter of Economic Rights and Duties of States, G.A. Res. 3281 (XXIX), UN GAOR, 29th Sess., 2315th plen. mtg., UN Doc. A/RES/3281(XXIX) (Dec. 12, 1974), reprinted in 14 I.L.M. 251 (1975). 61. Charter of Economic Rights and Duties of States, G.A. Res. 3281 (XXIX), UN GAOR, 29th Sess., 2315th plen. mtg., UN Doc. A/RES/3281(XXIX) (Dec. 12, 1974), reprinted in 14 I.L.M. 251 (1975). 62. The sixteen states that voted in opposition were Austria, Belgium, Canada, Denmark, Federal Republic of Germany, France, Ireland, Italy, Luxembourg, Japan, the Netherlands, Norway, Spain, Sweden, the United Kingdom and the United States. The six states that abstained were Australia, Barbados, Finland, Israel, New Zealand and Portugal. Id. 63. Id. at art. 2, p. 2(c). 64. See generally Charles N. Brower & John B. Tepe, Jr., “The Charter of Economic Rights and Duties of States: A Reflection or a Rejection of International Law?,” 9 Int’l Law. 295 (1975); Burns H. Weston, “The Charter of Economic Rights and Duties of States and the Deprivation of Foreign Owned Wealth,” 75 Am. J. Int’l L. 437 (1981).

a brief history of international investment agreements 13

Developed countries responded to the threat of uncompensated expropriation by creating the bilateral investment treaty (BIT).65 The UN Charter, adopted at the end of the war, had prohibited the use of military force66 except in selfdefense,67 which rendered the use of force to collect debts or protect investment illegal under international law.68 Given the serious deficiencies of customary international law as a means of protecting international investment,69 treaties offered potentially the most effective means for preventing uncompensated expropriations. Germany was the first to conclude such an agreement. Having lost its foreign investment as a result of its defeat in the Second World War, Germany was

65. The BIT programs as they emerged in the early years are described in Rudolf Dolzer & Margrete Stevens, Bilateral Investment Treaties (1995); Vandevelde, supra note 20, Adeoye Akinsanya, “International Protection of Direct Foreign Investments in the Third World,” 36 Int’l & Comp. L.Q. 58 (1987), Eileen Denza & Shelagh Brooks, “Investment Protection Treaties: United Kingdom Experience,” 36 Int’l & Comp. L.Q. 908 (1987); Pamela B. Gann, “The U.S. Bilateral Investment Treaty Program,” 21 Stan. J. Int’l L. 373 (1985); Mohamed I. Khalil, “Treatment of Foreign Investment in Bilateral Investment Treaties,” 7 ICSID Rev. 339 (1992); Palitha T.B. Kohona, “Investment Protection Agreements: An Australian Perspective,” 21 J. World Trade L. 79 (1987); T. Modibo Ocran, “Bilateral Investment Protection Treaties: A Comparative Study,” 8 N.Y.L. Sch. J. Int’l & Comp. L. 401 (1987); Robert K. Paterson, “Canadian Investment Promotion and Protection Treaties,” 29 Can. Y.B. Int’l L. 373 (1991); Jeswald W. Salacuse, “BIT by BIT: The Growth of Bilateral Investment Treaties and Their Impact on Foreign Investment in Developing Countries,” 24 Int’l Law. 655 (1990); M. Sornarajah, “State Responsibility and Bilateral Investment Treaties,” 20 J. World. Trade L. 79 (1986); Kenneth J. Vandevelde, “The Bilateral Investment Treaty Program of the United States,” 21 Cornell Int’l L.J. 201 (1988) [hereinafter Vandevelde, The Bilateral Investment Treaty Program]; and Kenneth J. Vandevelde, ”U.S. Bilateral Investment Treaties: The Second Wave,” 14 Mich. J. Int’l L. 621 (1993). 66. UN Charter art. 2, para. 4 (Providing in part that “[a]ll members shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state . . . .”). 67. UN Charter art. 51 (Declaring that “[n]othing in the present charter shall impair the inherent right of individual or collective self-defense if an armed attack occurs against a Member of the United Nations until the Security Council has taken the measures necessary to maintain international peace and security.”). 68. Even before the adoption of the UN Charter, the use of force to collect debts had become increasingly controversial. For example, Article I of the Hague Convention of 1907 had outlawed the use of force to collect contract debts owed to private citizens of one state by the government of another state unless the debtor state refused to submit the dispute to arbitration. Pacific Settlement of International Disputes (Hague, I), Oct. 18, 1907, T.S. No. 536, 1 Bevans 577. The American delegation to the conference that drafted the convention had supported this provision. See 6 Green Hackworth, Digest of International Law 152 (1943). 69. See supra text accompanying notes 14–17.

14 kenneth j. vandevelde

especially sensitive to the political risks to which foreign investment was exposed.70 In 1959, Germany concluded the first two BITs, one with Pakistan and the other with the Dominican Republic.71 Other Western European countries quickly followed Germany’s lead. France72 concluded its first BIT in 1960, Switzerland73 in 1961, the Netherlands74 in 1963, Italy75 and the Belgium-Luxembourg Union76 in 1964, Sweden77 and Denmark78 in 1965, and Norway79 in 1966. Additional expropriations during the 1970s80 and the adoption of the NIEO and CERDS resolutions triggered additional BIT programs in the 1970s. The United Kingdom81 concluded its first BIT in 1975, Austria82 in 1976, and Japan83 in 1977. The United States made the decision to inaugurate its BIT program in 1977, the first year of the Carter administration, although it did not successfully complete a negotiation until the 1980s.84 These new bilateral investment treaties were remarkably uniform in content and contained several distinctive features.85 First, the BITs, as their name implied, dealt exclusively with investment. The developed countries recognized that trade was now within the province of the GATT. Other factors, however, also militated in favor of agreements restricted solely to investment. For example, the inclusion of noninvestment issues, it was feared, would make the agreements too complex and difficult to conclude.86 Thus, the establishment of investment protection might be delayed by disagreements over unrelated issues. At the same time, the desire of developing countries to obtain concessions in noninvestment areas might induce them to conclude BITs containing investment obligations that they could not honor, with the result that conclusion of the agreements might offer investors false security.87 At least in the United States, it was hoped

70. Salacuse & Sullivan, supra note 29, at 73. 71. United Nations Conference on Trade and Development, Bilateral Investment Treaties in the Mid-1990s at 8, 177, UN Sales No. E.98.II.D.8 (1998). 72. Id. at 175 (agreement was with Chad). 73. Id. at 205 (agreement was with Tunisia). 74. Id. at 192 (agreement was with Tunisia). 75. Id. at 184 (agreement was with Guinea). 76. Id. at 163 (agreement was with Tunisia). 77. Id. at 204 (agreement was with Ivory Coast). 78. Id. at 172 (agreement was with Madagascar). 79. Id. at 193 (agreement was with Madagascar). 80. See supra text accompanying note 52. 81. UNCTAD, supra note 71, at 211 (agreement was with Egypt). 82. Id. at 161 (agreement was with Romania). 83. Id. at 185 (agreement was with Egypt). 84. Vandevelde, The Bilateral Investment Treaty Program, supra note 65, at 209–11. 85. See generally UNCTAD, supra note 71; Dolzer & Stevens, supra note 65. 86. Vandevelde, The Bilateral Investment Treaty Program, supra note 65, at 210. 87. Vandevelde, supra note 20, at 26, 31–32.

a brief history of international investment agreements 15

that the BITs would reflect a genuine commitment to investment protection on which investors could rely.88 Second, BITs were negotiated principally between a developed and a developing country.89 The underlying assumption was that the agreement would protect the investment of the developed country in the territory of the developing country.90 Typically, the agreement was drafted by the developed country and offered to the developing country for signature, with the final agreement reflecting only minor changes from the original draft.91 This persistent pattern added an ideological dimension to the agreements. Although both parties formally assumed the same obligations, the agreements were perceived as nonreciprocal because in practice the obligations all fell on the developing country.92 Third, the motivation for the developing country to conclude the agreements in most cases was to attract foreign investment.93 The theory was that offering legal protections to foreign investment would induce foreign investors to invest. Fourth, the motivation for the developed country to conclude the agreements was to obtain protection for its foreign investment.94 Whereas the FCN agreements had been concerned principally with establishing economic relations, the BITs were a defensive reaction to past expropriations of existing investments without payment of fair market value.95 In the United States in particular, the decision to negotiate BITs was very much motivated by a desire to create a network of treaties adopting the standard of prompt, adequate, and effective compensation for expropriation, a standard that, among other things, required payment of fair market value.96 Indeed, the United States hoped that the conclusion of a sufficiently large network of treaties embracing that standard 88. Vandevelde, The Bilateral Investment Treaty Program, supra note 65, at 211–12. 89. UNCTAD, supra note 71, at 8–19. 90. Id. at 1. 91. For a discussion of the U.S. experience in this regard, see Vandevelde, The Bilateral Investment Treaty Program, supra note 65, at 211–13. For a general discussion of BIT negotiations, see Kenneth J. Vandevelde, Treaty Interpretation from a Negotiator’s Perspective, 21 Vand. J. Transnat’l L. 281 (1988). 92. Sornarajah, supra note 48, at 227. 93. UNCTAD, supra note 71, at 5. 94. Id. at 2–6. 95. Vandevelde, supra note 20, at 20. 96. Id. at 21. The standard of prompt, adequate and effective compensation for expropriation originally was articulated in 1938 by U.S. Secretary of State Cordell Hull in a note to the Mexican government. Id. at 118. This language soon became the definitive formulation in the view of the United States of the standard of compensation required by customary international law in the event of an expropriation of foreign owned property, supplanting earlier formulations, such as ‘just compensation.’Id. The standard was understood to require payment without delay of fair market value in a freely convertible currency. Id.

16 kenneth j. vandevelde

would provide evidence that the standard was a norm of customary international law and thus applied to expropriations even in the absence of a treaty.97 The United States therefore refused to conclude any BIT unless it adopted that standard explicitly.98 Establishing the principle of prompt, adequate, and effective compensation was more important to the United States than obtaining protection for any specific asset of foreign investment.99 Fifth, presumably because of the perceived nonreciprocal nature of the agreements and the skepticism of many developing countries toward foreign investment, the network of BITs remained somewhat limited. From 1959 to 1969, only seventy-five BITs worldwide were concluded, fewer than seven per year.100 From 1970 to 1979, ninety-two BITs were concluded, the rate increasing slightly to more than nine per year.101 The rate more than doubled in the 1980s to more than twenty-one agreements each year, for a total of 219 during that decade.102 Still, from 1959 until 1989, only 386 agreements were concluded, an average of about one per month worldwide.103 Sixth, the protections provided by the BITs were similar to those that had been provided in the modern FCNs concluded by the United States. Typically, they contained a guarantee of national and most favored nation treatment for covered investment,104 a promise of “fair and equitable treatment” for covered investment,105 a commitment to pay prompt, adequate, and effective compensation

97. Id. 98. Id. at 25–26. 99. Id. at 125. 100. UNCTAD, supra note 71, at 9. 101. Id. 102. Id. 103. For a list of BITs concluded from the inception of the program through 1989, see id. at 159–217. 104. For example, the BIT between the United States and Estonia provides: Each Party shall permit and treat investment . . . on a basis no less favorable than that accorded in like situations to investment . . . of its own nationals or companies, or of nationals or companies of any third country, whichever is the most favorable, subject to the right of each Party to make or maintain exceptions falling within one of the sectors or matters listed in the Annex to this Treaty. Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Est., art. II(1), Apr. 19, 1994, S. Treaty Doc. No. 103–38 (1994). 105. For example, the BIT between the United States and Trinidad and Tobago provides: “Each Party shall at all times accord to covered investments fair and equitable treatment and full protection and security, and shall in no case accord treatment less favorable than that required by international law.” Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Trin. & Tobago, art. II(3)(a), Sept. 26, 1994, S. Treaty Doc. No. 104–14 (1995).

a brief history of international investment agreements 17

for expropriation of covered investment,106 and restrictions on exchange controls.107 Although the U.S. FCNs had guaranteed nationals of each party national and MFN treatment with respect to the right to “engage in” various commercial enterprises,108 most of the BITs did not guarantee to investors of one party the right to establish investment in the territory of the other party.109 The U.S. BITs were an exception, promising national and MFN treatment with respect to the right to establish investment.110 Like the modern FCNs, the BITs contained a provision for settlement of disputes between the parties, although the mechanism was different. Whereas the modern FCNs had provided for submission of disputes to the International Court of Justice, the BITs provided for submission of disputes to an ad hoc arbitral tribunal.111

106. For example, the U.S.-Romania BIT provides: Investments shall not be expropriated or nationalized either directly or indirectly through measures tantamount to expropriation or nationalization (“expropriation”) except: for a public purpose; in a nondiscriminatory manner; upon payment of prompt, adequate and effective compensation; and in accordance with due process of law . . . . Treaty Concerning the Reciprocal Encouragement and Protection of Investment, U.S.-Rom., art. III(1), May 28, 1992, S. Treaty Doc. No. 102–36 (1992). 107. For example, the BIT between the United States and Uzbekistan provides: Each Party shall permit all transfers relating to a covered investment to be made freely and without delay into and out of its territory. Such transfers include: (a) contributions to capital; (b) profits, dividends, capital gains, and proceeds from the sale of all or any part of the investment or from the partial or complete liquidation of the investment; (c) interest, royalty payments, management fees, and technical assistance and other fees; (d) payments made under a contract, including a loan agreement; and (e) compensation pursuant to Articles III [relating to expropriation] and IV [relating to losses due to armed conflict], and payments arising out of an investment dispute. Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Uzb., art. V(1), Dec. 16, 1994, S. Treaty Doc. No. 104–25 (1996). 108. See supra text accompanying note 39. 109. In the U.S. BITs, the guarantees of national and MFN treatment apply to the right to establish investment. Thus, covered investors are entitled to establish investment in the host state as long as nationals of the host state or of any third state were permitted to do so. Such a provision, however, is not typical of most BITs. See UNCTAD, supra note 71, at 46. 110. See Treaty Concerning the Encouragement and Reciprocal Protection of Investment, supra note 104, where the agreement cited requires each party to “permit” investment on a national and MFN basis. 111. The BIT between the United States and Ecuador, for example, provides at art. VII: 1. Any dispute between the Parties concerning the interpretation or application of the Treaty which is not resolved through consultations or other diplomatic channels, shall be submitted, upon the request of either Party, to an arbitral tribunal

18 kenneth j. vandevelde

One major innovation in the BITs was the inclusion, starting in the mid1960s, of a provision in which the host country consented to arbitration of certain disputes with investors, typically those involving the provisions of the agreement.112 This provision was prompted by the conclusion of a 1965 convention

for binding decision in accordance with the applicable rules of international law. In the absence of an agreement by the Parties to the contrary, the arbitration rules of the United Nations Commission on International Trade Law (UNCITRAL), except to the extent modified by the Parties or by the arbitrators, shall govern. 2. Within two months of receipt of a request, each Party shall appoint an arbitrator. The two arbitrators shall select a third arbitrator as Chairman, who is a national of a third State. The UNCITRAL Rules for appointing members of three member panels shall apply mutatis mutandis to the appointment of the arbitral panel except that the appointing authority referenced in those rules shall be the Secretary General of the [International Centre for the Settlement of Investment Disputes]. 3. Unless otherwise agreed, all submissions shall be made and all hearings shall be completed within six months of the date of selection of the third arbitrator, and the Tribunal shall render its decisions within two months of the date of the final submissions or the date of the closing of the hearings, whichever is later. Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Ecuador, art. VII, Aug. 27, 1993, S. Treaty Doc. No. 103-15 (1993). 112. The BIT between the United States and Albania, for example, provides at article IX: 1. For purposes of this Treaty, an investment dispute is a dispute between a Party and a national or company of the other Party arising out of or relating to an investment authorization, an investment agreement or an alleged breach of any right conferred, created or recognized by this Treaty with respect to a covered investment. 2. A national or company that is a party to an investment dispute may submit the dispute for resolution under one of the following alternatives: (a) to the courts or administrative tribunals of the Party that is a party to the dispute; or (b) in accordance with any applicable, previously agreed dispute-settlement procedures; or (c) in accordance with the terms of paragraph 3. 3. (a) Provided that the national or company concerned has not submitted the dispute for resolution under paragraph 2(a) or (b), and that three months have elapsed from the date on which the dispute arose, the national or company concerned may submit the dispute for settlement by binding arbitration: (i) to the Centre, if the Centre is available; or (ii) to the Additional Facility of the Centre, if the Centre is not available; or (iii) in accordance with the UNCITRAL Arbitration Rules; or (iv) if agreed by both parties to the dispute, to any other arbitration institution or in accordance with any other arbitration rules. (b) A national or company, notwithstanding that it may have submitted a dispute to binding arbitration under paragraph 3(a), may seek interim injunctive

a brief history of international investment agreements 19

establishing the International Centre for Settlement of Investment Disputes (ICSID),113 an entity affiliated with the World Bank that was intended to provide a venue for arbitration of disputes between investors and host country governments. For the first time, investors had an effective remedy for unlawful actions by host countries that injured their investments that did not depend upon military action or espousal of their claim by their home country.114 Further, the BITs typically did not require the investor to exhaust local remedies before resorting to international arbitration, though they sometimes required the investor to pursue local remedies for a limited period of time, after which the investor could arbitrate the dispute.115 In providing the investor with a legal remedy that did not depend upon espousal, these BIT provisions depoliticized investment disputes. That is, they placed investment protection in the realm of law rather than politics.116 C. The Global Era The global era in the history of international investment agreements begins at the end of the 1980s. This era reflects profound changes in the context in which international investment agreements were negotiated.117 One of the most important changes was the intermingling of trade and investment provisions in international agreements. The completion of the Uruguay

relief, not involving the payment of damages, before the judicial or administrative tribunals of the Party that is a party to the dispute, prior to the institution of the arbitral proceeding or during the proceeding, for the preservation of its rights and interests. 4. Each Party hereby consents to the submission of any investment dispute for settlement by binding arbitration in accordance with the choice of the national or company under paragraph 3(a)(i), (ii), and (iii) or the mutual agreement of both parties to the dispute under paragraph 3(a)(iv). . . . 5. Any arbitration under paragraph 3(a)(ii), (iii) or (iv) shall be held in a state that is a party to the United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards, done at New York, June 10, 1958. 6. Any arbitral award rendered pursuant to this Article shall be final and binding on the parties to the dispute. Each Party shall carry out without delay the provisions of any such award and provide in its territory for the enforcement of such award. Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Alb., Jan. 11, 1995, S. Treaty Doc. No. 104–19 (1995). 113. Convention on the Settlement of Investment Disputes Between States and Nationals of Other States, March 18, 1965, 17 U.S.T. 1270, T.I.A.S. No. 6090. 114. Vandevelde, supra note 20, at 22–25. 115. UNCTAD, supra note 71, at 93. 116. Vandevelde, supra note 20, at 22–25. 117. For a more extended discussion of these changes, see Kenneth J. Vandevelde, Sustainable Liberalism and the International Investment Regime, 19 Mich. J. Int’l L. 373 (1998).

20 kenneth j. vandevelde

Round of GATT negotiations resulted in the creation of the World Trade Organization (WTO) in 1995 to administer the GATT and related agreements and the explicit injection of investment-related issues into the jurisdiction of the WTO.118 The most important way in which this occurred was through the conclusion of the General Agreement on Trade in Services (GATS),119 which was intended to remove barriers to cross-border trade in services. Services are sometimes provided internationally through the establishment by a service provider of one country of an office or subsidiary in the territory of another country, in effect, by the establishment of foreign investment in the country where the services are to be consumed. The GATS explicitly applies to the delivery of services through such a “commercial presence.”120 Thus, a GATS commitment to allow trade in a certain service sector through a commercial presence amounts to a commitment to allow the establishment of foreign investment. Further, GATS commitments concerning the treatment that the service provider will receive constitute commitments to protect foreign investment.121 Thus, the WTO potentially has jurisdiction over all foreign investment in the service sector of the economy. The potential significance of this agreement is reflected in the fact that, as of 2005, the stock of foreign direct investment in the services sector was $6.11 trillion, compared with the $2.98 trillion stock of foreign direct investment in the manufacturing sector.122 Other agreements further expanded the jurisdiction of the WTO with respect to investment matters beyond the service sector. The Agreement on Trade Related Investment Measures123 prohibits the imposition on foreign investment of certain trade distorting performance requirements. The Agreement on Trade-Related

118. See Hoekman & Kostecki, supra, note 30. 119. General Agreement on Trade in Services, Apr. 15, 1994, Marrakesh Agreement Establishing the World Trade Organization, Annex 1B, Legal Instruments - Results of the Uruguay Round, 33 I.L.M. 44 (1994) [hereinafter GATS]. 120. GATS, supra note 119 at art. I.2, provides that “[f]or purposes of this Agreement, trade in services is defined as the supply of a service . . . (d) by a service supplier of one Member, through commercial presence in the territory of another Member.” 121. For example, Article II requires most favored nation treatment of trade in services, Article III imposes certain obligations of transparency with respect to trade in services, Article VI imposes restrictions on domestic regulation of trade in services, and Article XI limits states’ ability to restrict payments for current transactions relating to trade in services. GATS, supra note 119. 122. UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. 225. 123. Agreement on Trade Related Investment Measures, Apr. 15, 1994, Agreement Establishing the World Trade Organization, Annex 1B, Legal Instruments - Results of the Uruguay Round, vol. 31, 33 I.L.M. 1125 (1994).

a brief history of international investment agreements 21

Intellectual Property Rights124 obligates the parties to provide certain protection for intellectual property, a form of investment. The global era also witnessed an explosion in the number of BITs. This explosion seems to have been rooted in two major causes. The first cause was the victory of market ideology. The economic success of several Asian economies that had high rates of private investment and promoted the production of goods for export, relative to that of other developing countries that had pursued import substitution policies, demonstrated the constructive role that foreign investment and global integration could play in a developing economy.125 Specifically, during the period from 1965 to 1990, eight Asian economies126 grew at a rate three times that of Latin America and twenty-five times that of sub-Saharan Africa.127 Further, the collapse of the Soviet bloc128 discredited the principal alternative to market capitalism as an economic policy.129 The second cause was a loss of alternatives to foreign investment as a source of capital. The debt crisis of the 1980s had reduced the availability of private lending,130 which, by 1980, had accounted for half of all capital flows to developing countries.131 In addition, the massive federal deficits created during the Reagan administration had prompted extensive borrowing by the United States Government, which absorbed much of the available capital,132 further crowding

124. Agreement on Trade-Related Aspects of Intellectual Property Rights, Apr. 15, 1994, Agreement Establishing the World Trade Organization, Annex 1C, Legal Instruments Results of the Uruguay Round, vol. 31, 33 I.L.M. 1125 (1994). 125. World Bank, The East Asian Miracle: Economic Growth and Public Policy 40–42 (1993); Alex E. Fernandez Jilberto & Andre Mommen, Setting the Neoliberal Development Agenda, in Liberalization in the Developing World 1, 3–4 (Alex E. Fernandez Jilberto & Andre Mommen eds., 1996). 126. The eight economies were Japan, the four “Asian Tigers” (Hong Kong, Republic of Korea, Singapore and Taiwan) and Indonesia, Malaysia and Thailand. 127. World Bank, The East Asian Miracle: Economic Growth and Public Policy 2 (1993). 128. On the disintegration of the Soviet bloc, see Michael Mandelbaum, “Coup de Grace: The End of the Soviet Union,” 71 Foreign Aff. 164 (1992), Michael Mandelbaum, “The Bush Foreign Policy,” 70 Foreign Aff. 5 (1991), Coit D. Blacker, “The Collapse of Soviet Power in Europe,” 70 Foreign Aff. 88 (1991). 129. Rapley, supra note 53, at 70; Mark Kramer, “Eastern Europe Goes to Market,” 86 Foreign Pol’y 134 (1992). 130. See Rapley, supra note 53, at 37–38; Jahangier Amuzegar, “Dealing with Debt,” 68 Foreign Pol’y 140, 141–42 (1987); Samuel Britain, “A Very Painful World Adjustment,” 61 Foreign Aff. 541, 541–48 (1983); Pedro-Pablo Kucynski, “Latin American Debt,” 61 Foreign Aff. 344, 350–51 (1982). 131. Poulson, supra note 51, at 450–51 (1994). 132. During the three years between 1982 to 1985, the United States changed from the world’s largest creditor to the world’s largest debtor. See Steven Husted & Michael Melvin, International Economics 314 (3d ed. 1995); Lester C. Thurow & Laura D’Andrea Tyson, “The Economic Black Hole,” 67 Foreign Pol’y 3 (1987).

22 kenneth j. vandevelde

developing countries out of the private market for credit.133 Reductions in developmental assistance at the behest of the Reagan administration during that same period had reduced the availability of public financing.134 For example, between 1980 and 1987 the United States reduced its contributions to multilateral development banks from $2.3 billion annually to $1.1 billion.135 Developing countries seeking capital to finance development increasingly had little alternative but to seek foreign private investment. For both these reasons, then, by the late 1980s developing countries were abandoning the hostility to foreign investment that had characterized the postcolonial era and seeking openly to attract foreign investment by creating a favorable environment for such investment.136 It appeared that the policies of hostility to foreign investment, import substitution, and closed markets to foreign goods, services, and capital that they had employed during the postcolonial era had been a mistake.137 Discussions of a New International Economic Order and the right to expropriate without payment of compensation disappeared.138 Latin American countries abandoned the Calvo Doctrine and agreed to the imposition of international minimum standards for the protection of foreign investment.139 Developing countries rushed to attract foreign investment by demonstrating their support for market capitalism in general and a secure investment climate in particular. Concluding BITs that guaranteed protection for foreign investment offered a mechanism for signaling a desire to attract foreign investment by providing a more secure environment for such investment.140

133. Jeffrey E. Garten, “Gunboat Economics,” 63 Foreign Aff. 538, 546 (1984). 134. Id. at 552–54. 135. John W. Swell & Christine E. Contee, “Foreign Aid and Gramm-Rudman,” 65 Foreign Aff. 1015, 1022 (1987) (total U.S. foreign assistance increased during this time, but only because of increase in military assistance). 136. See Kenneth J. Vandevelde, “Investment Liberalization and Economic Development: The Role of Bilateral Investment Treaties,” 36 Colum. J. Transnat’l L. 501, 502–03 (1998). 137. Jeswald W. Salacuse, ”From Developing Countries to Emerging Markets: A Changing Role for Law in the Third World,” 33 Int’l Law. 875, 882–86 (1999). 138. See Thomas Waelde, “Requiem for the “‘New International Economic Order,’” in Festschrif Fuer Ignaz Seidl-Hohenveldern 771 (Gerhard Hafner et al. eds., 1998). 139. They did so by the conclusion of BITs, in which they agreed to certain standards for the treatment of foreign investment and to the submission of disputes with investors involving the BITs to binding arbitration. For example, Bolivia and Uruguay concluded BITs in 1987, Argentina and Venezuela in 1990, and Chile and Peru in 1991. UNCTAD, supra note 71, passim. 140. Id. at 6.

a brief history of international investment agreements 23

Starting in the late 1980s, the number of BITs concluded accelerated dramatically. While fewer than 400 BITs had been concluded in the thirty years from 1959 to 1989,141 during the next fifteen years some 2,000 BITs would be concluded.142 The BITs concluded in the era of globalization were not much changed in content from the BITs of the postcolonial era.143 They still principally addressed the traditional problem of investment protection. The U.S. BITs and a few others have incorporated some changes that were largely in reaction to arbitral claims filed under the NAFTA investment chapter, but that did not alter the basic nature of the treaty. These changes included the addition of language specifying that the fair and equitable treatment standard merely incorporates the international minimum standard,144 clarifying the extent to which the expropriation provision applies to regulatory actions by the host country145 and modifying the procedures employed in arbitrations under the investor-state dispute resolution provision.146 As the trend toward liberalization continued and countries sought deeper integration than they believed that they were able to achieve within the WTO framework or with agreements limited to investment,147 bilateral and regional trade agreements with investment-related provisions greatly proliferated in number. As of June 2007, at least 251 preferential trade agreements with investment-related provisions had been concluded, 89% of which had been concluded since the 1990s. The United States, for example, has concluded such free trade

141. Id. at 9. 142. UNCTAD, supra note 1, at 1–2. 143. See UNCTAD, supra note 71. 144. See text infra, at note 192. Such language appears in the 2004 U.S. model BIT at article 5 and in Annex A. The 2004 model BIT is available online at http://www.state.gov/ documents/organization/38710.pdf. 145. See text infra, at note 192. Such language appears in the 2004 model BIT, supra note 144, at Annex B. 146. Among the changes to the investor-state dispute resolution mechanism in the 2004 model BIT, supra note 144, are a three year limitations period for bringing claims, id. at art. 26, a provision for expedited consideration of challenges to the legal sufficiency of a claim, id. at art. 28, a provision to make the documents submitted in an arbitral proceeding public and to open the hearings, id. at art. 29, and a provision by which the parties to the agreement may provide binding interpretations of the treaty and its annexes to the tribunal, id. at arts. 30–31. 147. See Jong-Wha Lee, et al., Proliferating Regional Trade Agreements: Why and Whither? (2004); Richard E. Feinberg, “The Political Economy of United States’ Free Trade Agreements,” 26 World Econ. 1019 (2003).

24 kenneth j. vandevelde

agreements with Singapore,148 Chile,149 Australia,150 Morocco,151 the Central American States,152 Peru,153 Oman,154 Columbia,155 Panama,156 and the Republic of Korea.157 Free trade agreements between the United States and Jordan and the United States and Bahrain did not include an investment chapter because the United States already had concluded BITs with each country.158

148. United States-Singapore Free Trade Agreement, signed, January 15, 2003, available online at http://www.ustr.gov/Trade_ Agreements/Bilateral/Singapore_FTA/Final_ Texts/Section_Index.html. 149. Free Trade Agreement between the Government of the United States of America and the Government of the Republic of Chile, signed June 6, 2003, available online at http://www.ustr.gov/Trade_Agreements/Bilateral/Chile_FTA/Final_Texts/Section_ Index.html. 150. United States-Australia Free Trade Agreement, signed May 18, 2004, available online at http://www.ustr.gov/Trade_Agreements/Bilateral/Australia_ FTA/Final_Text/ Section_Index.html. 151. United States-Morocco Free Trade Agreement, signed June 15, 2004, available online at http://www.ustr.gov/Trade_Agreements/Bilateral/Morocco_ FTA/FInal_Text/ Section_Index.html. 152. Central America--Dominican Republic--United States Free Trade Agreement, signed August 5, 2004, available online at http:// www.ustr.gov/Trade_Agreements/ Bilateral/CAFTA/CAFTA-DR_Final_Texts/Section_ Index.html. 153. United States-Peru Trade Promotion Agreement, signed Apr. 12, 2006, available online at http://www.ustr.gov/Trade_Agreements/Bilateral/Peru_TPA/Final_Texts/ Section_Index.html. 154. Agreement between the Government of the United States of America and the Government of the Sultanate of Oman on the Establishment of a Free Trade Area, signed Jan. 19, 2006, available online at http://www.ustr.gov/assets/Trade_Agreements/Bilateral/ Oman_FTA/Final_Text/asset_upload_file345_8820.pdf. 155.United States-Colombia Trade Promotion Agreement, signed Nov. 22, 2006, available online at http://www.ustr.gov/Trade_Agreements/Bilateral/Colombia_FTA/Final_ Text/Section_Index.html, 156. United States-Panama Trade Promotion Agreement, signed June 28, 2007, available online at http://www.ustr.gov/Trade_Agreements/Bilateral/Panama_FTA/Final_ Text/Section_Index.html 157. Free Trade Agreement between the United States of America and the Republic of Korea, signed June 30, 2007, available online at http://www.ustr.gov/Trade_Agreements/ Bilateral/Republic_of_Korea_FTA/Final_Text/Section_Index.html 158. Agreement between the United States of America and the Hashemite Kingdom of Jordan on the Establishment of a Free Trade Area, signed January 15, 2003, and Agreement between the Government of the United States of America and the Government of the Kingdom of Bahrain on the Establishment of a Free Trade Area, signed September 14, 2004. The U.S.-Jordan BIT dates from 1997. See Treaty Between the Government of the United States of America and the Government of the Hashemite Kingdom of Jordan Concerning the Encouragement and Reciprocal Protection of Investment, with Annex and Protocol, signed July 2, 1997, Treaty Doc. 106-30, 106th Cong. 2d Sess. (2000). The Bahrain BIT dates from 2000. Treaty Between the Government of the United States of

a brief history of international investment agreements 25

The mixing of trade and investment provisions in the same agreement reflected changes in the nature of economic activity. Trade and investment are no longer seen merely as substitutes but as complements. The traditional view was to regard the establishment of a foreign subsidiary as a means for delivering goods or services to a foreign market, particularly when high tariffs made export to that market unattractive.159 In that view, foreign investment was an alternative to trade. In the global era, however, investment increasingly has been seen not as a means of replacing trade but of promoting it. Foreign subsidiaries, once established, were often links in a large chain of production, importing raw materials and parts from other subsidiaries and exporting a product to still other subsidiaries perhaps for further refinement.160 Deeper economic integration thus required lowering barriers both to trade and to investment. As a result, states negotiated bilateral and regional trade agreements that included investment related provisions. This, in turn, meant the return of the “package deal.” The inclusion of investment in a larger group of concessions allowed the parties to offer concessions on investment in exchange for concessions in other areas.161 For example, a state might offer to open its economy to foreign investment in exchange for another party’s offer of market access to goods. This process of deeper economic integration began to occur among states with dissimilar economic circumstances. In the postcolonial era, economic integration agreements typically were concluded among states at similar levels of economic development, with the most notable example being the European Community,162 which later evolved into the European Union.163 In 1990, two highly developed countries, the United States and Canada, launched the negotiation of the North American Free Trade Agreement (NAFTA) with Mexico,

America and the Government of the State of Bahrain Concerning the Encouragement and Reciprocal Protection of Investment, Sept. 29, 1998, Treaty Doc. 106-25, 106th Cong. 2d Sess. (2000). 159. United Nations Conference on Trade and Development, World Investment Report 1996: Investment, Trade and International Policy Arrangements 79 (1996). 160. Id. at 103. 161. For discussions of the use of the package deal in the negotiation of the NAFTA, see Frederick W. Mayer, Interpreting NAFTA: The Science and Art of Political Analysis (1998). 162. The European Community was formed in the 1950s consisting of six West European states: Belgium, France, the Federal Republic of Germany, Italy, Luxembourg, and the Netherlands. Denmark, Ireland, and the United Kingdom joined in 1973, Greece in 1981 and Spain and Portugal in 1986. Thus, in the postcolonial era, the European Community consisted solely of developed, West European countries. See http://europa. eu.int/abc/keyfigures/eu_work_ progress/index_animated_en.htm (European Union membership lists). 163. The European Community became the European Union following the entry into force of the Treaty of European Union in 1993. See id.

26 kenneth j. vandevelde

a developing country.164 Subsequently, the European Union undertook to expand its membership to include the transitional economies that formerly were part of the Soviet bloc.165 After the conclusion of NAFTA, free trade agreements with investment-related provisions between developed and developing countries became increasingly common. By mid-2005, they accounted for 39% of all preferential trade agreements with investment provisions.166 Part of the explanation for this trend was that the old distinction between capital-exporting developed countries and capital-importing developing countries was blurring. A number of developing countries had achieved considerable economic success and were becoming significant exporters of capital. For example, in 2003, Singapore’s stock of direct investment abroad was larger than that of several developed countries, including Austria, Denmark, Finland, Greece, Ireland, Norway, and Portugal.167 At the same time, developed countries were becoming major capital importers. For example, the total stock of foreign direct investment in the United States grew from $83 billion in 1980 to $1.8 trillion in 2006.168 The old distinctions of the postcolonial era further dissolved as the transitional economies of Central and Eastern Europe began to join the major developed economies in the European Union.169 And, whereas economic integration agreements in the past generally had been regional, these new agreements often were between states in different regions.170 Interregional agreements now account for more than half of all preferential trade agreements with investment-related provisions.171 Thus, in the global era, the distinctions that had characterized the postcolonial era had collapsed. Investment provisions appeared in agreements at all 164. See generally Maxwell A. Cameron & Brian W. Tomlin, The Making of NAFTA: How the Deal Was Done (2000); Mayer, supra note 161. 165. In 2004, the European Union added eight members that were former Soviet bloc nations. These were the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic, and Slovenia. In 2007, Bulgaria and Romania became members. Former Soviet bloc states currently seeking membership are Croatia and the Former Yugoslav Republic of Macedonia. Two other states joined the European Union in 2004: Cyprus and Malta. Turkey is also currently seeking membership. See http://www.eurunion.org/ states/offices.htm. 166. UNCTAD, supra note 1, at 11. 167. UNCTAD, supra note 122, at 382–85. 168. UNCTAD, World Investment Report 2007: Transnational Corporations, Extractive Industries and Development (New York and Geneva: United Nations, 2007), p. 255. 169. As discussed supra at note 165, eight former communist states joined the European Union in 2004 and two more in 2007. 170. For example, the United States has concluded free trade agreements outside the Western Hemisphere with Israel, Jordan, Singapore, Australia and Morocco. 171. UNCTAD Research Note, “Recent developments in international investment agreements,” UNCTAD/WEB/ITE/IIT/2005/1, available at: http://www.unctad.org/ sections/dite_dir/docs//webiteiit20051_en.pdf.

a brief history of international investment agreements 27

levels—bilateral, regional, and multilateral. The agreements could address trade in goods, trade in services, investment, or any combination of the three. The parties to the agreement may be developed states, developing states, or both. These states may or may not be in the same region. Although the substance of the BITs of the global era is mostly unchanged from the substance of the postcolonial era BITs, to the extent that trade and investment provisions are being intermingled in free trade agreements the global era resembles the colonial era rather than the postcolonial era. In other respects as well, the investment regime of the global era seems to resemble the colonial era more than the postcolonial. For example, investment policy has lost the ideological division that had characterized the postcolonial era. While in the postcolonial era many had seen international investment agreements as unequal treaties to which developed countries reluctantly and perhaps unwisely adhered to in order to attract foreign investment,172 in the global era states almost universally adopted the view that foreign investment could promote economic prosperity and set about jointly creating legal frameworks that would promote and protect international investment flows. In international economic relations, ideology was replaced with pragmatism and cooperation.173 Similarly, while in the post-colonial era, investment agreements, which is to say primarily BITs, were between a capital-exporting developed country and a capital-importing developing country, in the global era, the number of investment agreements between developing countries has grown remarkably as developing countries have become capital exporters, often to other developing countries.174 At the end of 2006, more than one fourth of all BITs were concluded between developing countries.175 Finally, the very purpose of investment agreements is shifting. While in the postcolonial era investment agreements were intended to protect investment of developed countries in the territory of developing countries primarily against expropriation, in the global era investment agreements increasingly are intended to liberalize investment flows. They have become instruments of globalization, removing barriers to trade and investment, much in the same way that the FCNs

172. See, e.g., Andrew T. Guzman, Why LDCs Sign Treaties That Hurt Them: Explaining the Popularity of Bilateral Investment Treaties, 38 Va. J. Int’l L. 639 (1998). (Ed. Note: An updated version of this article is included as Chapter 3 of this volume.) 173. Vandevelde, supra note 117, at 386–90. 174. By 2004, 49.8% of all foreign direct investment in developing countries came from other developing countries. UNCTAD, World Investment Report 2006: FDI from Developing and Transition Economies: Implications for Development (New York and Geneva: United Nations, 2007), p. 119. 175. UNCTAD, supra note 1, at 5.

28 kenneth j. vandevelde

of the eighteenth and nineteenth centuries sought to establish commercial relations between countries. The global era is also characterized by the number of countries that are now part of the framework of international investment agreements. For example, at least 177 countries were party to at least one BIT by the end of 2006.176 The framework of international investment agreements is truly a global one. D. Looking Forward The global era of investment agreements has raised several issues, the resolution of which could form the basis for the emergence of a fourth era in the history of international investment agreements. First, a great deal of energy has been devoted to the conclusion of international investment agreements. One issue this presents is whether they are effective. To the extent that the purpose of the agreements is to protect foreign investment, one is almost forced to concede their effectiveness. Some 290 arbitrations to enforce provisions of the investment agreements have been filed.177 Because of the confidentiality of proceedings, it is difficult empirically to evaluate data about judgments and awards, if any, but the success of an investment agreement in protecting foreign investment can be measured by more than just the number of favorable arbitral awards. Investors who have disputes with host countries may use the existence of the treaty and the possibility of arbitration as leverage to negotiate a satisfactory resolution of the dispute. In such a case, the treaty undeniably has contributed to the protection of foreign investment, even if no arbitration occurred. To the extent that the purpose of the agreements is to promote investment flows, the evidence is less clear. Perhaps no one contends that investment agreements alone will result in increased investment flows.178 As discussed in the Introduction to this volume, they are simply one factor among many in creating a favorable investment climate.179 Still, the issue has arisen as to whether one can identify a statistical correlation between the number of investment agreements concluded and the amount of foreign investment attracted. Several studies have investigated this issue, with inconsistent results. These studies are, of course, the subject of this volume, and for that reason will not be summarized here.

176. The United Nations Conference on Trade and Development, which contains the most comprehensive BIT database on its website at www.unctad.org, counts 177 economies as being party to one or more BITs. 177. UNCTAD IIA Monitor No. 1 (2008), Latest Developments in investor-State Dispute Settlement, UNCTAD/WEB/ITE/IIA/2008/3, available at http://www.unctad.org/en/ docs/iteiia20083_en.pdf 178. In the early days of the U.S. BIT program, U.S. negotiators were very candid with potential BIT partners in noting that there was no evidence to prove that concluding a BIT would increase investment flows. Vandevelde, supra note 20, at 32. 179. See the discussion in the Introduction to this volume.

a brief history of international investment agreements 29

The same large arbitral awards in favor of investors that may convince investors of the value of the agreements may also cause host countries to conclude that the agreements are too costly. This could be true to the extent that international investment agreements over time are not perceived as attracting increased investment flows. Developing countries may come to see the agreements as poor bargains in which countries surrender portions of their sovereignty and subject themselves to costly arbitrations with investors, without having gained appreciable new investments as a result. The result could be a decision to cease negotiating more agreements or at least to diminish some of the investment protections afforded by the agreements in order to decrease the cost of implementing them. They may also seek to amend those that already have been concluded.180 The impulse to weaken agreements may come from the developed countries as well as the developing countries. The United States, after being named as the respondent in several claims brought under the investment chapter of NAFTA,181 changed its investment agreements to limit certain types of claims. Specifically, it has added language to some agreements that is intended to clarify the scope of the expropriation provision in order to reduce claims for compensation as a result of regulatory actions by the host country.182 It has also added language

180. At least 109 BITs already have been renegotiated, though not necessarily to reduce the protections provided. In fact, a typical purpose has been to strengthen old agreements. UNCTAD, supra note 1, at 3. Some BITs were renegotiated because the prior agreement expired or because one of the parties had joined the European Union, necessitating that BIT obligations be harmonized with EU obligations. 181. As of January 31, 2008, the United States had been named as a respondent in 13 arbitral claims filed under the NAFTA investor-state dispute resolution mechanism. A list of claims may be found at http://www.state.gov/s/l/c3741.htm. 182. For example, Annex 10-D of the free trade agreement with Chile, supra note 149, provides that The Parties confirm their shared understanding that: 1. Article 10.9(1) is intended to reflect customary international law concerning the obligation of States with respect to expropriation. 2. An action or series of actions by a Party cannot constitute an expropriation unless it interferes with a tangible or intangible property right or property interest in an investment. 3. Article 10.9(1) addresses two situations. The first is direct expropriation, where an investment is nationalized or otherwise directly expropriated through formal transfer of title or outright seizure. 4. The second situation addressed by article 10.9(1) is indirect expropriation, where an action or series of actions by a Party has an effect equivalent to direct expropriation without formal transfer of title or outright seizure. (a) The determination of whether an action or series of actions by a Party, in a specific fact situation, constitutes an indirect expropriation, requires a case-by-case, fact-based inquiry that considers, among other factors: (i) the economic impact of the government action, although the fact that an action or series of actions by a Party has an adverse effect on the

30 kenneth j. vandevelde

clarifying that the commitment to “fair and equitable treatment” does no more than incorporate the international minimum standard183 imposed by customary international law.184 Similarly, one of the relatively few recent investment agreements to omit a provision for investor-state arbitration was the 2004 free trade agreement between the United States and Australia.185 Further, Canada has amended its model BIT to include an annex exempting prior agreements from the MFN obligation, thus giving the parties leeway to reduce the level of protection afforded in later agreements.186 Ironically, the disappearance of the ideological division between the developed and developing countries may have made it easier for developed countries to soften the protective coverage of the agreements. In their early history, the BITs had a strong ideological component and were intended to counter the claim of developing countries and Soviet bloc nations that customary international law

economic value of an investment, standing alone, does not establish that an indirect expropriation has occurred; (ii) the extent to which the government action interferes with distinct, rea sonable investment-backed expectations; and (iii) the character of the government action. (b) Except in rare circumstances, nondiscriminatory regulatory actions by a Party that are designed and applied to protect legitimate public welfare objectives, such as public health, safety, and the environment, do not constitute indirect expropriations. 183. For example, article 15.5.(2) of the free trade agreement with Singapore, supra note 148, provides: For greater certainty, paragraph 1 prescribes the customary international law minimum standards of treatment of aliens as the minimum standard of treatment to be afforded to covered investments. The concepts of “fair and equitable treatment” and “full protection and security” do not require treatment in addition to or beyond that which is required by that standard, and do not create additional substantive rights. (a) The obligation in paragraph 1 to provide “fair and equitable treatment” includes the obligation not to deny justice in criminal, civil, or administrative adjudicatory proceedings in accordance with the principle of due process embodied in the principal legal systems of the world; and (b) The obligation in paragraph 1 to provide “full protection and security” requires each Party to provide the level of police protection required under customary international law. 184. See supra text accompanying notes 13–16. 185. United States-Australia Free Trade Agreement, signed May 18, 2004. The parties explained this omission on the ground that the advanced state of the legal system in both countries had diminished the need for such a provision. The agreement, however, does include language to authorize negotiations directed at adding an investor-state dispute resolution clause in the future, if the parties change their view of its necessity. 186. The language may be found at Annex III of the Canadian model BIT, available online at http://www.dfait-maeci.gc.ca/tna-nac/fipa-en.asp.

a brief history of international investment agreements 31

did not require payment of fair market value for expropriated investment.187 As the need to score points in an ideological debate diminishes, developed countries may be more willing to compromise. Indeed, it is easy to foresee cases in which the developed home country of an investor and the developing host country will join forces against the investor in a particular dispute. Some new investment agreements concluded by the United States already provide in certain cases for arbitral tribunals hearing claims brought by investors to consult with the treaty parties concerning the proper interpretation of the agreement.188 The assumption is that the country parties may disagree jointly with an investor’s interpretation of the agreement. In fact, the NAFTA parties already have issued an interpretive notice indicating that the NAFTA‘s guarantee of “fair and equitable treatment” for investment requires no more than customary international law,189 a narrower interpretation of the protection than had been urged by some commentators.190 As has been noted,191 this interpretation is now being incorporated into the text of U.S. agreements. Although large arbitral awards against host countries could dampen the enthusiasm of host countries, including developed countries, for these agreements, they might also signal to investors the value of the agreements and lead to increased investment flows. Because most concerns about the favorability of the investment climate center on the climate in developing countries, investment agreements are least likely to attract investment to developed countries. Any enhanced investment flows as a result of the investment agreements are likely to be in the direction of developing countries. Thus, concerns by developing countries about the issuance of large arbitral awards could be more than offset by the belief that investment agreements had contributed to increased investment flows.

187. See text supra at notes 95–99. 188. Such language appears in the 2004 U.S. model BIT at article 30 and 31. See supra note 144. 189. The NAFTA Free Trade Commission, on July 31, 2001, adopted the following interpretation of NAFTA Chapter 11: 1. Article 1105(1) prescribes the customary international law minimum standard of treatment of aliens as the minimum standard of treatment to be afforded to investments of investors of another Party. 2. The concepts of “fair and equitable treatment” and “full protection and security” do not require treatment in addition to or beyond that which is required by the customary international law minimum standard of treatment of aliens. 3. A determination that there has been a breach of another provision of the NAFTA, or of a separate international agreement, does not establish that there has been a breach of Article 1105(1). 190. See United Nations Conference on Trade and Development, Fair and Equitable Treatment 10–15 (1999): Rudolf Dolzer, Fair and Equitable Treatment: A Key Standard in Investment Treaty Law, 39 Int’l Law. 87 (2005). 191. See supra text accompanying note 144.

32 kenneth j. vandevelde

Further, as the existence of investment agreements becomes more widely known in the investor community and as more countries adopt larger numbers of them, the absence of an investment agreement may be seen as sending a strong negative signal to investors, thus reducing the ability of developing countries to refuse to conclude them or to weaken those that already exist.192 It seems unlikely that the support of developed countries for investment agreements is likely to diminish soon. The interests of the business community in protecting its investment abroad will ensure that capital exporting countries continue to insist that agreements include substantial protections for investments.193 Although developed countries now recognize that such agreements can impose costs on them, the costs at present are largely associated with agreements in which other developed countries are parties.194 Rather than abandoning or significantly weakening the agreements, developed countries may simply adopt the solution employed in the U.S.-Australia free trade agreement and exclude investor-country arbitration from agreements among developed countries. The intermingling of investment and trade provisions in the same agreements inevitably pushes the investment provisions in the direction of liberalization. Trade agreements deal almost entirely with opening markets to trade by eliminating tariffs, nontariff barriers, and discriminatory treatment. As has been noted,195 liberalizing trade increasingly requires liberalizing investment. As countries seek to integrate their economies by concluding free trade agreements, to the extent that they consider investment at all, the notion of liberalizing investment flows will seem to be a critical element. Thus, as investment protections are reined in slightly, investment liberalization provisions are becoming more prevalent. Investment agreements increasingly will be seen not merely as a method of protecting existing investment, but of creating and structuring future international economic relations. The explosion in the number of bilateral agreements relating to investment and with largely very similar provisions has prompted the question whether states should negotiate a multilateral agreement on investment that would simplify the process of negotiation by enabling a state to create a treaty-based 192. It has been theorized that one of the reasons that developing countries conclude investment agreements is simply that other developing countries already have. The inducement to conclude the treaties may not be that the treaties attract investment, but rather that their absence will discourage it. See Zachary Elkins, Andrew T. Guzman, and Beth Simmons, Competing for Capital: The Diffusion of Bilateral Investment Treaties (1996–2000). 193. In the United States, for example, the business community played a major role in prompting the inauguration of the BIT program. Vandevelde, supra note 20, at 20. 194. For example, the United States thus far has been a respondent to claims made only under the NAFTA. It has not been a respondent in any case brought under a BIT. UNCTAD, supra note 1, at 13. 195. See supra text accompanying notes 154–56.

a brief history of international investment agreements 33

investment regime with a large number of parties by acceding to a single agreement. A multilateral agreement that displaced or obviated the need for a large number of bilateral agreements could also reduce the complexity of the international investment regime, which increasingly is characterized by overlapping agreements at the multilateral, regional, and bilateral levels.196 The OECD countries attempted to negotiate a multilateral agreement on investment (MAI) in the 1990s, the idea being that such an agreement would be concluded among the OECD countries initially, but opened to signature by other countries at a later date.197 The negotiations, however, failed.198 It was ironic that the countries that have, perhaps, the greatest consensus among themselves concerning the provisions that should be included in a bilateral investment treaty were unable to agree on a multilateral version of the agreement. The failure of that effort may have been in part the result of that very consensus: that is, because these countries already provide a favorable environment for investment as a matter of national policy, most of the participants had little to gain from the agreement and thus, once negotiations were underway, the focus shifted to that which they would be conceding.199 As the MAI came to be seen as involving mainly concessions rather than gains, the impetus to continue diminished. Meanwhile, various nongovernmental organizations opposed to economic globalization initiated a campaign against the MAI, raising the political price of achieving an agreement.200 Simultaneously, the business community, which similarly perceived that more would be lost than gained by the agreement, failed to provide countervailing support for the MAI.201 Ultimately, the parties simply abandoned the negotiations. It seems unclear how such an agreement might emerge in the near future. The OECD shows no inclination to restart the MAI negotiations. One possible alternative forum is the WTO. At its First Ministerial Meeting in Singapore in 1996, the WTO directed the formation of a Working Group on

196. The complexity of the network of agreements has become a matter of increasing concern. See, e.g., Mary-France Houde and Katia Yannaca-Small, The Relationship Between International Investment Agreements (Organization for Economic Co-operation and Development, Working Papers on International Investment No. 2004/1, 2004), available at http:// www.oecd.org/dataoecd/8/43/31784519.pdf. 197. On the initiation of the OECD negotiations, see Organisation for Economic Cooperation and Development, Towards Multilateral Investment Rules (1996). 198. On the failure of the MAI negotiations, see Edward M. Graham, Fighting the Wrong Enemy: Antiglobal Activists and Multinational Enterprises (2000); United Nations Conference on Trade and Development [UNCTAD], Lessons from the MAI, UN Doc. UNCTAD/ITE/IIT/Misc.22, UN Sales No. E.99.II.D.26 (1999). 199. UNCTAD, supra note 207, at 23–25. 200. Id. 201. Id.

34 kenneth j. vandevelde

Trade and Investment.202 At its Fourth Ministerial Meeting in Doha in 2001, which launched the Doha Round of multilateral trade negotiations, the WTO undertook negotiation of a multilateral investment agreement.203 The WTO General Council decided in August 2004, however, that such an agreement would not be negotiated during the Doha Round.204 Another possible alternative forum is the United Nations Conference on Trade and Development (UNCTAD), which has considerable experience in investment matters.205 Thus far, however, UNCTAD has been given no mandate to conduct negotiations on such an agreement. Some recent trends, moreover, would seem to militate against the conclusion of a multilateral investment agreement. To the extent that countries are beginning to want to link trade and investment, any possible multilateral agreement becomes far more complex and difficult to imagine outside the WTO context, and the WTO is not currently pursuing negotiations in this area. Further, the growing variation in agreements also complicates the negotiations. Balanced against these trends are the fact that nearly every country now accepts the desirability of concluding investment agreements in at least some circumstances, the fact that an increasingly complex framework of agreements seems to call for simplification through a multilateral agreement that could displace many or all of the bilateral agreements, and the growing convergence between the views of developed and developing countries. Regardless, the reality is that little active work toward a multilateral agreement is being performed anywhere, while the number of BITs and bilateral and regional free trade agreements continues to grow.206

conclusion The content of international investment agreements has been, and continues to be, shaped by the political, economic, and legal contexts in which they are negotiated. In the colonial era, when the community of nations was largely limited to the European powers and newly independent countries in the Americas and 202. World Trade Organization, Ministerial Declaration of 13 December 1996, p. 20, WT/ WT/MIN(96)/DEC, 36 I.L.M. 218 (1997). 203. World Trade Organization, Ministerial Declaration of 14 November 2001, pp. 20–22, WT/MIN(01)/DEC/1, 41 I.L.M. 746 (2002). 204. Doha Work Programme, WT/L/579 (Aug. 1, 2004), available at http:// docsonline. wto.org/DDFDocuments/u/WT/L/579.doc. 205. UNCTAD, for example, publishes the widely-cited annual World Investment Report. It has also published the most comprehensive study of bilateral investment treaties yet written, supra note 71, and a series of books on key issues in international investment agreements. For a more complete description of UNCTAD’s activities in this area, see its website at www.unctad.org. 206. UNCTAD, supra note 1, at 10.

a brief history of international investment agreements 35

when international law recognized the use of force as a legitimate means of protecting country interests, customary law or even military force was the primary means of protecting foreign investment. Investment-related provisions did appear in some agreements intended to establish commercial relations between two countries. The international investment regime, however, included relatively few agreements and these agreements provided limited protection for investment. In the postcolonial era, the former colonies achieved political independence and military force was delegitimized as a basis for protecting international economic interests. Fears of economic exploitation on the part of developing countries and an ideological skepticism about the value of free markets, however, caused many developing countries to resist integration into the international economy. At the same time, the hostility of developing countries and Soviet bloc countries toward foreign investment caused developed countries to strengthen their efforts to provide legal protection for foreign investment through international investment agreements. Trade relations were left primarily to the multilateral GATT. International investment agreements, however, remained relatively few in number. They were largely divorced from the concept of liberalization and focused on providing increased protection for foreign investment against political risk. In the global era, the disintegration of the Soviet bloc and the acceptance on the part of developing countries of the value of foreign investment have led to the emergence of an international investment regime that has become virtually universal as nearly every country has concluded at least one investment agreement, the great majority of which contain provisions of remarkable uniformity. Investment agreements continue to protect investment, generally with as much rigor as they did in the postcolonial era, but as in the colonial era they are more frequently being seen as elements of economic integration rather than mere investment protection, sometimes combining trade and investment provisions. In short, history has led us to an increasingly universal international investment regime that seeks to integrate national economies through the removal of barriers to investment flows and through the protection of established investment. This is but another era in the continuing evolution of international investment agreements. The impact of these agreements, particularly the results of arbitrations conducted pursuant to the agreements, in tandem with the evolution in the political and economic context in which these agreements operate, will determine the shape of the next era.

This page intentionally left blank

2. the framework of investment protection: the content of bit s ∗ peter muchlinski introduction This chapter offers an overview of the most common provisions found in International Investment Agreements (IIAs).1 The main focus is on bilateral investment treaties (BITs), as they represent the most common type of IIA. It is beyond the present work to offer a comprehensive and detailed analysis of the principal provisions of IIAs. This has been done elsewhere.2 Nevertheless, a general

∗ This chapter, derived from Peter T Muchlinski Multinational Enterprises and the Law (Oxford, Oxford University Press, 2nd Ed, 2007), is reprinted with permission from Oxford University Press. 1. This chapter is a revised and updated version of Peter T Muchlinski Multinational Enterprises and the Law (Oxford, Oxford University Press, 2nd Ed, 2007) at 674–698 and 701–702. It also draws upon pages 595–596, 621–622, 628–630, 638–639 and 648. 2. See in particular the United Nations Conference on Trade and Development (UNCTAD) Series on issues in international investment agreements published between 1999 and 2004. All volumes have been brought together in Karl P. Sauvant and Joerg Weber, eds., International Investment Agreements: Key Issues Vols I–III (New York and Geneva, United Nations, 2004) and are also available at www.unctad.org/iia. See further, for examples of major IIAs, UNCTAD International Investment Agreements: A Compendium Vols I–XIV also available at www.unctad.org/iia. On BITs see UNCTAD Bilateral Investment Treaties in the Mid 1990s (New York and Geneva, United Nations, 1998) and Bilateral Investment Treaties 1995–2005: Trends in Investment Rulemaking (New York and Geneva, United Nations, 2007); Rudolph Dolzer and Margrete Stevens Bilateral Investment Treaties (The Hague, Matrinus Nijhoff, 1995); J.P.Laviec Protection et Promotion des Investissements: Etude de Droit International Economique (Paris, PUF, 1985); Rudolph Dolzer and Christoph Schreuer Principles of International Investment Law (Oxford, Oxford University Press, 2008) . The texts of BITs are published by ICSID in its periodically updated collection Investment Treaties (Oceana/Oxford University Press). There is an extensive electronic collection available through UNCTAD at www.unctad.org/iia. For analysis, see further G. Sacerdoti “Recent Developments in Bilateral Treaties on Investment Protection” 269 Hague Recueil 251–460 (1997) and A. Akinsanya “International Protection of Foreign Investment in the Third World” 36 ICLQ 58 (1987). For British practice see: F.A.Mann “British Treaties for the Promotion and Protection of Investment” 52 BYIL 241 (1981); Denza and Brooks “Investment Protection Treaties: United Kingdom Practice” 36 ICLQ 908 (1987). For French practice, see Patrick Juillard “Les Conventions Bilaterales d’Investissement Conclues par la France” 2 JDI 274 (1979) and “Les Conventions Bilaterales d’Investissement Conclues par la France Avec les Pays n’Appartenant pas a la Zone Franc” AFDI 760 (1982). For German practice, see Justus Alenfeld Die Investitionsforderungsvertrage der Bundesrepublik Deutschland

38 peter muchlinski

analysis of the main trends in IIA practice is essential for an understanding of their role in the regulation of relations between investors and host countries. The major provisions of BITs and other IIAs are similar, although there are some significant variations in national and regional practice.3 Most BITs, and other IIAs with investment provisions, follow the pattern used in the text below. A. Preamble The preambles to IIAs set down the general objects and purposes of such treaties. Though not legally binding, they may be relevant to the interpretation of the agreement. Usually, such provisions emphasize the desirability of greater economic integration between the contracting parties through improved conditions of investment as laid down in the treaty. There is little variation between the formulations used. A significant variation occurs when the contracting states use the preamble to define specific sectors in which investment is to be protected or promoted. For example, BITs concluded by Switzerland with Sudan and Egypt highlight investment promotion in the fields of “production, commerce, tourism, and technology.”4 Another variation involves reference to respect for particular public policy goals, especially the protection of health, safety, the environment, or consumers, or the promotion of internationally recognized labor rights.5 Preambles are becoming increasingly important as IIAs are

(Frankfurt, Antenaum Verlag, 1970) and Heinrich Klebes Encouragement et Protection des Investissements Prives Dans Les Pays en Developpement - Les Traites Bilateraux de la Republique Federalle Allemagne dans Leur Contexte (Thesis, Strasbourg, 1983). For U.S. practice, see Pattison “The United States-Egypt Bilateral Investment Treaty: A Prototype for Future Negotiation” 16 Cornell Int’l Law Jo. 305 (1983), Kunzer “Developing a Model Bilateral Investment Treaty” 15 Law & Pol’y.Int’l. Bus.273 (1983), Bergman “Bilateral Investment Protection Treaties: An Examination of the Evolution and Significance of the U.S. Prototype Treaty” 16 NYU.J.Int.L & p.1 (1983), Cody “United States Bilateral Investment Treaties: Egypt and Panama” Ga.JIL 491 (1983), Coughlin “The U.S. Bilateral Investment Treaty: An Answer to Performance Requirements” in B.S.Fisher and J.Turner (Eds) Regulating the Multinational Enterprise: National and International Challenges (Praeger, 1983) ch.7. 129–142, and Kenneth J. Vandevelde United States Investment Treaties Policy and Practice (The Hague, Kluwer Law International, 1992). 3. See further UNCTAD Investment Provisions in Economic Integration Agreements (New York and Geneva, United Nations, 2006). 4. UNCTAD 1998 above n.2 at 31. 5. See, for example, the Preambles to the Korea-Trinidad and Tobago BIT 2002 and the U.S.-Uruguay BIT 2004 both in UNCTAD 2007 n.2 above at 5. The U.S.-Uruguay Treaty is also reproduced in 44 ILM 268 (2005) also available at http://ustr.gov/assets/ World_Regions?Americas/South_America/asset_upload_file440_6728.pdf

the framework of investment protection: the content of bits 39

subjected to greater public scrutiny for conformity with goals other than investment promotion and protection.6 B. Provisions Defining the Scope of Application of the Treaty IIAs contain provisions that define the scope of application of the treaty by reference to subject matter, covered persons and entities, territory, and temporal effect. It is in the negotiation of these provisions that the host country can influence the effect of the treaty upon its economy. For example, the government may restrict the treaty’s impact to specific economic sectors.7 Some agreements use a general scope of application clause, though such clauses do not obviate the need for definition clauses.8 1. Subject matter covered a. Definition of investments to which the IIA applies. IIAs usually commence with a provision that defines the investments of the contracting parties that are covered by the treaty. More recent agreements tend to favor broad, asset-based definitions that include not only physical assets, equity, and choses in action, but also incorporate intellectual property rights and contractual concessions.9 The aim is to ensure sufficient flexibility to encompass not only equity, but also nonequity investments, and to allow for the evolution of new forms of investment between the parties. Furthermore, the notion of direct investment is not taken as the starting point.10 Nonetheless, certain treaties concluded by Germany contain a clause which makes clear that the general categories of investment mentioned in the standard definition clause are to be linked to a direct investment.11 In addition to such definitions, some agreements contain an enterprise-based definition focusing on “the business enterprise” or the “controlling interests in a business enterprise,”12 while other agreements take a “transaction-based” definition.

6. See UNCTAD 2007 above n.2 at 3. 7. See further UNCTAD 1998 above n.2 at 32–37 and UNCTAD Key Issues Vol. I. n.2 above at 119–122. 8. See UNCTAD 2007 above n.2 at 5. 9. See, for example, ASEAN Agreement for the Promotion and Protection of Investments, Article 1(3) in UNCTAD Compendium Vol II above n.2 at 294. 10. Laviec above n.2 at 31. 11. See, e.g., Germany-Israel BIT 1976 Article 1(1) (a) where the general clause is preceded by a clause stating that the term investment means: “(i) investment in an enterprise involving active participation therein and the acquisition of assets ancillary thereto, or (ii) the enterprise or assets acquired as a result of such an investment.” UNCTAD 1998 above n.2 at 33. 12. See, for example, the Canada-U.S. Free Trade Agreement 1988 cited in UNCTAD Key Issues Vol.I. above n.2 at 125. Canada has now adopted a “closed-list” definition of investment, which, rather than containing a chapeau defining the concept of investment, has a long but finite list of tangible and intangible assets to be covered by the agreement.

40 peter muchlinski

For example, the Organisation of Economic Cooperation and Development (OECD) Code of Liberalisation of Capital Movements does not define “investment” or “capital” as such, but contains a list of covered transactions in Annex A, including direct investment.13 Limitations may be made to the definition of “investment.” Thus some agreements exclude portfolio investment,14 on the grounds that it is less stable than direct investment and so should not be given equal protection, or limit protection to investments permitted under the law of the host country. This formulation is common in Chinese and Association of Southeast Asian Nations (ASEAN) agreements,15 in investments of a certain size, and in investments in specific sectors.16 b. The admission of investments. Under general international law, countries have the unlimited right to exclude foreign nationals and companies from entering their territory. There is no international standard requiring countries to adopt an “open door” to inward direct investment.17 On the other hand, countries are free to set the limits of permissible entry in their national laws as they see fit. Thus, countries are free to agree to provisions in IIAs that secure access to their territory for investors from another contracting party. In this regard, two main models of agreements are emerging: a “controlled entry” model that reserves the right of the host country to regulate the entry of foreign investments into its territory, and a “full liberalization” model that extends the non-discrimination standard (both national treatment and most-favored-nation (MFN) treatment) in the agreement to the pre-entry stage of the investment.18 Such an approach is

This approach evolved out of the “enterprise-based” definition. See Canada Model Foreign Investment Protection Agreement (FIPA) 2004 Article 1 at http://ita.law.uvic.ca/ documents/Canadian2004-FIPA-model-en.pdf or UNCTAD 2007 above n.2 at 10–11. 13. UNCTAD Key Issues above n.2 at 118. 14. See, for example, the Denmark-Poland BIT Article 1(1)(b) where investment refers to “all investments in companies made for the purpose of establishing lasting economic relations between the investor and the company and giving the investor the possibility of exercising significant influence on the management of the company concerned” quoted in ibid. at 123. 15. Ibid. at 122–123. 16. Ibid. at 125. 17. Laviec above n.2 at 77. See further Muchlinski above n.1 chs.5 and 6. 18. UNCTAD, in its study Admission and Establishment (New York and Geneva, United Nations, 1999 also in UNCTAD Key Issues Vol.I. above n.2 at 142–160), identifies five models of admission and establishment clauses: investment control, which corresponds to the controlled entry approach in the text; selective liberalization, based on the GATs type “opt-in” sectoral liberalization approach; the regional industrialization program approach, based on certain developing country regional integration agreements; the mutual national treatment approach of the EU; and the combined national treatment MFN approach, which corresponds to the full liberalization approach in the text. See too Thomas Pollan Legal Framework for the Admission of FDI (Utrecht, Eleven International Publishing, 2006) Ch.4. who uses a modified version of the UNCTAD classification.

the framework of investment protection: the content of bits 41

particularly favored in the practice of NAFTA19 and in the bilateral treaty practice of the United States and Canada. The majority of BITs follow a “controlled entry” approach. Therefore, the application of the treaty to an investment is made conditional on its being approved in accordance with the laws and regulations of the host country.20 On the other hand, the U.S.-Uruguay BIT of October 25, 2004, illustrates the “full liberalization” approach. By Article 3(1): Each Party shall accord to investors of the other Party treatment no less favourable than it accords, in like circumstances, to its own investors with respect to the establishment, acquisition, expansion, management, conduct, operation, and sale or other disposition of investments in its territory.21 This is followed by Article 3(2), which extends the same protection to investments, and Articles 4(1) and 4(2) which cover, respectively, the MFN protection of investors and of investments. These provisions make entry to the host country subject to the principle of non-discrimination, and, to that extent, represent a restriction on the host country’s sovereign power to regulate the entry of foreign investors. Nevertheless, the application of this principle is subject to the right of the host country to exclude certain sectors from foreign investment.22 Therefore, the U.S. model accepts restrictions on entry, provided they are applied in a manner that does not discriminate against U.S. investors.23 Under the Russian Federation-U.S. BIT, the contracting states agreed that, for a period of five years, the Russian Federation would be able to require permission for large-scale investments that exceeded the threshold amount in the Russian Federation Law on Foreign Investments of July 4, 1991, provided that this power was not

19. See NAFTA Articles 1102–1108 in UNCTAD Compendium Vol III above n.2 at 73. 20. See examples cited in UNCTAD 1998 above n.2 at 46–47 and UNCTAD 2007 above n.2 at 21–22. See too Asian African Legal Consultative Committee (AALCC) Model BIT “Model A” Art.3 which states inter alia that, “[e]ach Contracting Party shall determine the mode and manner in which investments are to be received in its territory”. “Model B” is more restrictive of investor’s rights of entry in that its Art.3 makes the screening of investment proposals by the host country a mandatory treaty requirement. See 23 ILM 237 (1984) or UNCTAD Compendium Vol.III above n.2 at 115. 21. 44 ILM 268 (2005) at 217 also available at http://ustr.gov/assets/World_Regions? Americas/South_America/asset_upload_file440_6728.pdf 22. See ibid. Article 14(2). 23. See Pattison’s discussion of the U.S.-Egypt BIT in this respect: n.2 above at 318–319. Pattison is critical of the broad exclusion of industries from the treaty that, in his opinion, created a substantial void in the protection offered. The current version of this treaty retains these exceptions: U.S.-Egypt BIT Supplementary Protocol of 11 March 1986 para.3 at http:// www.unctad.org/sections/dite/iia/docs/bits/us_egypt.pdf

42 peter muchlinski

used to limit competition or to discourage investment by U.S. companies and nationals.24 c. Applicability to investments made prior to the conclusion of the treaty. Most BITs extend their protection to investments already made by nationals of the contracting parties prior to the conclusion of the treaty, however, not all host countries accept such a clause. Thus, some treaties extend to prior investments only on condition that the host country approves a special request to that effect in each case, while others only apply to investments made at a specified date prior to the conclusion of the treaty. Some treaties contain a blanket exclusion of investments made prior to the entry into force of the treaty.25 2. Covered persons and entities The protection offered by IIAs is limited to investors who invest in the territory of the host contracting country and who possess a link of nationality with the home contracting country. Consequently, the IIA must define both the persons and the corporations that are to be treated as nationals of the other contracting party for the purposes of the treaty. The attribution of nationality to natural persons is not generally a problem. Most BITs simply refer to the country’s citizenship laws as governing the matter.26 However, further criteria might be introduced, such as a combined requirement of citizenship and residence.27 In British practice, the right of residence has been used as an alternative to British citizenship in defining British nationals for the purpose of a BIT.28 By contrast the definition of nationality in the case of corporations is more complex. The nationality of a company may be determined by reference to one or more of the following criteria: the place of incorporation, the location of the registered office or seat of the company, or by reference to the nationality of the controlling interest in the company. Each of these criteria has been used, alone or in combination, in BITs.29 Furthermore, the term “companies” has not been

24. U.S.-Russian Federation BIT 17 June 1992 Protocol para.4(a). 31 ILM 794 at 810 (1992). 25. See UNCTAD 1998 above n.2 at 42–43 and UNCTAD 2007 above n.2 at 19–20 for examples of each type of provision. 26. See, e.g., the U.S.-Uruguay BIT 2004 above n.5 Article I “National” of a Party means a natural person who is a national of a Party under its applicable law; “investor of a Party” means “a Party or state enterprise thereof, or a national or an enterprise of a Party, that attempts to make, is making, or has made an investment in the territory of the other Party”; “enterprise of a Party” means “an enterprise constituted or organized under the law of a Party, and a branch located in the territory of a Party and carrying out business activities there.” See too examples in UNCTAD 1998 ibid. Table III.3, at 40. 27. See, e.g., Germany-Israel BIT Article I(3)(b): UNCTAD ibid. at 38; Armenia-Canada BIT 1997 in UNCTAD 2007 above n.2 at 13. 28. See UK-Philippines BIT in Mann above n.2 Article I(3)(b). 29. For examples see: UNCTAD 1998 above n.2 at 39–41, UNCTAD 2007 above n.2 at 15–17.

the framework of investment protection: the content of bits 43

restricted to legally incorporated entities alone; often it includes references to partnerships and other forms of business associations. The most significant issue in the context of the present study is the extent to which IIAs cover the case of an investment made by a parent company from the home contracting country through a subsidiary company incorporated under the laws of the host contracting country. Can the agreement in question extend its protection to the local subsidiary of the foreign parent company notwithstanding its possession of host country nationality? The provisions of BITs have approached this problem in a number of ways.30 First, some treaties are silent on the matter. This is the case with the majority of early BITs concluded by France and Belgium, although some more recent French agreements now include a “control” criterion.31 Similarly, UK practice has been not to use a test of control in determining the nationality of a locally incorporated subsidiary.32 The legal effect of such an approach is to remove the protection of the treaty from the locally incorporated subsidiary. Nonetheless, the provision defining the nationality of the corporation should be read alongside the definition of investments covered by the treaty. This may be broad enough to cover investments carried out through foreign-owned local companies. Thus the U.S.-Uruguay BIT of 2004 defines an “enterprise” as, “any entity constituted or organized under applicable law, whether or not for profit, and whether privately or governmentally owned or controlled, including a corporation, trust, partnership, sole proprietorship, joint venture, association, or similar organization; and a branch of an enterprise.”33 Without more it would be unclear from this definition whether locally incorporated foreign-owned subsidiaries were within the scope of the treaty. However the term “investment” is defined as, “every asset that an investor owns or controls, directly or indirectly, that has the characteristics of an investment, including such characteristics as the commitment of capital or other resources, the expectation of profit, or the assumption of risk.”34 This is sufficient to protect the investment of a national or company of the home contracting country that takes the form of a company established under the law of the host contracting country. In such a case, the application of the treaty is not based on the nationality of the subsidiary, but on the nationality of its controlling interest. 30. See Laviec above n.2 at 43–48 on which the following paragraphs are based. 31. See, e.g., France-Nepal BIT 1984 Article I in ICSID Investment Treaties Vol.2. above n.2. 32. See, e.g., UK-Philippines BIT above n.28 Article I(4): “The term ‘company’ of a Contracting Party shall mean a corporation, partnership or other association, incorporated or constituted and actually doing business under the laws in force in any part of the territory of that Contracting Party wherein a place of effective management is situated.” According to Dr. Mann, under this definition, “[a] Philippine company cannot claim the protection of the Agreement even if all the shares are owned by a British national”: ibid. at 242. 33. U.S.-Uruguay BIT 2004 above n.5 Article I. 34. Ibid.

44 peter muchlinski

A second approach to this issue is for the treaty to contain a definition of corporate nationality that includes a test based on controlling interests as an exclusive or an alternative test. Treaties using this approach include those concluded by the Netherlands, Sweden and Switzerland.35 This ensures that locally incorporated foreign-owned subsidiaries are covered by the treaty. However, the control test is exceptional. It makes more of the links with the home country, based on capital, than of the legal links of the subsidiary company with the host country, by reason of its incorporation there. In this, the test does not represent the existing state of international law.36 A third approach is to infer the nationality of the home country to the subsidiary incorporated in the host country by means of an agreement between the subsidiary and the host country.37 It is reminiscent of the approach taken under Article 25(2)(b) of the Washington Convention in relation to the jurisdiction of the International Centre for Settlement of Investment Disputes (ICSID) over disputes between the host contracting country and a local subsidiary owned by investors possessing the nationality of another contracting country.38 The fourth approach is one of specifically extending the protection of the BIT to subsidiaries owned by or under the effective control of nationals or companies from the other contracting country. For example, the German model treaty extends the national treatment and MFN standards to, “investments in [the territory of a Contracting Party] owned or controlled by nationals or companies of the other Contracting Party.”39 According to Laviec, in such a case, the protection of the treaty is extended to the local subsidiary without undermining the test 35. UNCTAD 1998 above n.2 at 39; Laviec above n.2 at 45. 36. Laviec ibid. See further the Barcelona Traction Case ICJ Reports (1970) 3, which rejected a test of corporate nationality based on the nationality of the controlling interest, favoring a test based on the nationality of the place of incorporation or principal seat of management. 37. See, e.g., Netherlands-Kenya Agreement on Economic Co-operation 11 June 1979 Tractatenblad 1970 No.166 Art XIV: “For the purpose of the present Agreement: . . . (b) a legal person which is lawfully established in the territory of a Contracting Party shall be a national of that Contracting Party in conformity with its legislation; except where any such legal person, established in the territory of a Contracting Party is controlled by a national or nationals of the other Contracting Party and it has been agreed between the legal person, and the first mentioned Contracting Party that it should be treated for the purposes of this Agreement as a national of the other Contracting Party.” 38. The Washington Convention on Settlement of Investment Disputes Between States and Nationals of Other States 1965: 4 ILM 524 (l965); 575 UNTS 159. For detailed analyses see Christoph Schreuer The ICSID Convention: A Commentary (Cambridge, Cambridge University Press, 2001) and Emmanuel Gaillard La Jurisprudence du CIRDI (Paris, Editions Pedone, 2004). For further references to the ICSID Convention see ICSID ICSID Bibliography (Doc.ICSID/13/Rev.2, March 1, 1992, and subsequent revisions available at www.worldbank.org ). 39. Germany-St Lucia BIT 1985 Art.3(1) ICSID Investment Treaties Vol.2. above n.2.

the framework of investment protection: the content of bits 45

of corporate nationality as established under international law. The link with home country nationals is established on the basis of their economic interests in the subsidiary, without the need to designate the subsidiary as possessing home country nationality in opposition to its nationality of incorporation.40 Finally, some BITs provide limited protection for shareholders from the home contracting country in companies incorporated in the host contracting country, even though they do not extend their protection to locally incorporated foreign owned companies. Thus in the UK-Philippines BIT,41 Article V(2) extends the protection against unlawful expropriation to cases where the assets of a company incorporated under the law of the expropriating country are taken, and in which nationals or companies of the other contracting country own shares. Furthermore, by Article X(2) companies incorporated under the law of one contracting party, in which a majority of the shares are owned by nationals or companies of the other contracting party, shall be treated, for the purposes of disputes under the Washington Convention, as companies of the other contracting party, in accordance with Article 25(2)(b) of that Convention. These provisions represent departures from the law on the diplomatic protection of shareholders, as laid down in the Barcelona Traction case.42 Consequently, these provisions represent specific treaty-based standards of treatment and not general principles of international law. 3. Territorial application It is unnecessary to go into detail on these provisions. The territorial scope of the treaty will be defined by its terms. Usually, the treaty will apply throughout the territory of all the contracting countries, however, it may contain territorial extension or restriction clauses. A particular problem that has been encountered concerns the applicability of the treaty to maritime exclusive economic zones. This may be particularly important in relation to the treatment of offshore natural resource concessions.43 4. Temporal application The entry into force of a treaty occurs on the date designated by the parties for this purpose. The usual practice is that the IIA enters force either upon the exchange of instruments of approval or ratification, or upon reciprocal notification that the relevant constitutional requirements of each contracting country have been fulfilled, or at a set date after such notification has been given, usually after one month.44 The termination of the treaty is usually provided for in that the treaty is designated to last for a period of years. The usual period is ten years, although shorter terms of five years and longer terms of up to twenty years have been agreed upon. Some treaties are open-ended, providing for termination after the giving of the required period of notice. In addition, there may be clauses that extend the operation of the protection granted by the treaty 40. Laviec above n.2 at 46–47. 41. Above n.28. 42. See n.36 above. 43. See further UNCTAD 1998 above n.2 at 41–42, UNCTAD 2007 above n.2 at 17–19. 44. UNCTAD 1998 ibid. at 43–44.

46 peter muchlinski

beyond the date of termination in cases where individual investments, entered into while the treaty was in force, continue after the date of termination.45 C. Standards of Treatment The core provisions of BITs concern the standards of treatment to be applied to investments made by nationals and companies of one contracting party in the territory of the other. The applicable standards can be classified into general standards, encompassing both international minimum standards of treatment and standards evolved in earlier Friendship Commerce and Navigation (FCN) treaties,46 and specific standards applicable to particular incidents of investment activity, such as the transfer of funds and taxation. 1. General standards of treatment General standards of treatment have been divided into those recognized by general international law and those that have evolved out of commercial treaty practice.47 The principal general standards of treatment referred to in IIAs are those of fair and equitable treatment, national treatment, and the MFN standard. Many treaties also contain clauses protecting the observance of obligations entered into between the investor and the host country. a. Fair and equitable treatment. This concept is not normally precisely defined in IIAs. Usually it is simply stated either alone or in conjunction with a reference to international law. It has been suggested that fair and equitable treatment

45. Ibid. at 44 and UNCTAD 2007 above n.2 at 20–21. 46. FCN treaties were concluded primarily by the United States with its advanced trading partners, though other advanced countries also concluded such treaties. The first treaty of amity and commerce signed by the United States was a treaty of 6 February 1778 with France: see Wilson “Postwar Commercial Treaties of the United States” 43 AJIL 262 at 277 (1949). Such treaties included provisions on the protection of corporate investments, based on the national treatment standard. See, for example, Art.III(2) of the U.S.Italy FCN Treaty of 2 February 1948; Art.III(3) of the U.S.-China FCN of 4 November 1946: see Wilson at 265. Art.III(2) of the U.S.-Italy Treaty was the subject of litigation before the ICJ in the Case Concerning Elettronica Sicula S.p.A. (ELSI) (United States v. Italy) Judgment 20 July 1989 ICJ Reports (1989) 15. FCN treaties are no longer being negotiated, although such treaties remain in force for a number of countries. The principal reason for their demise is that such treaties contain a breadth of provisions dealing not only with commercial matters, but also with the protection of the rights of individuals and with general good relations between the signatory states. They do not have a specific business orientation., nor do they deal in adequate detail with the most pressing problems likely to be faced by foreign investors, such as funds transfers, unfair treatment, or dispute settlement. Furthermore, FCN treaties were concluded primarily between the economically advanced states. Their purpose was to further good relations between them. Such treaties were not, therefore, suited to the task assigned to BITs, namely, the control of less-developed, capital-importing states in their treatment of foreign investors. 47. See generally Laviec above n.2 ch III.

the framework of investment protection: the content of bits 47

represents a classical international law standard which embodies international minimum standards of treatment.48 According to Laviec, a reference to fair and equitable treatment should not be read as a reference to international minimum standards. If the intention is to assimilate the two concepts, this should be made explicit in the text.49 Otherwise, the fair and equitable treatment standard should stand on its own.50 Indeed, the U.S.-Uruguay BIT of 2004 contains, in Article 5(1) thereof, a provision that mentions fair and equitable treatment alongside international minimum standards.51 Article 5(2) goes further and expressly states that, “[f]or greater certainty, paragraph 1 prescribes the customary international law minimum standard of treatment of aliens as the minimum standard of treatment to be afforded to covered investments.” It continues by saying that, “[t]he concepts of ‘fair and equitable treatment’ and ‘full protection and security’ do not require treatment in addition to or beyond that which is required by that standard and do not create additional substantive rights.” This is a direct response to the suggestion made in NAFTA arbitral decisions that this standard is additive to the international law standard.52 It also makes clear that the agreement applies international minimum standards of treatment as part of the fair and equitable standard. The practice of capital-importing countries concerning consent to the fair and equitable treatment standard is uneven. It is not included in the Asian-African Legal Consultative Committee (AALCC) model treaties,53 nor has it been accepted by certain Asian and African countries in their negotiations.54 On the other hand, it does appear in some agreements with Central and South American countries.55 Furthermore, China has accepted the standard in some, but not all,

48. See UNCTAD 1998 above n.2 at 53, UNCTAD 2007 above n.2 at 28. See further UNCTAD Key Issues Vol I above n.2 at 209–234. 49. For example, agreements concluded by France relate fair and equitable treatment to the general principles of international law and require that the rights of states must not be impaired de jure or de facto (Article 3: see French Model BIT in UNCTAD 1998 ibid. at 254 and see too France-Mexico BIT 1998 Article 4 and France-Uganda BIT 2002 Article 3 both in UNCTAD 2007 above n.2 at 31). Agreements concluded by Belgium-Luxembourg also normally specify that fair and equitable treatment, “may in no case be less favourable than that recognized by international law.” See Article 3 Belgium-Luxembourg Economic UnionMalaysia BIT 8 February 1982 Moniteur Belge 30 March 1982. 50..Laviec above n.2 at 94. 51. Above n.5. 52. See further Muchlinski above n.1 at 637–638. 53. Above n.21. 54. For example, most, though not all, treaties signed by Rwanda, Pakistan, Singapore, and Saudi Arabia: UNCTAD 1998 above n.2 at 54. 55. Ibid.

48 peter muchlinski

of the agreements.56 The standard is accepted in the ASEAN multilateral investment framework.57 To date, arbitral decisions covering the interpretation of the fair and equitable treatment standard have concentrated on the responsibilities of the host country in its conduct towards the investor.58 As a result, certain elements of an emergent standard of review of administrative action are taking shape. This standard of review reflects contemporary approaches to good governance. Thus, tribunals have noted that the original customary international law standard, with its emphasis on outrageous mistreatment of the alien, is no longer sufficient.59 Instead, the correct question is, whether, at an international level and having regard to generally accepted standards of administration of justice, a tribunal can conclude in the light of all the available facts that the impugned decision was clearly improper and 56. See Belgium-Luxembourg Economic Union-People’s Republic of China (PRC) BIT 4 June 1984 Art.3(1): accords “equitable treatment” to direct or indirect investments: 24 ILM 537 at p. 540 (1985). This is supplemented by a reference to treatment not less favorable than that provided for in the, “generally recognised principles and rules of international law adopted by the Contracting Parties.” ibid. Protocol Art.7 at ILM p. 549. France-PRC BIT 30 May 1984 Art 3(1): accords “fair and equitable treatment to the investments made by investors of the other party” ibid. ILM p. 552. See too Australia-PRC BIT 11 July 1988 Art.3(a): 28 ILM 121 at 127 (1989). Not all BITs concluded by China contain the equitable treatment standard. Thus the Japan-PRC BIT of 27 August 1988 contains only the national treatment and MFN standards: 28 ILM 575 (1989). 57. Above note 112 Art.III(2). See too the 1998 Framework Agreement on the ASEAN Investment Area, Article 5(a) which requires Member States to, “ensure that all programmes are undertaken on a fair and mutually beneficial basis” and (b) which requires transparency: UNCTAD Compendium Vol.VI above n.2 at 230. 58. See further UNCTAD Fair and Equitable Treatment Series on issues in international investment agreements (New York and Geneva, United Nations, 1999) also in UNCTAD Key Issues Vol I above n.2 at 209–234; Steven Vasciannie “The Fair and Equitable Treatment Standard in International Investment Law and Practice” 70 BYIL 99 (1999); OECD Fair and Equitable Treatment Standard in International Investment Law Working Papers on International Investment Law No. 2004/3 (Paris, OECD, September 2004) also reproduced as Ch.3 of OECD International Investment Law: A Changing Landscape (Paris, OECD, 2005); Patrick G. Foy and Robert J. Deane “Foreign Investment Protection under Investment treaties: Recent Developments under Chapter 11 of the North American Free Trade Agreement” 16 ICSID Rev-FILJ 299 (2001); J. C. Thomas “Reflections on Article 1105 of NAFTA: History, State Practice and the Influence of Commentators” 17 ICSID Rev-FILJ 21 (2002); Patrick Dumberry “The Quest to define ‘Fair and Equitable treatment’ for Investors under International Law – The Case of the NAFTA Chapter 11 Pope and Talbot Awards” 3 JWI 657 (2002); Christoph Schreuer “Fair and Equitable Treatment in Arbitral Practice” 6 JWIT 357 (2005); Rudolf Dolzer “Fair and Equitable Treatment: A Key Standard in Investment Treaties” 39 Int’L. Law. 87 (2005). 59. For example, in USA (L.F. Neer) v. Mexico (Neer Claim) (1927) AJIL 555 at 556 such treatment was defined as treatment amounting to an “outrage, to bad faith, to wilful neglect of duty, or to an insufficiency of governmental action so far short of international

the framework of investment protection: the content of bits 49

discreditable, with the result that the investment has been subjected to unfair and inequitable treatment.60 In the light of this contemporary perspective, it is now reasonably well-settled that the standard requires a particular approach to governance, on the part of the host country, that is encapsulated in the obligations to act in a consistent manner, free from ambiguity and in total transparency, without arbitrariness and in accordance with the principle of good faith.61 In addition, investors can expect due process in the handling of their claims,62 and to have the host authorities act in a manner that is non-discriminatory and proportionate to the policy aims involved.63 These will include the need to observe the goal of creating favorable investment conditions and the observance of the legitimate commercial expectations of the investor.64 It is not necessary to show bad faith or a deliberate intention to injure the investment to show a breach of the standard, though the presence of such factors may aggravate the breach.65 On the other hand, the

standards that every reasonable and impartial man would readily recognize its insufficiency.” See too International Thunderbird Gaming Corporation v. Mexico NAFTA Arbitration under UNCITRAL Rules award of 26 January 2006 (available at www.asil. org/ilib) where the majority of the tribunal suggests that under NAFTA, and the international minimum standard embodied in its fair and equitable treatment provision, Article 1105, only a gross denial of justice or manifest arbitrariness would fall below acceptable international standards (at para.194). This would appear to be more restrictive of the standard than other recent awards. See also the Separate Opinion of Professor Thomas Waelde for an extensive discussion of the scope of the doctrine of legitimate expectations in relation to the standard and its application to the particular facts of the case (available at www.asil.org/ilib). For comment see Stephen Fietta “The ‘Legitimate Expectations’ Principle under Article 1105 NAFTA – International Thunderbird Gaming Corporation v. The United Mexican States” 7 JWIT 423 (2006) who considers that legitimate expectations may be becoming a self-standing basis of claim under the fair and equitable treatment standard. 60. Mondev International Ltd v. U.S. ICSID Case No Arb (AF)/99/2 Award of 11 October 2002, 42 ILM 85 (2003) at para.127. 61. See Tecmed v. Mexico ICSID Case No Arb (AF)/00/2 Award of 29 May 2003, 43 ILM 133 (2004) at para.154–155. 62. See for example Loewen v. U.S. ICSID Case No Arb (AF)/98/3 Award of 26 June 2003, 42 ILM 811 (2003). 63. See Loewen v. United Sates ibid.; Waste Management Inc v. Mexico ICSID Case No. ARB (AF)/00/3 award of 30 April 2004: 43 ILM 967 (2004) at para.98; MTD Equity v Chile ICSID Case No Arb/01/7 Award of 25 May 2004, 44 ILM 91 (2005) at para.109; Eureko v. Poland Partial Award 19 August 2005 at paras.231–235. 64. Tecmed v. Mexico above n.61 at para.156–157; Saluka v. Czech Republic UNCITRAL Rules Arbitration, Permanent Court of Arbitration, award of 17 March 2006 at paras. 302–309 (available at www.ita.law.uvic.ca); Thunderbird Gaming Corporation above n.59. 65. Azurix v. Argentina ICSID Case No.ARB/01/12 award of 14 July 2006 at para.372 (available at www.ita.law.uvic.ca).

50 peter muchlinski

standard is case specific and requires a flexible approach, given that it offers a general point of departure in formulating an argument that the foreign investor has not been well treated by reason of discriminatory or other unfair measures that have been taken against the investor’s interests.66 Such case specific flexibility may require an examination not only of governmental but also of investor conduct in a given case.67 b. National treatment. The national treatment standard requires that foreign investors should receive treatment no less favorable than that accorded to nationals of the host country engaged in similar business activity.68 It seeks to ensure equality of competitive conditions between foreign investors and domestic investors in a like situation. Following from the above, the basis of a claim brought under the national treatment standard lies in the allegation that the investor and/or their investment have been treated less favorably than a comparable domestic investor/ investment. The essential elements of the claim are as follows. The aggrieved foreign investor must show, first, that there exist domestic investors in the same economic or business sector; second, that the foreign investor and/or investment is being treated less favorably than their domestic counterparts and, third, that the design and nature of the measures that lead to this difference in treatment have a discriminatory effect. This may be an express intention of the measure, or a factual effect thereof. Thus, both de jure and de facto discrimination will be sufficient to found the claim. The burden of proof is usually that of establishing a presumption and a prima facie case that the claimant has been treated in a different and less favorable manner than their domestic counterpart(s), whereupon the burden shifts to the respondent country to show that no such effect has arisen.69 66. Mondev International Ltd v. United States above n.60 at para.118 and Waste Management above n.63 at para.99. 67. See further Peter Muchlinski “‘Caveat Investor’? The Relevance of the Conduct of the Investor under the Fair and Equitable Treatment Standard” 55 ICLQ 527 (2006). 68. See for example the OECD Declaration on International Investment and Multinational Enterprises 27 June 2000, “National Treatment” (Paris, OECD, 2000) at p.11–12 or at http:// www.oecd.org/dataoecd/56/36/1922428.pdf. See further OECD National Treatment for Foreign-Controlled Enterprises (Paris, OECD, 1985, revised 1993 and 2005). For further examples of national treatment provisions see UNCTAD National Treatment Series on issues in international investment agreements (New York and Geneva, United Nations, 1999) or in UNCTAD Key Issues Vol. I above n.2 at 161–189 on which this section is based. The author prepared the original manuscript for the UNCTAD paper. See too UNCTAD 2007 above n.2 at 33–38. 69. See Feldman v. Mexico ICSID Case No. ARB(AF)/99/1 award 16 December 2002 available on www.naftaclaims.com or 18 ICSID Rev-FILJ 488 (2003) at paras.176–178. For a full analysis of the main cases on national treatment, which have arisen mainly in the context of NAFTA claims, see Muchlinski above n.1 at 622–626. The main cases are S.D.Myers v. Canada UNCITRAL Award of 12 November 2000 available at www.naftaclaims.com or 40 ILM 1408 (2001); Pope and Talbot v. Canada Award on the Merits of Phase 2, 10 April 2001 available on www.naftacliams.com; ADF Group Inc v. United States

the framework of investment protection: the content of bits 51

National treatment provisions are drafted according to a common basic pattern, but with some significant variations.70 Some agreements do not refer expressly to national treatment, so as to avoid extending preferential treatment reserved for national enterprises to their foreign counterparts. This was the policy of China in its early agreements, though references to national treatment appear in its more recent agreements usually with the qualification that such treatment will be accorded subject to national laws and regulations. More recent Chinese agreements contain an unreserved national treatment provision.71 On the other hand, the majority of IIAs do contain a national treatment clause. This covers a number of specific matters. First the scope of application will be defined. As noted above, agreements such as NAFTA and the U.S. and Canadian Model BITs apply the standard both to pre- and post-entry treatment, while BITs concluded by European capitalexporting countries tend to apply only to post-entry treatment. By contrast, the World Trade Organization (WTO) General Agreement on Trade in Services (GATS) has a hybrid approach allowing national treatment under Article XVII only in those sectors that the members have included in their schedules under Article XVI, and subject to any restrictions reserved by the member in question in their schedule to Article XVII. In addition some agreements, notably again the North American regional and bilateral agreements, specify that national treatment will apply to the political subdivisions of the contracting parties, however, silence on this point does not mean that such subdivisions are excluded from the standard. It is implicit in the standard itself that it applies to all forms of discriminatory treatment unless expressly excluded. The national treatment clause will then determine the factual situations to which it applies. Some agreements refer to “the same” or “identical” circumstances to limit the application of the standard, or to designate specific sectors or activities to which it applies.72 The greatest subject matter coverage will be available under clauses that refer to “like circumstances” or “similar circumstances.” Here the factual comparison to be made does not need to be as exact as under the more precise formulations. This comparison should be limited to genuine likeness, as where the domestic and foreign investors both operate in the same economic sector.73 The widest scope for comparison will come with clauses that

of America ICSID Case No.ARB(AF)/00/1 award of 9 January 2003 available at www. naftaclaims.com or 18 ICSID Rev-FILJ 195 (2003); Feldman v. Mexico previous note and Methanex v. United States of America award of 3 August 2005 available at http://www.state. gov/documents/organization/51052.pdf or 44 ILM 1345 (2005) 70. See generally UNCTAD Key Issues Vol. I. above n.68. 71. Ibid. at 187 n.6. 72. See ibid. at 171 for examples. 73. On the other hand, the London Court of International Arbitration in Occidental Exploration and Production Co v. Ecuador Administered Case No. UN 3467 award of 1 July

52 peter muchlinski

do not refer to any factual comparisons, but merely state the standard. This approach is favored among Chilean, French, German, Swiss, and United Kingdom BITs, though the latter also retains a functional delimitation formula in relation to investors.74 The national treatment clause will then define the standard of treatment itself. While some agreements, notably earlier regional economic integration agreements between developing countries, included a “same” or “as favorable as” standard, allowing for treatment no better than that accorded to domestic investors, the majority of agreements use the “no less favorable” standard. This permits treatment more favorable than that accorded to domestic investors, where the host country so chooses, and for the application of international minimum standards of treatment where national standards fall below this. In addition, some agreements extend the standard to both de jure and de facto discrimination.75 Where the provision is silent on the matter, it is possible to interpret the standard to include both types of discrimination, as is the case in relation to NAFTA. Finally, the national treatment provision may be a “stand alone” provision or be combined in one clause with the most-favored-nation standard and/or fair and equitable treatment. It may also be subject to exceptions. These may be of a general kind such as national security, public health and public morality exceptions, subject-specific exceptions, industry-specific exceptions and, possibly, development exceptions.76 c. Most-favored-nation (MFN) standard. The inclusion of an MFN clause in a BIT, or other type of IIA, has the effect of extending, to the home contacting country, more favorable terms of investment granted to a third state by the host contracting country. Thereby, discriminatory terms of investment, operating against investors from the home contracting country, are prevented, and the equal treatment of all foreign investors by the host country is ensured. The MFN standard thus ensures equality of competitive conditions in the host country market as between foreign investors of different nationalities, thereby complementing the national treatment standard, which, as noted above, does the same in relation to competition between domestic and foreign investors. The MFN standard presents few drafting issues. It is a treaty-based standard that may be conditional or unconditional, though the latter formulation is now

2004 available at http://ita.law.uvic.ca/documents/Oxy-EcuadorFinalAward_001.pdf felt able to compare the treatment accorded to a U.S. oil company with the treatment accorded to exporters in general, rather than to Ecuadorian oil companies: see paras. 173–176. For a critical assessment see the comment by Susan D. Franck in 99 AJIL 675 (2005). This would appear to be at odds with the policy of the OECD which is to encourage comparison within the same industrial sector. 74. UNCTAD Key Issues Vol. I above n.68 at 173. 75. See for example, GATS Article XVII(2) and (3). 76. See UNCTAD Key Issues Vol. I above n.68 at 177–180.

the framework of investment protection: the content of bits 53

the norm. It will be subject to exceptions, especially in relation to taxation and regional economic integration commitments entered into by the contracting parties under other agreements. It may appear on its own or be combined with the national treatment standard in a single provision. There may also be provision for the more favorable of the two standards to prevail in any given case.77 Despite recent arbitral decisions, it remains uncertain how far the MFN standard can allow a claim that more favorable treatment, accorded by a host contracting party to nationals of a home contracting country under another IIA (the “thirdparty agreement”), must extend to nationals of the home contracting country under the IIA pertaining to the claimant (the “base agreement”).78 Such an argument was put forth by the claimant in Asian Agricultural Products Ltd v. Republic of Sri Lanka.79 It was rejected on the ground that the claimant had failed to show that the other BIT did, in fact, offer more favorable treatment; however, the tribunal did not rule out such an argument in principle, and it would appear to be consistent with the aims of the MFN standard. In this light, more recent tribunals have revisited the issue. The application of the MFN clause in an IIA will depend upon its interpretation in accordance with the accepted canons of treaty interpretation under international law.80 Given the considerable differences in the drafting of such clauses, this approach has produced significant differences of outcome in the cases. Thus, some awards offer an expansive interpretation of the MFN clause while others are more cautious.81 In addition, a distinction has been made between the application of the MFN clause to substantive protection standards contained in an IIA and procedural standards applicable to the dispute settlement provisions of the agreement.82 77. See for example NAFTA Article 1104 above n.19 at 74. 78. For a review of the case law see further Jurgen Kurtz “The MFN Standard and Foreign Investment: An Uneasy Fit?” 6 JWIT 861 (2004); Locknie Hsu “MFN and Dispute Settlement – When the Twain Meet” 7 JWIT 25 (2006); Rudolph Dolzer and Terry Myers “After Tecmed: Most-Favored-Nation Clauses in Investment Protection Agreements” 19 ICSID Rev-FILJ 49 (2004); Dana H. Freyer and David Herlihy “Most-Favored-Nation Treatment and Dispute Settlement in Investment Arbitration: Just how ‘Favored’ is ‘MostFavored’?” 20 ICSID Rev-FILJ 58 (2005). The terminology in the text is taken from Dolzer and Myers at 50. 79. Asian Agricultural Products Ltd v. Republic of Sri Lanka (ICSID Tribunal, Final Award, June 21, 1990) 30 ILM 577 (1991). 80. See Freyer and Herlihy above n.78 at 62–63. 81. For discussion of the cases see further Muchlinski above n.1 at 630–635. 82. See in particular Maffezini v. Spain ICSID Case NO. ARB/97/7 Decision on Objections to Jurisdiction 25 January 2000: 16 ICSID Rev-FILJ 212 (2001) which held that a more favorable dispute settlement clause in the third party agreement could extend to the investors rights under the dispute settlement clause in the base agreement. This case has not been uniformly followed due to the reasons given in the text. It was followed in Siemens AG v. Argentina ICSID Case NO.ARB/02/8 award on jurisdiction of 3 August 2004 available at www.worldbank.org/icsid/cases or 44 ILM 138 (2005); Cammuzi International

54 peter muchlinski

d. The observance of obligations. BITs usually contain provisions that ensure the observance of laws and regulations, which provide for more favorable treatment of investors than the standards contained in the treaty, and of obligations entered into under investment agreements or authorizations.83 The former preserves any benefits owed to investors under host country laws, regulations, or other international agreements that go beyond the standards of treatment provided for in the treaty, thereby avoiding the use of the BIT as a statement of the maximum possible protection that the investor can enjoy. The latter ensures that the observance of any obligations entered into by the host country towards the foreign investor, either under a state contract or within the terms of an authorization for the admission of the investment, becomes an obligation under the treaty.84 Such a provision, known as an “umbrella clause,” ensures an additional measure of contractual stability by making the observance of obligations owed to the investor a treaty standard, thereby reinforcing the existing contractual duties owed to the investor. “Umbrella clauses” have been the subject of some controversial arbitral decisions before ICSID. In particular, two decisions, interpreting, respectively, the Swiss BITs with Pakistan and the Philippines, sought to determine whether the umbrella clause in those agreements provided jurisdiction for an international arbitral tribunal to determine the effects of an alleged breach of the underlying investment agreement between the investor and the host country.85 Each case was brought by the same claimant, the Swiss based Societe Generale de

SA v. Argentina ICSID Case NO.ARB/03/7 decision on jurisdiction 10 June 2005 available at http://www.asil.org/pdfslcvajd050614.pdf ; Gas Natural SDG SA v. Argentina ICSID Case No. ARB/03/10 Decision on Jurisdiction 17 June 2005 available at http://ita.law. uvic.ca; Suez, Sociedad General de Aguas de Barcelona S.A., and Vivendi Universal S.A. v. Argentina ICSID Case No. ARB/03/19 Decision on Jurisdiction 3 August 2006 available at www.ita.law.uvic.ca. Maffezini was not followed in Salini Construttori SPA and others v. Jordan ICSID Case NO/ARB/02/13 decision on jurisdiction 29 November 2004 available at www.worldbank.org/icsid/cases or 44 ILM 573 (2005) and Plama Consortium Limited v Bulgaria. ICSID Case No.ARB/03/24 decision on jurisdiction 8 February 2005 available at www.worldbank.org/icsid/cases or 44 ILM 721 (2005). This case was distinguished in Suez v Argentina above at paras.64–67. 83. See UNCTAD 1998 above n.2 at 56, 86–87; U.S.-Uruguay BIT 2004 above n.5 Article 16. 84. UNCTAD ibid. at 56 n.66. 85. See for detailed analysis Christoph Schreuer “Travelling the BIT Route—of Waiting Periods, Umbrella Clauses and Forks in the Road” 5 JWIT 231 (2004); Thomas Waelde “The ‘Umbrella’ Clause in Investment Arbitration—a Comment on the Original Intentions and Recent Cases” 6 JWIT 183 (2005); Anthony Sinclair “The Origins of The Umbrella Clause in the International Law of Investment Protection” 20 Arb. Int’l 411 (2004); Stanimir A. Alexandrov “Breaches of Contract and Breaches of Treaty: The Jurisdiction of Treaty-Based Arbitration Tribunals to Decide Breach of Contract Claims in SGS v. Pakistan and SGS v. Philippines” 5 JWIT 555 (2004); Yuval Shany “Contract Claims vs. Treaty

the framework of investment protection: the content of bits 55

Surveillance (SGS), and covered the same question, namely, whether the applicable BIT could be used to found a claim against the host country for alleged breaches of the investment contract between the disputing parties concerning the provision of pre-shipment customs inspection services. In SGS v. Pakistan,86 Pakistan argued that the claim could not be heard under the BIT as it arose entirely out of the contract between the parties, and Article 11 of that contract provided for all disputes to be heard before a local arbitral tribunal in the host country. Pakistan relied on the decision of the Annulment Committee in the Vivendi Case 87 where it was held that where the basis of the claim is a breach of contract, the tribunal will give effect to any valid choice of forum clause. The tribunal upheld Pakistan’s objection to jurisdiction. It followed the Vivendi Case and accepted that contract-based and treaty-based claims were not identical.88 Accordingly, contract claims stood to be determined by the tribunal selected under the forum selection clause in the contract. This did not, however, oust the jurisdiction of the international arbitral tribunal to hear claims based on the BIT, especially as the dispute settlement clause in the BIT, Article 9, did not state that the jurisdiction of the tribunal to determine claims in violation of the BIT was not exclusive.89 The tribunal then turned its attention to the claimant’s argument that the umbrella clause in the Swiss-Pakistan BIT gave the tribunal jurisdiction over the contract claims, noting that this was the first time an international arbitral tribunal had been asked to examine the legal effect of such a clause. Article 11 of the BIT states, “Either Contracting Party shall constantly guarantee the observance of the commitments it has entered into with respect to the investments of the investors of the other Contracting Party.” The Tribunal rejected the claimant’s argument, both on a textual basis and for policy reasons. The term “commitments” was potentially susceptible to almost infinite expansion and the claimant’s interpretation went against the generally accepted view that a breach of contract alleged by an investor in relation to a contract concluded with a state was a matter of municipal law and could not automatically be elevated into a breach of international law. The clause had to show that the contracting parties intended it to cover breaches of contract and the text of Article 11 did not do so. Nor was any other evidence of such an Claims: Mapping Conflicts Between ICSID Decisions on Multisourced Investment Claims” 99 AJIL 835 (2006). 86. ICSID Case No.ARB/01/13 decision on objections to jurisdiction 6 August 2003 available at www.worldbank.org/icsid/cases or 18 ICSID Rev-FILJ 307 (2003). 87. Compania des Aguas del Aconquija SA and Vivendi Universal (Formerly Compagnie Generale des Eaux) v. Argentina ICSID Case ARB/97/3 decision on annulment 3 July 2002) available at www.worldbank.org/icsid/cases or 41 ILM 1135 (2002). On the relationship between contract claims and treaty-based claims see further Muchlinski above n.1 at 722–724. 88. Above n.86 at para. 147. 89. Ibid. at para. 152.

56 peter muchlinski

intention adduced by the claimant.90 Furthermore, the consequences of adopting such an interpretation pointed against it. It would incorporate unlimited numbers of state contracts and municipal law instruments and render the substantive protection provisions in Articles 3 to 7 of the BIT superfluous. It could also nullify any freely negotiated dispute settlement clause in a state contract and would allow the investor to dictate the choice of forum, leaving the state at a procedural disadvantage.91 Finally, the tribunal noted that Article 11 was placed at the end of the treaty and not among the main substantive obligations. It was thus not meant to project a substantive obligation. This did not mean that it was devoid of meaning. It could be invoked, for example, to ensure that the host country enacts any rules and obligations necessary to give effect to the investment, or where it interferes with the ability of the investor to pursue an international claim once consent to international arbitration has been given by the host country.92 By contrast, in SGS v. Philippines 93 an umbrella clause was held to cover the contractual dispute between the parties, notwithstanding the presence of an exclusive jurisdiction clause in favor of local trial courts for such disputes. The tribunal considered that it was clear from the general language of Article 25(1) of the ICSID Convention that jurisdiction may extend to disputes of a purely contractual character, and Article 42(1), which covers the issue of applicable law, did not preclude such jurisdiction. However, the umbrella clause in the SwissPhilippines BIT was phrased differently from the clause in the Pakistan case. Article X(2) of the Swiss-Philippines BIT states, “Each Contracting Party shall observe any obligation it has assumed with regard to specific investments in its territory by investor of the other Contracting Party.” This was a crucial point of distinction between the two cases. The tribunal noted that the language of Article X(2) was mandatory and it applied to any legal obligation undertaken for a specific investment. Thus, it was unlike the formulation on the Pakistan BIT which applied to “all commitments” and was less clear.94 The tribunal continued by saying that, given the aim of the BIT, any uncertainty should be resolved in favor of the further protection of the covered investment. Thus, the umbrella clause ought to be effective within the BIT and there was no presumption against the inclusion of contractual issues under international law. It all depended on the interpretation of the actual clause before the tribunal. In the present case, the clause was effective in bringing a binding

90. Ibid. paras. 164–167. 91. Ibid. at para. 168. 92. Ibid. at paras. 169–172. 93. ICSID Case No.ARB /02/6 decision on objections to jurisdiction 29 January 2004 available at www.worldbank.org/icsid/cases. 94. Ibid. at paras. 116–117 and 121.

the framework of investment protection: the content of bits 57

legal obligation within the framework of the BIT, given its object and purpose.95 Having thus given effect to the umbrella clause, the tribunal then backed away from applying it to allow the claimant to bring their claim.96 It held that the exclusive jurisdiction clause was not rendered ineffective by the umbrella clause. The BIT was a framework agreement that could not override the specific provisions of particular contracts freely negotiated by the parties to the dispute. Nor could a later BIT replace an earlier investment contract, as these were legal instruments of differing types. In addition, although an exclusive jurisdiction clause could not affect the jurisdiction of a tribunal over a contractual claim, it could affect admissibility. In the present case SGS had to comply with the exclusive jurisdiction of the Philippine’s courts in relation to contract-based claims.97 Finally, the Tribunal inquired whether there were any independent claims based on the breach of the BIT as such. It held that there were not as the dispute was only about money owed under a contract. There was no expropriation claim raised on the facts, and the fair and equitable treatment claim, like the claim under the umbrella clause, was premature as it had to await the resolution of the amount payable under the contract before it could be addressed under the BIT.98 The above awards offer little guidance on the precise scope of the umbrella clause, other than the obvious fact that the effect of the clause depends on its actual wording.99 Given the absence of a doctrine of precedent in international arbitral jurisprudence, neither decision is conclusive or binding.100 Which approach is preferred will depend on the degree to which the discretion of the host country is sought to be controlled. A wider discretion for the host country flows from the approach of the Pakistan case while the Philippines case takes a more investor-centred approach.101 Given the aim of BITs, and other IIAs, 95. Ibid. paras. 116–122. 96. See for critical analysis Waelde above n.85. 97. Above n.93 at paras. 141–155. 98. Ibid. at paras. 159–163. 99. According to Shany, above n.85, the differences in the decisions can also be explained by different ideological positions taken by tribunals on the issue of how to resolve competition between different norms applicable to an investment dispute. The Pakistan Case is described as “disintegrationalist,” seeking to limit harmonization of norms by reference to limiting jurisdictional rules, while the Philippines Case is seen as “integrationalist” seeking harmonization of different rules before a widened ICSID jurisdiction. 100. See SGS v Philippines above n.93 at para.97 where the Tribunal explains that it is not bound to follow the interpretation of the umbrella clause in the Pakistan case. 101. See UNCTAD State Contracts Series on issues in international investment agreements (New York and Geneva, United Nations, 2004) at 23 and Schreuer above n.85 at 255 who feels the umbrella clause would be rendered devoid of effect if the Pakistan Case approach is taken. But see, for an opposite view, Pan American Energy LLC v. Argentina ICSID Case No. ARB/03/13 Decision on Preliminary Objections 27 July 2006 at www.ita. law.uvic.ca and El Paso International Company v. Argentina ICSID Case No.ARB/03/15 Decision on Jurisdiction 27 April 2006 at www.ita.law.uvic.ca. In both cases, the ICSID

58 peter muchlinski

to create a good investment climate, and to control excesses of governmental abuse of power in relation to investors, the correct approach may be to use the umbrella clause to cover only those breaches of contract that stem from such an abuse of power, and not from an ordinary commercial breach.102 This distinction may not resolve satisfactorily the issue of venue for the dispute. The investor may try to color any breach of contract in the language of maladministration, in order to convert the commercial breach of contract into a breach of the applicable IIA, so as to secure international arbitration over any local dispute settlement procedures stipulated in the contract. This was precisely the fear expressed by the tribunal in the Pakistan case when it referred to the risk of unequal procedural advantages accruing to the investor under an expansive reading of the umbrella clause. Nor is it clear what such a reading of the umbrella clause would add to existing protection available to the investor under the expropriation, fair and equitable treatment, and non-discrimination provisions of the applicable BIT or other IIA, which would appear to cover most, if not all, foreseeable types of maladministration on the part of the host country. If the umbrella clause means anything, it must mean that contractual obligations are protected as obligations under the BIT.103 If so, then the inexorable conclusion must be that all substantial breaches of contract are justiciable as breaches of the BIT, unless the parties to the investment contract expressly exclude them from admissibility in

tribunals felt that such a reading of the umbrella clause would allow any minor contract claim to become a BIT claim, in their view an inappropriate remedy that ICSID tribunals had to control. See Pan American para. 110 and El Paso para. 82. 102. See Waelde above n.85 at pp. 235–236. 103. See Schreuer above n.85 at 255. But see Salini Construction v. Jordan ICSID Case No.ARB/02/13 decision on jurisdiction 29 November 2004, 44 ILM 573 (2005) or www. icsid.org/cases at paras. 123–126 where the umbrella clause in Art. 2(4) of the Italy-Jordan BIT was held incapable of incorporating contractual obligations into the BIT, only obligations protective of a legal framework favorable to investors. See too Pan American Energy LLC v. Argentina above n.101 where the tribunal endorsed the SGS v. Pakistan decision and held that: “it is especially clear that the umbrella clause does not extend its jurisdiction over any contract claims when such claims do not rely on a violation of the standards of protection of the BIT, national treatment, MNF [sic] clause, fair and equitable treatment, full protection and security, protection against arbitrary and discriminatory measures, protection against expropriation or nationalisation either directly or indirectly, unless some requirements are respected. However, there is no doubt that if the State interferes with contractual rights by a unilateral act, whether these contractual rights stem from a contract entered into by a foreign investor with a private party, a State autonomous entity or the State itself, in such a way that the State’s action can be analysed as a violation of the standards of protection embodied in a BIT, the treaty-based arbitration tribunal has jurisdiction over all the claims of the foreign investor, including the claims arising from a violation of its contractual rights.” Para. 112. See too, on identical terms, El Paso International v. Argentina above n.101 para. 84.

the framework of investment protection: the content of bits 59

their contract, as suggested in the Philippines case. Clearly, tribunals must offer greater clarity and uniformity of approach on this issue. 2. Specific standards of treatment BITs contain standards on specific aspects of the treatment of a foreign investment by the host country. In particular, the following matters have been commonly dealt with by provisions in BITs: the free transfer of payments relating to the investment out of the host country; compensation for losses due to expropriation, armed conflict, or internal disorder; the promotion of investments; and the rights of entry and sojourn of individuals in connection with the investment and performance requirements. Taxation is not dealt with under BITs, save in the North American models that extend the protection against expropriation to tax measures.104 Some treaties contain provisions expressly excluding this area from their scope.105 a. Free transfer of payments. Provisions on the free transfer of currency are among the most important in BITs.106 They seek to ensure that the investor can transfer the income from its investment out of the jurisdiction of the host country and repatriate its capital on the termination of the investment. At the same time, such provisions may include limits on the free transfer of currency in the interests of the balance-of-payments position of the host country. Thus a balance is generally struck between the interests of the investor and of the host country. The right to transfer currency is usually subject to the requirements that it shall be made without delay, in convertible currency, and at the official rate of exchange at the date of transfer.107 Some treaties go on to specify the types of transfers protected, often by way of a non-exhaustive list,108 while others simply refer to transfers “in respect of investments.”109 104. See U.S.-Uruguay BIT 2004 above n.5 Article 21; Canada Model BIT Article 16 above n.12.The U.S. BIT applies non-discrimination provisions to indirect taxation as well. Both contain procedural requirements that must be complied with by the investor before any dispute in relation to tax treatment can be brought under these agreements. On taxation issues under the U.S.-Argentina BIT of 1991see further Pan American v. Argentina above note 101 at paras. 117–139 and El Paso v Argentina above note 101 para.101–116. 105. UNCTAD 1998 above n.2 at 63. 106. UNCTAD 2007 above n.2 at 56. 107. See, e.g., U.S.-Uruguay BIT 2004 above n.5 Article 7; Netherlands-Philippines BIT 1985 Article 7 UNCTAD 1998 above n.2 at 77. See further UNCTAD 2007 above n.2 at 57–61. 108. The U.S.-Uruguay BIT 2004 Article 7(1) lists: contributions to capital; profits, dividends, capital gains, and proceeds from the sale of all or any part of the covered investment or from the partial or complete liquidation thereof; interest, royalty payments, management fees and technical assistance, and other fees; payments under a contract, including a loan agreement; compensation for losses due to armed conflict or civil strife or for expropriation of property; payments arising out of a dispute (above n.5). 109. See, e.g., UK-Philippines BIT above n.28 Art VII(1): “Each Contracting Party shall in respect of investments permit nationals or companies of the other Contracting Party the free transfer of their capital and of the earnings from it . . . .”

60 peter muchlinski

Most BITs contain restrictions on free transfer based on the interests of the host country. Thus the UK-Philippines BIT states, in Article VII(1): Each Contracting Party shall in respect of investments permit nationals or companies of the other Contracting Party the free transfer of their capital and of the earnings from it, subject to the right of the former Contracting Party to impose equitably and in good faith such measures as may be necessary to safeguard the integrity and independence of its currency, its external financial position and balance of payments, consistent with its rights and obligations as a member of the International Monetary Fund.110 The reference to the rights and obligations of the host country as a member of the International Monetary Fund (IMF) introduces by implication the standards of the IMF Articles of Agreement into the BIT regime.111 These aim at the progressive elimination of restrictions on the free transfer of funds arising from current transactions, while accepting the right of countries to impose restrictions on the free movement of capital where this is necessary in the interests of that country.112 However, should the BIT provide for a greater freedom of transfer for the investor than the IMF Articles of Agreement, it is arguable that the terms of the BIT should prevail, as these represent a specialized regime of investor protection that is more elaborate than the IMF regime on the transfer of funds and capital.113 Where the BIT does not mention the IMF regime, the presumption is that it is ousted by the regime of the bilateral treaty. Many BITs provide for the progressive phasing out of the repatriation of capital where this is warranted by the host country’s foreign exchange situation. Thus, the UK-Hungary BIT provides that while each contracting party has the right, in exceptional balance of payments difficulties and for a limited period, to exercise equitably and in good faith powers conferred by its laws, “[s]uch powers shall not however be used to impede the transfer of profit, interest, dividends, royalties or fees; as regards investments and any other form of return, transfer of a minimum of 20 per cent a year is guaranteed.”114 The AALCC Models also introduce restrictions on the free transfer of capital and returns. Model “A” contains provisions similar to the UK provisions cited above, permitting “reasonable restrictions” and allowing for a 20% guaranteed minimum right of transfer 110. Above n.28. See further UNCTAD 2007 above n.2 at 62–63 where three main types of exceptions are identified, namely, compliance with regulatory measures such as bankruptcy, securities trading criminal acts or compliance with tribunal decisions, balance of payments crises and financial services regulation. 111. Laviec above n.2 at 142. 112. Ibid. at 138–142. 113. Ibid. at 143. 114. UK-Hungary BIT Treaty Series No.3 (1988) Cm 281. Article 7(1). See too UK-Jamaica BIT Article 7(a) and (b) cited in UNCTAD 1998 above n.2 at 79. See too Japan-Vietnam BIT 2003 Article 16 in UNCTAD 2007 above n.2 at 62–63.

the framework of investment protection: the content of bits 61

in any year period where exceptional financial or economic conditions are experienced.115 Model “B” leaves the repatriation of capital and returns subject to the terms set out by the host country at the time of the reception of the investment. These shall be set out in the letter of authorization issued by the host country on the approval of the investment and will remain unchanged throughout the period of the investment unless the investor and the host country agree to any alterations.116 b. Compensation for losses due to armed conflict or internal disorder. Most BITs contain clauses providing for the compensation of the investor for losses due to armed conflict or internal disorder. These do not establish an absolute right to compensation. Rather, they lay down that the investor shall be treated in accordance with the national treatment and/or MFN standard in the matter of compensation. Some treaties, for example the German model treaty, deal with such compensation in a single provision, which also deals with compensation for expropriation.117 Others deal with these two types of compensation in separate provisions.118 This may be preferable, as compensation for expropriation is not measured by reference to national treatment or MFN standards, but by reference to the standard of “prompt, adequate and effective compensation.”119 Most BITs guarantee the free transfer of compensation to the investor’s home country.120 According to the tribunal in the case of Saluka v. Czech Republic, the standard obliges the host country to adopt all reasonable measures to protect assets and property from threats or attacks that may particularly target foreigners or certain groups of foreigners. Nonetheless, the “full security and protection” clause is not meant to cover any kind of impairment of an investor’s investment, but to protect more specifically the physical integrity of an investment against interference by use of force.121 c. Compensation for expropriation. One of the most contentious disputes between capital-exporting and capital-importing countries over international minimum standards of treatment has been that relating to compensation for expropriation. To ensure that their preferred approach prevails, capital-exporting countries have embodied their understanding of the applicable rules of international law into the expropriation provisions of BITs.122 Thus, practically all BITs 115. Above n.21 Article 6. 116. Ibid. Article 6. 117. See Article 4. 118. See, e.g., UK-Philippines BIT n.28 above Article V (compensation for expropriation), Article VI (compensation for loss due to war, other armed conflicts, revolution, national emergency, revolt, insurrection, or riot). 119. See Asian Agricultural Products Ltd v. Republic of Sri n.79 above at 653–654. (Dissenting opinion of Dr S.K.B.Asante). 120. UNCTC Bilateral Investment Treaties (1988 UN Doc ST/CTC/65) para.2 01 at p. 47. 121. Saluka v. Czech Republic above n.64 at paras. 483–484. 122. For a full analysis, see: UNCTAD 1998 above n.2 at 65–73; UNCTAD 2007 above n.2 at 44–52; Laviec above n.2 ch. V.

62 peter muchlinski

contain a provision that permits expropriation or nationalization of assets owned by the investor from the other contacting country only where this is done for a public purpose, under due process of law, without discrimination, and upon the payment of “prompt, adequate and effective,” “adequate” or “just” compensation in accordance with the fair market value of the assets immediately before expropriation.123 Nevertheless, variations from this model do exist. For example, while the AALCC Model “A” BIT accepts this formulation,124 Model “B” refers to “principles for the determination of appropriate compensation.”125 Furthermore, the UK-Peoples Republic of China BIT specifies a standard of “reasonable compensation” which is defined as the real value of the expropriated investment immediately before the expropriation became public knowledge.126 This amounts to derogation from the usual British practice that specifies the market value of the assets expropriated, or, in the absence of a determinable market value, the actual loss sustained, on or immediately before the date of expropriation.127 It was necessitated by the refusal of China to accept a reference to “market” in the treaty in view of its command economy.128 China has adopted a similar stance in treaties with other capital-exporting countries. Thus the treaty with the BelgiumLuxembourg Union speaks of compensation amounting to, “the value of the property and assets invested on the date immediately preceding the date of expropriation, or on the date on which the expropriation was made public.”129 In the BIT with France the amount of compensation payable, “shall reflect the true value of the investments in question.”130 In the BIT with Japan the standard of compensation is different again. Here the compensation payable “shall be such as to place the nationals and companies in the same financial position as that in which the nationals and companies would have been if the expropriation, nationalization or any other measures the effects of which would be similar to expropriation or nationalization . . . had not been taken.”131 Thus a restitutionary

123. See, e.g., U.S.-Uruguay BIT above n.5 Article 6(1). The U.S. model also adds due process, fair and equitable treatment, and observance of international minimum standards of treatment as requirements for a lawful expropriation. See too Angola-United Kingdom BIT 2000 Article 5 and Belarus-Croatia BIT 2001 Article 5, both in UNCTAD 2007 above n.2 at 47. 124. Above n.21 Article 7. 125. Ibid. Article 7 Alternative 1. Note that Alternative 2 refers to the “prompt, adequate and effective” formula. 126. Article 5. See Denza and Brooks above n.2 at 919. 127. See, e.g., UK-Philippines BIT above n.28 Article V(1). 128. Denza and Brooks n.126 above. 129. Protocol to the Belgium-Luxembourg-PRC BIT 4 June 1984 Art.2: 24 ILM 537 at 546 (1985). 130. France-PRC BIT Annex para.2. Ibid. at p. 561. 131. Above n.56 Art.5(3).

the framework of investment protection: the content of bits 63

standard is adopted. It can hardly be said that China has taken a uniform attitude to this question. However, she does not accept the traditional formula of “prompt, adequate and effective” compensation.132 Concerning the modalities of payment, most BITs provide for the payment of interest and for payment without delay in fully realizable and transferable form. Furthermore, many treaties provide for judicial review, before the courts and tribunals of the host country, of the legality of the expropriation, and of the amount of compensation being offered.133 Finally, the majority of BITs cover not only direct expropriation but also indirect measures that have the effect of neutralizing the value of the investor’s assets, while leaving their formal ownership intact.134 In this context, the distinction between compensable takings and legitimate governmental regulation becomes significant. This issue has created some uncertainty due to the lack of proper definitions of such terms in IIAs. While most expropriation provisions will state that a direct or indirect taking will require compensation, they do not exemplify measures that qualify as such. More recent agreements are taking tentative steps to rectify this problem. Thus the United States-Uruguay BIT of 25 October 2004, which is based on the 2004 U.S. Model BIT, offers the following definition of an indirect taking in Annex B paragraph 4: The second situation addressed by Article 6(1) [expropriation] is known as indirect expropriation, where an action or series of actions by a Party has an effect equivalent to direct expropriation without formal transfer of title or outright seizure. (a) The determination of whether an action or series of actions by a Party, in a specific fact situation, constitutes an indirect expropriation, requires a case-by-case, fact-based inquiry that considers, among other factors: (i) the economic impact of the government action, although the fact that the action or series of actions by a Party has an adverse effect on the economic value of an investment, standing alone, does not establish that an indirect expropriation has occurred; 132. According to Sornarajah, the fact that the “prompt, adequate and effective” formula is not used consistently in treaty practice weakens the argument that it represents the applicable international standard of compensation: “State Responsibility and Bilateral Investment Treaties” 20 JWTL 79 at 91–92 (1986). See, for a more recent reiteration of this position, M. Sornarajah The International Law on Foreign Investment (Cambridge, Cambridge University Press, 2nd Ed, 2004) ch.10. According to UNCTAD, most agreements today include language that has the effect of applying the standard of prompt adequate and effective compensation: UNCTAD 2007 above n.2 at 48–49. 133. See, e.g., UK-Philippines BIT above n.28 Article V(1); Japan-PRC BIT above n.56 Article 5(4). 134. For examples of the formulations used, see: UNCTAD 1998 above n.2 at pp.65–66 and UNCTAD 2007 above n.2 at 44–47. The usual formulation is to refer to measures having an equivalent effect to, or which are tantamount to, expropriation.

64 peter muchlinski

(ii) the extent to which the government action interferes with distinct, reasonable investment backed expectations; and (iii) the character of the government action. (b) Except in rare circumstances, non-discriminatory regulatory actions by a Party that are designed and applied to protect legitimate public welfare objectives, such as public health, safety, and the environment, do not constitute indirect expropriations.135 This type of provision was first introduced into United States bilateral Free Trade Agreements (FTAs) pursuant to deliberations in Congress on the meaning of expropriation under the Trade Act of 2002.136 It reflects the standards of review contained in U.S. domestic judicial review case law.137 This is a welcome development in so far as it allows for more specific, and structured, discussion of how the scope of a country’s right to regulate will be assessed in the case of regulatory acts that impair the economic value of an investment. d. Other specific standards. Among other, less common, specific standards to be found in BITs are general statements committing the home and/or host country to providing investment promotion incentives;138 provisions protecting the right of entry and sojourn of individuals in connection with the investment, subject to the laws of the host country and the hiring of local personnel;139 restrictions on the imposition of performance requirements on investors by the host country;140 and transparency obligations.141 In addition, the most recent U.S. BITs contain provisions asserting that it is inappropriate for host countries to seek investment through the lowering of environmental or labor standards,

135. U.S.-Uruguay BIT 2004, 44 ILM 268 at 294 (2005). 136. See, for example, U.S.-Chile FTA Annex 10-D; U.S.-Singapore FTA Ch.15 (exchange of letters on expropriation) both available on http://www.ustr.gov/new/fta or www.unctad. org/iia. 137. See further Gary H.Sampliner “Arbitration of Expropriation Cases under U.S. Investment Treaties—a Threat to Democracy or the Dog that Didn’t Bite?” 18 ICSID RevFILJ 1 at pp. 35–42 (2003). 138. UNCTAD 1998 above n.2 at p. 50–51; UNCTAD 2007 above n.2 at 26–28. 139. UNCTAD 1998 Ibid. at pp. 83–84 UNCTAD 2007 ibid. at 69–73. 140. UNCTAD 1998 above n.2 at p. 81–82 and see U.S.-Uruguay BIT 2004 above n.5 Article 8. See for a full discussion UNCTAD Host Country Operational Measures Series on issues in international investment agreements (New York and Geneva, United Nations, 2001) and see too UNCTAD 2007 at 64–69. 141. See, for example, U.S.-Uruguay BIT 2004 Articles 10–11. See for a full discussion UNCTAD Transparency Series on issues in international investment agreements (New York and Geneva, United Nations, 2004) and see examples for transparency provisions in UNCTAD 2007 above n.2 at 76–80.

the framework of investment protection: the content of bits 65

while the Canadian counterpart applies to health safety and the environment.142 The Canadian model is also notable for the inclusion of a general exceptions clause protecting the rights of the contracting parties to regulate in the fields mentioned by its terms. The clause follows the general pattern of Article XX GATT 1994 by listing areas in which regulation is consistent with the provisions of the BIT and adds a “chapeau” requiring such regulation not to be arbitrary, discriminatory, or a disguised restriction on trade and investment.143 By contrast, the U.S. model reserves only measures aimed at the maintenance or restoration of international peace or security, or the protection of essential security interests.144 Finally, it should be noted that the extension of future BITs and other IIAs to include obligations upon investors and home countries has been proposed. This is seen as a means of ensuring a fair balance of rights and duties between the investor, the host country, and the home country, thereby ending the prioritization of investor rights and host country obligations characteristic of existing agreements.145 D. Dispute Settlement Clauses Dispute settlement clauses in BITs, and other types of IIAs, can be divided between those dealing with disputes between the contracting countries as to the observance and interpretation of the treaty, and those dealing with disputes between the investor and the host country.

142. U.S.-Uruguay BIT 2004 Articles 12, 13 above n.5, Canadian Model FIPA Article 11 above n.12. The areas covered are: protection of human, animal, and plant life and health; compliance with laws not inconsistent with the agreement; the conservation of living or non-living exhaustible natural resources; prudential financial regulation; monetary credit and exchange rate policies; essential security interests; the upholding of UN obligations and international peace and security interests; confidentiality laws; cultural industries; and measures taken in conformity with WTO decisions. 143. Above n.12 Article 10. 144. U.S.-Uruguay BIT 2004 above n.5 Article 18. For further examples to exception provisions see UNCTAD 2007 above n.2 at 80–99. The main exceptions covered by the UNCTAD study relate to taxation issues (though some agreements, notably the U.S. and Canadian agreements, allow for certain taxation measures to be covered by the protective provisions of the agreement); essential security and public order; protection of health and natural resources; cultural exceptions; prudential measures in financial services; MFN exceptions; environmental protection; labor rights protection; and miscellaneous exceptions such as intellectual property rights exceptions. 145. See the IISD Model International Agreement on Investment for Sustainable Development Negotiators Handbook (April 2005 revised April 2006) http://www.iisd.org/pdf/2005/ investment_model_int_handbook.pdf. The 2005 version of the Model Agreement is reproduced in 20 ICSID Rev-FILJ 91 (2005) and see too Howard Mann “Introductory Note” ibid. 84. See also UNCTAD World Investment Report 2003 (New York and Geneva, United Nations, 2003) ch.VI.

66 peter muchlinski

1. Disputes between the contracting parties Two settlement procedures are used in all BITs, namely, negotiation and ad hoc arbitration.146 The usual procedure is for a dispute to be settled by negotiation between the contracting countries, or, if this is not possible, to go to arbitration. The arbitration procedure most commonly followed is that of a three-member tribunal, each contracting party appointing one member, with the Chairman appointed by these two from among nationals of third countries.147 There is usually a time limit specified within which such choices must be made. If they are not, then recourse may be had to the President of the International Court of Justice, the Secretary-General of the United Nations, the Secretary-General of ICSID, or to some other specified international figure, to make the necessary appointments.148 The decision of the tribunal is reached by majority vote and is binding on the parties. The tribunal fixes its own rules of procedure although, the U.S. model treaty specifies the use of the United Nations Commission on International Trade Law (UNCITRAL) Arbitration Rules in the absence of agreement between the parties.149 As to the applicable law, few BITs contain specific clauses on this matter. Those that do tend to refer to “international law” or to “general principles of law.”150 A particular provision found in virtually all BITs, which can give rise to disputes between the contracting countries, concerns the subrogation of the investor’s claims against the host country to the home country, after that investor has been compensated for its losses out of “political risks” insurance taken out with the investment insurance agency of the home country.151 In such a case, the home country is entitled to benefit from any compensation that the host country is liable to pay to the investor, up to the amount that has been paid out to the investor under the insurance cover. The normal formulation excludes any acceptance of liability on the part of the host country merely by virtue of its recognition of the home country’s rights of subrogation, a right that arises from

146. UNCTAD 1998 above n.2 at 100, UNCTAD 2007 above n.2 at 126–130. See for a full discussion UNCTAD Dispute Settlement: State-State Series on issues in international investment agreements (New York and Geneva, United Nations, 2003) also in UNCTAD Key Issues Vol. I above n.2 at 315–345. 147. See, e.g., UK-Philippines BIT above n.28 Article XI; Japan-PRC BIT above n.56 Article13. 148. UNCTAD 1998 above n.2 at 101; Australia-India BIT 1999 Article 13 in UNCTAD 2007 above n.2 at 127. 149. See, for example U.S.-Uruguay BIT 2004 above n.5 Article 37(1). 150. UNCTAD 1998 above n.2 at 101–102; Hungary-Lebanon BIT 2001 Article 9 in UNCTAD 2007 above n.2 at 127. 151. UNCTAD 1998 above n.2 at 88–89. The U.S. does not use such a clause in its BITs as it concludes special investment guarantee agreements with host countries as part of its overseas investment insurance program. See further Muchlinski above n.1 at 615.

the framework of investment protection: the content of bits 67

the insurance contract between the home country and the investor, and not from any delictual conduct on the part of the host country.152 2. Disputes between the host state and the foreign investor Early BITs did not cover the issue of disputes between the host country and the investor, however, the conclusion of the Washington Convention of 1965, setting up ICSID, prompted the inclusion of ICSID jurisdiction clauses in BITs.153 All the model agreements in their current versions include an ICSID clause.154 Earlier treaties refer to ICSID as the only system of dispute settlement to be used between investors and host countries,155 while more recent ones offer a choice between ICSID and other systems of international arbitration.156 More recently the U.S. and Canadian model BITs have included detailed provisions of dispute settlement that seek to tailor procedures to the specific concerns of these countries for more effective and transparent arbitral procedures in the investment field.157 As with dispute settlement between the contracting parties, virtually all investor-country dispute settlement clauses require that the dispute be settled first by amicable means including consultation negotiation and other non-binding third party means.158 152. See, e.g., UK-Philippines BIT above n.28 Article VIII(1): “If either Contracting Party makes payment under an insurance or guarantee agreement with its own nationals or companies in respect of an investment or any part thereof in the territory of the other Contracting Party, the latter Contracting Party shall recognise the assignment of any right or claim arising from the indemnity paid, by the party indemnified to the former Contracting Party, and that the former Contracting Party is entitled by virtue of subrogation to exercise the rights and assert the claims of such nationals or companies. This does not necessarily imply, however, a recognition on the part of the latter Contracting Party of the merits of any case or the amount of any claim arising therefrom.” 153. Washington Convention above n.38. See generally UNCTAD Dispute Settlement: Investor-State Series on issues in international investment agreements (New York and Geneva, United Nations, 2003) also in UNCTAD Key Issues Vol. I above n.2 at 347–380 and see Muchlinski above n.1 ch. 18. 154. UNCTAD 1998 above n.2 at 94–95. 155. This is true in UK practice: Denza and Brooks above n.2 at 920; and see UKPhilippines BIT above n.28 Article X. 156. See, e.g., AALCC Model “A” above n.21 Article 10(iii)-(v) which offers ICSID as the first choice for the parties to the dispute, with the UNCITRAL Arbitration Rules 1976 applying where ICSID procedures are inapplicable. See too the ASEAN Agreement on Promotion and Protection of Investments of 15 December 1987 above note 112 Art.X(2) which offers a choice of ICSID, UNCITRAL, the Regional Centre for Arbitration at Kuala Lumpur, or any other regional centre for arbitration in ASEAN. For further examples see UNCTAD 1998 above n.2 at 95–96, UNCTAD 2007 at 100–101, 110–12, and UNCTAD Dispute Settlement: Investor-State above n.153 at pp. 37–44. 157. See U.S.-Uruguay BIT 2004 Section B above n.5; Canadian Model FIPA Section C above n.12. For a full discussion on the innovations introduced by the U.S. and Canadian Model Agreements see UNCTAD 2007 above n.2 at 119–126. 158. See, for example, U.S.-Uruguay BIT 2004 ibid. Art.23 and UNCTAD Dispute Settlement: Investor-State above n.153 at pp. 23–25.

68 peter muchlinski

Where a BIT refers to ICSID arbitration the precise wording of the clause is important. If the wording shows complete and unconditional consent on the part of the contracting country to the use of ICSID arbitration, this can amount to a unilateral offer of such arbitration to investors from the other contracting country, which can be turned into a binding consent to ICSID jurisdiction where the investor requests such arbitration in writing.159 On the other hand, where the clause is conditional on further agreement to submit the dispute to ICSID it will be ineffective, on its own, to found the jurisdiction of the Centre. The host country will then retain the choice whether or not to consent to ICSID jurisdiction. In order to avoid disputes over the effect of clauses purporting to establish ICSID jurisdiction, ICSID has drawn up model clauses for BITs.160 However these have not been widely used and consent clauses have grown up in a rather haphazard fashion, requiring interpretation of their precise terms.161 In more recent years, the question has arisen whether choice of dispute settlement clauses in investor-country agreements can operate to prevent the institution of international dispute settlement procedures where an option to use domestic dispute settlement procedures is available and the investor makes a choice to that effect.162 Such “fork-in-the-road” clauses have been the subject

159. Thus, by Article X(1) of the UK-Philippines BIT n.28 above: “The Contracting Party in the territory of which a national or company of the other Contracting Party makes or intends to make an investment shall assent to any request on the part of such national or company to submit, for conciliation or arbitration, to the Centre established by the Convention on the Settlement of Investment Disputes between States and Nationals of Other States opened for signature at Washington on 18 March 1965 any dispute that may arise in connection with the investment.” By Article X(2) access to ICSID arbitration is extended to, “A company which is incorporated or constituted under the law in force in the territory of one Contracting Party and in which before such a dispute arises the majority of shares are owned by nationals or companies of the other Contracting Party,” provided that the requirements of Article 25(2)(b) of the Washington Convention are complied with. See too UK-Sri Lanka BIT 13 February 1980 Article 8(1) which formed the legal basis for the ICSID arbitration in Asian Agricultural Products Ltd v. Republic of Sri Lanka above n.79. On the requirements of Article 25(2)(b) see Muchlinski above n.1 at 726–731. 160. See ICSID/6 (1969): 8 ILM 1341 (1969). 161. See Schreuer above n.38 at 210–221 and see Muchlinski above n.1 at 719–722. 162. An example of such a choice is found in Article 8 of the China-Vietnam BIT 1992 cited in UNCTAD 1998 above n.2 at 91. See too Argentina-France BIT 1991 Article 8(2): “Once the investor has submitted the dispute either to the jurisdictions of the Contracting Party involved or to international arbitration, the choice of one or the other of these procedures shall be final” cited in Schreuer above n.38 at 240; Spanish original is in ICSID Investment Treaties Vol. 3 (Dobbs Ferry, New York, Oceana Publications, Release 92–4, March 1993); U.S.-Uruguay BIT 2004 above n.5 Annex C (2): “For greater certainty, if an investor of the United States elects to submit a claim [under Section B, the investor-State dispute settlement provisions of the Treaty] to a court or administrative tribunal of

the framework of investment protection: the content of bits 69

of recent arbitral interpretation before ICSID.163 The key issue concerns whether the dispute submitted by the investor to the national courts or tribunals is one that invokes a claim of a breach of the applicable BIT, or whether it is an independent cause of action that raises no such issue.164 If the national claim is identical to the international claim made before the ICSID Tribunal, then the jurisdiction of the latter body has been excluded by the investor’s choice under the fork-in-the-road provision. But, questions that can only be determined under the applicable national law of the host country, such as whether a license has been properly refused or a tax properly charged, can only be considered before national courts or tribunals in the first instance. Such determinations do not, of themselves, raise any issues as to the breach of a BIT, or other applicable IIA. Accordingly, a choice of national tribunal by the investor in such a case will not operate to oust ICSID jurisdiction where the investor subsequently brings a separate claim alleging breach of the applicable BIT or IIA. The reference of an issue of national law to the relevant national dispute settlement body should not preclude a subsequent international claim being brought under the treaty for an alleged violation of its protection provisions, on the ground that such a claim relates to the investment protected under the applicable treaty and is separate from the underlying national legal dispute. In such a case there can be no fork-in-the-road, as there is no identity of subject-matter in the two legal proceedings. In addition, it is not open to the host country to avoid its responsibility under international law by relying on an exclusive jurisdiction clause in a contract with the investor if the national claim is not one based exclusively on the investment contract but also raises issues as to the conformity of governmental action with the BIT.165 A second factor that emerges from recent ICSID decisions is that the dispute brought before the national courts or tribunals, and the international claim, must be brought by the same claimant for the fork-in-the-road provision to preclude international jurisdiction.166 Thus separate national and international

Uruguay, that election shall be definitive, and the investor may not thereafter submit the claim to arbitration under Section B.” 163. See the discussion in Schreuer above n.38 at pp. 239–249. 164. See Genin v. Estonia ICSID Case No.ARB/99/2 award of 25 June 2001 available at www.worldbank.org/icsid/cases or 17 ICSID Rev-FILJ 395 (2002) at paras. 331–333. 165. See further Compania de Aguas Aconquija SA and Compagnie Generale des Eaux (Vivendi) v. Argentina ICSID Case No.ARB/97/3 award of 21 November 2000 available at www.worldbank.org/icsid/cases or 40 ILM 426 (2001) at paras. 53–55 upheld by the Ad Hoc Annulment Committee above note 171 at paras. 38–42, but see paras. 93–115 where the Committee concluded that having accepted jurisdiction over the claimant’s claims the tribunal manifestly exceeded its powers for failing to determine those claims on their merits. 166. See Lauder v. Czech Republic Final Award 3 September 2001 available at www. mfcr.cz/Arbitraz/en/FinalAward.doc. at paras. 162–163.

70 peter muchlinski

claims brought by different affiliates in a corporate group, or by the parent, or ultimate owner, and their affiliates, will not be regarded as definitive choices of national over international jurisdiction, binding all members of the group to that choice. Corporate groups will not be seen as a single entity for these purposes.167 This raises concerns that the group as a whole might seek to use its ability to bring multiple claims as a means of putting pressure on the host country to admit liability.168 On the other hand, different entities in the group will have different interests to protect. Thus the local affiliate is most likely to bring claims under national law relating to the application of legal rules and administrative decisions to the investment, while the parent, or ultimate owner, may be more likely to bring claims based on the BIT. To assimilate these parties and these claims may well deprive Multinational Enterprises (MNEs) of important legal rights that ensure the proper operation of the investment.169

conclusion The development of codified standards in IIAs has gone through a number of major phases, however, no agreed multilateral investment rules have resulted. Instead, there is a patchwork of voluntary non-governmental, and intergovernmental, codes and guidelines and legally binding BITs and other IIAs. The latter provide an increasingly settled content which has lead some to argue that these now represent customary international law.170 Though compelling, this view has itself been subjected to criticism, not least because such treaties are the outcome 167. See CMS Gas Transmission Company v. Argentina ICSID Case No.ARB/01/8 decision on jurisdiction 17 July 2003 available at www.worldbank.org/icsid/cases or 42 ILM 788 (2003) at para. 80. See too Azurix v. Argentina ICSID Case No.ARB/01/12 decision on jurisdiction 8 December 2003 available at www.worldbank.org/icsid/cases or 43 ILM 262 (2004) at paras. 86–90. 168. This point has been made in relation to the bringing of multiple claims against the Czech Republic by Netherlands based CME and its owner Ronald Lauder, however, in that proceeding the Czech Republic decided against arguing for a consolidation of the claims. See further “Who Wins and Who Loses in Investment Arbitration? Are Investors and Host States on a Level Playing Field?—the Lauder/Czech Republic Legacy” in 6 JWIT pp. 59–77 (2005); and Charles Brower and Jeremy Sharpe “Multiple and Conflicting International Arbitral Awards” 4 JWI 211 (2003). 169. See Schreuer above n.38 at 249. 170. See for example F.A.Mann above n.2; Patrick Juillard 1979 and 1982 above n.2. For a more recent restatement of this position see: S.Hindelang “Bilateral Investment Treaties, Custom and a Healthy Investment Climate—the Question of Whether BITs Influence Customary International Law Revisited” 5 JWIT 789 (2004); S. Schwebel, “The Influence of Bilateral Investment Treaties on Customary International Law” 98 ASIL Proceedings 27–30 (2004); Dominique Carreau and Patrick Juillard Droit International Economique (Paris, Dalloz, 2nd Ed, 2005) Part II Title II Ch.3.

the framework of investment protection: the content of bits 71

of individual bargaining relationships, which cannot act as a source of general legal obligation.171 Given this objection, the elevation of certain treaty-based standards of treatment to norms of customary international law may be mistaken. BITs and other IIAs constitute a special juridical regime designed to restate, in treaty form, international minimum standards of treatment of foreign investors as accepted by the capital-exporting countries, and to merge these with established, treaty-based standards of commercial conduct that do not possess the character of customary international law, despite their widespread usage over many centuries. The result is an integrated system of norms for the international regulation of investment relationships between, primarily, developed capitalexporting countries and less developed capital-importing countries, in a manner conducive to efficient capital accumulation by investors from the capital-exporting countries. Thus BITs and other IIAs represent a particular legal instrument seeking to solve a specific problem of international economic relations, and creating an effective lex specialis between the parties. They are not the equivalent of a codifying convention. On the other hand, it could be argued that the increased use of BITs between developing countries themselves172 shows that they have an impact upon the legal regime now favored by the traditional opponents of binding international minimum standards of treatment. Perhaps the key question is what might be gained by elevating treaty-based standards to customary law. In effect, it would bind all countries to what may remain contested international minimum standards of treatment, regardless of whether such countries have signed IIAs. This would prevent freedom of choice for countries as to the extent and nature of their commitments in relation to foreign investment law. Given the widespread application of otherwise contested standards as treaty based obligations, it would appear unnecessary to do so and, in this very sensitive policy area, it could produce an unfavorable political response which it is desirable to avoid.

171. See M. Sornarajah The International Law on Foreign Investment above n.132 at 204–217. 172. See World Investment Report 2003 above n.145 at p. 89.

This page intentionally left blank

3. explaining the popularity of bilateral investment treaties∗ andrew t. guzman introduction A serious analysis of bilateral investment treaties (BITs) and their implications both for investment levels and for the distribution of the gains from investment is timely. BITs have become the dominant international vehicle through which investment is regulated. As of 1996, there were 1,010 BITs in existence around the globe,1 more than half of which have been signed or brought into force since the start of 1990.2 The number of countries that have signed at least one BIT has reached 149 (including some countries that have ceased to exist, such as the USSR), leaving very few countries without such treaties. Although a substantial academic literature related to these treaties exists, there has been surprisingly little analysis of the impact of BITs on the welfare of the countries that have signed them. This chapter seeks to address this large gap in the literature and contribute to a more coherent understanding of BITs, their impact on foreign investment, and their effect on the welfare of nations. In recent years foreign direct investment (FDI) has grown at an unprecedented rate.3 Between 1986 and 1990, global FDI flows increased from $88 billion dollars to $234 billion, representing an average rate of increase of 26% in nominal terms and 18% in real terms. From 1980 to 1993 the stock of foreign investment increased at an average annual rate of 11% in real terms, reaching a total of $2.1 trillion in 1993.4 A significant proportion of FDI flows has been

∗ This chapter is an update of “Why LDCs Sign Treaties That Hurt Them: Explaining the Popularity of Bilateral Investment Treaties,” 38 Virgina J. Int’l Law 639 (1998), reprinted with permission from the Virginia Journal of International Law. The author is grateful to Karis Gong for assistance in preparing the chapter for publication in its current form. For comments on the original article, the author is indebted to Stephen Choi, Jack Goldsmith, Steven Ratner, Dhananjai Shivakumar, Douglas Sylvester, and Alan Sykes. 1. “Recent Actions Regarding Treaties to Which the United States Is Not a Party,” 35 I.L.M. 1130, 1130 (1996). 2. id. 3. The simultaneous growth in FDI and the popularity of BITs should not be assumed to imply causation. Although BITs and FDI are obviously related, the sensitivity of FDI to the presence of BITs is an empirical question that must be left for future research. 4. See Ernest H. Preeg, Traders in a Brave New World: The Uruguay Round and the Future of the International System 13 (1995).

74 andrew t. guzman

directed at developing countries: FDI flows to these countries grew from $13 billion in 1987 to $22.5 billion in 1989 to $90.3 billion in 1995.5 BITs rose to prominence during a period in which the international regulation of foreign investment was the subject of great change, uncertainty, and controversy. Not long ago, when a host country expropriated a foreign investor’s property, the relevant rule of customary international law, known as the Hull Rule, required “prompt, adequate, and effective” compensation. In the years that followed World War II, however, developing countries questioned the Hull Rule, claiming the right to determine how they would treat investors and the standard of compensation that should apply if that treatment was sufficiently harmful. This challenge to the Hull Rule proved successful and, by the mid-1970s, the Hull Rule had ceased to be a rule of customary international law. Countries began to establish BITs even before the demise of the Hull Rule. These treaties, typically signed between developed and developing nations, are binding international agreements that govern the treatment of foreign investment. Even though developing countries as a group objected vociferously and repeatedly to the Hull Rule, these same countries have signed more than a thousand BITs that incorporate obligations similar to the Hull Rule. Indeed, most BITs offer investors even greater protection, at the expense of host countries, than the Hull Rule ever did. Most important, BITs typically include terms that protect the foreign investor against a “contractual breach” by the host. Thus, when a BIT is in force between a host and a home country, an agreement made between the home country investor and the host is binding for both. A breach of the agreement by the host is a violation of the BIT and, therefore, a violation of international law. BITs also give the aggrieved investor access to binding arbitration, thereby creating an enforcement mechanism that is much more effective, and thus better able to ensure compliance by the host, than was the Hull Rule. This chapter looks at why BITs have become the preferred method of governing the relationship between foreign investors and host governments in developing countries. Because BITs impose obligations that are similar—and, indeed, that exceed—those imposed by the Hull Rule, and because the legal position advocated by developing countries has always been for fewer such legal requirements, the simultaneous opposition to the Hull Rule and embracing of BITs is a paradox.6 This chapter offers a novel explanation of why developing states

5. See Malcolm D. Rowat, “Multilateral Approaches to Improving the Investment Climate of Developing Countries: The Cases of ICSID and MIGA,” 33 Harv. Int’l L.J. 103, 103–04 (1992); Anthony M. Vernava, “Latin American Finance: A Financial, Economic and Legal Synopsis of Debt Swaps, Privatizations, Foreign Direct Investment Law Revisions and International Securities Issues,” 15 Wis. Int’l L.J. 89, 145 n.199 (1996). 6. I am not the first to note this paradox. See, e.g., M. Sornarajah, The International Law of Foreign Investment 259 (1994) (“This duplicity can be explained on the basis that while these states subscribe to a particular norm of international law at the global level, they are

explaining the popularity of bilateral investment treaties 75

fought aggressively against the former rule of “prompt, adequate, and effective” compensation for expropriation and in favor of a more lenient standard, and yet contemporaneously flocked to sign treaties that offer investors much greater protection than did the old rule of customary international law. It is demonstrated that although an individual country has a strong incentive to negotiate with and offer concessions to potential investors—thereby making itself a more attractive location relative to other potential hosts—developing countries as a group are likely to benefit from forcing investors to enter contracts with host countries that cannot be enforced in an international forum, thereby giving the host a much greater ability to extract value from the investment. Put another way, developing countries as a group have sufficient market power in the “sale” of their resources that they stand to gain more when they act collectively than when they compete against one another. This chapter offers a comprehensive explanation for the behavior of developing countries and assesses the desirability of BITs. In addition, this chapter discusses the welfare implications of BITs as compared to the “appropriate compensation” standard that developing countries have advocated at the UN. The chapter demonstrates that although BITs increase global efficiency, they likely reduce the overall welfare of developing countries. Finally, the chapter discusses the impact of BITs on customary international law. The chapter argues that because BITs are signed by developing countries in pursuit of their economic self-interest rather than out of a sense of legal obligation, these treaties do not support a rule of customary international law that incorporates the Hull Rule. Before proceeding, a note of clarification is in order. The bulk of the literature on BITs and foreign investment protection has focused on expropriation, devoting much less attention to other types of disputes between investors and hosts. The analysis here, however, encompasses all potential disputes between investors and hosts. Indeed, disputes that do not involve a direct taking are more interesting and relevant today because outright takings are now quite rare. The most common source of tension between an investor and a host country is not expropriation but rather conflicts that fall short of a taking. Customary international law—even under the Hull Rule—provides little protection for the investor against these less extreme actions by the host. BITs, on the other hand, allow potential investors to negotiate for whatever protections and safeguards they feel are needed. In other words, BITs provide the investor with protections that are superior, in all forms of investor-host conflicts, to those of customary international law. yet prepared to accord a higher standard of protection to the nationals of states with which they conclude bilateral investment treaties in the hope of attracting investment.”); Rudolf Dolzer, “New Foundations of the Law of Expropriation of Alien Property,” 75 Am. J. Int’l L. 553, 567 (1981) (“This apparent contradiction can be easily explained in light of the special benefits that developing countries enjoy under such treaties.”).

76 andrew t. guzman

A. The fall of the Hull Rule and the Rise of the BIT Early in this century, the world’s principal nations shared the view that investors were entitled to have their property protected by international law and that the taking of an alien’s property by a host nation required compensation that was “prompt and adequate.”7 Consensus surrounding the Hull Rule was possible during the first half of this century because many of the countries that later opposed the rule were then colonies rather than sovereign countries. Before decolonization, the official views of these countries were controlled by their colonial masters, who supported a regime of full compensation. Furthermore, colonies were not recognized as independent nations, suggesting that even if they had an independent, publicly stated view of how international law should protect investors, that view would not have affected customary international law. As former colonies became sovereign countries, however, these newly minted countries were able to voice their own views, and those views became relevant to the formulation of customary international law. As their numbers grew, these countries carried greater weight in the international arena, and as they questioned existing international norms, including the Hull Rule, the status of those norms was threatened.8 In the battle for international legitimacy, both sides of the debate claimed that customary international law was on their side. The developed world pointed to the history of the Hull Rule and to the support it had received both in practice and in writings by commentators. In response, least developed countries (LDCs) pointed out that practice had not always accorded with the Hull Rule and that, in any event, the rule simply lacked the broad international support that customary international law requires.9 Although the developed world denied the point, 7. See Concerning the Factory at Chorzów (Ger. v. Pol.), 1926–29 P.C.I.J. (ser. A), Nos. 7, 9, 17, 19, excerpted in Henry J. Steiner et al., Transnational Legal Problems 451–54 (1994). The Permanent Court of International Justice stated that “there can be no doubt that the expropriation . . . is a derogation from the rules generally applied in regard to the treatment of foreigners and the principle of respect for vested rights.” Id. at 452. The Court also stated that “reparation must, as far as possible, wipe out all the consequences of the illegal act and reestablish the situation which would, in all probability, have existed if that act had not been committed.” Id. at 453. For other early international cases articulating the rules governing expropriation, see Norwegian Shipowners Claims Arbitration (U.S. v. Nor.) 1 Rep. Int’l Arb. Awards 307 (1922) and Spanish-Moroccan Claims Arbitration (U.K. v. Spain), 2 Rep. Int’l Arb. Awards 615 (1925). 8. A rule of customary international law requires two elements: (1) the general practice of states, and (2) state adherence to the rule based on a belief that such adherence is legally required (opinio juris). 9. See Ebrahimi v. Islamic Republic of Iran, Iran Award 560–44/46/47-3, at ∗51–52 (Iran-U.S. Cl. Trib. Oct. 12, 1994) (WESTLAW, INT-IRAN Database) (“[W]hile international law undoubtedly sets forth an obligation to provide compensation for property taken, international law theory and practice do not support the conclusion that the ‘prompt, adequate and effective’ standard represents the prevailing standard of compensation. . . . Rather,

explaining the popularity of bilateral investment treaties 77

it seemed that the debate itself was undermining the claim that the rule retained its status as customary international law. Beginning in the early 1960s, LDCs found a forum in the United Nations from which to announce their views and, in doing so, further undermined the position of developed countries. From 1962 through the mid 1970s, the United Nations General Assembly—dominated by LDCs—passed a series of resolutions intended to emphasize the sovereignty of nations with respect to foreign investment. The relevance of these resolutions is not that they themselves announced or created a rule of customary international law. Rather, because a large majority of countries made it clear that they felt no legal obligation to follow the Hull Rule, the resolutions demonstrated that “prompt, adequate, and effective” was no longer a rule of customary international law.10 Moreover, the UN resolutions demonstrate that developing countries, acting as a group, prefer a regime under which they are able to expropriate property when they feel it is justified and under which they need only pay what they determine to be appropriate compensation. Once it became clear that the Hull Rule was no longer recognized, which certainly occurred in the wake of Resolution 3171 and may have occurred much sooner, neither the traditional “prompt, adequate, and effective” standard nor the “appropriate compensation” standard had enough international support to be considered a rule of customary international law. In the absence of BITs, therefore, developing countries had won a clear victory. The international rules governing North-South investment were entirely uncertain, and individual countries were in a position to determine what constituted appropriate compensation.11 In light of the considerable and long-term efforts by LDCs to defeat the Hull Rule, one might conclude that developing countries oppose any form of investment protection at the international level. That conclusion, however, is contradicted by widespread and enthusiastic LDC support for BITs. In a remarkably short period of time, BITs have become an important part of the foreign investment landscape. Between 1959, the year the first BIT was signed, and 1991, customary international law favors an ‘appropriate’ compensation standard.”); Sornarajah, The International Law of Foreign Investment, supra note 6, at 365–65 (“It is safe to conclude that there is no customary practice supporting the norm of full compensation for nationalisation.”); Dolzer, supra note 6, at 561 (“[Although] compensation must be paid for expropriated alien property as a matter of international law . . . the evidence for the Hull rule’s continuing validity falls short of the mark.”); Oscar Schachter, “Compensation for Expropriation,” 78 Am. J. Int’l L. 121, 123 (19-84) (“[T]he ‘prompt, adequate and effective’ formula has not won general acceptance in cases or state practice.”). 10. Put differently, the UN resolutions provide evidence of the demise of the Hull Rule, not of the rise of an alternative rule of customary international law. 11. See Dolzer, supra note 6, at 553 (“[T]he present state of customary international law regarding expropriation of alien property has remained obscure in its basic aspects.”).

78 andrew t. guzman

over 400 BITs were signed worldwide. More than ninety developing countries and most developed countries were parties to at least one such treaty during this period. In the 1990s the pace of BIT signings increased dramatically and by mid-1996, over one thousand BITs had been signed, with almost every country on the globe party to at least one such treaty. Whatever impact these treaties may have on customary international law, they represent an important part of the international investment landscape in their own right. Furthermore, BITs offer foreign investors greater protection than the Hull Rule ever did. They do so primarily by providing a mechanism through which a potential investor and a potential host can establish a contract that is binding under international law. In addition, the provision of dispute settlement procedures offers investors a neutral forum in which disputes can be heard and a means to enforce settlement decisions. The other provisions of BITs offer substantive protections such as national treatment, most favored nation treatment, free transfer of assets, and a prohibition on performance requirements. Finally, BITs reproduce the Hull Rule’s requirement of prompt, adequate, and effective compensation for expropriation, including “expropriations” that fall short of a direct taking. B. The Dynamic Inconsistency Problem 1. Dynamic inconsistency and FDI Before turning to an explanation of the behavior of developing countries, it is helpful to introduce a phenomenon know as the “dynamic inconsistency problem.” Dynamic inconsistency exists when a preferred course of action, once undertaken, cannot be adhered to without the establishment of some commitment mechanism. The problem is akin to wanting to “tie oneself to the mast” but being unable to do so. More formally, dynamic inconsistency describes situations in which a “future policy decision that forms part of an optimal plan formulated at an initial date is no longer optimal from the viewpoint of a later date, even though no new information has appeared in the meantime.”12 In the domestic setting, the dynamic inconsistency problem is avoided in most private transactions through contract. Parties are able to commit to a certain behavior because private contracts are enforceable under domestic law. This ability to contract, in turn, allows parties to negotiate, subject to transaction costs, the most efficient possible agreement. In the international setting, however, the dynamic inconsistency problem is a significant barrier to efficient foreign direct investment. The central problem is that a sovereign country is not able, absent a BIT, to credibly bind itself to a particular set of legal rules when it negotiates with a potential investor. Regardless of the assurances given by the host before the investment and regardless of the intentions of the host at the time, the host

12. Olivier J. Blanchard & Stanley Fischer, Lectures on Macroeconomics 592 (1989).

explaining the popularity of bilateral investment treaties 79

can later change those rules if it feels that the existing rules are less favorable to its interests than they could be. Domestic legal structures, critical to the credibility of contractual promises among private parties under domestic law, are no longer adequate to ensure compliance with the initial agreement. These risks are particular to the foreign direct investment setting because the host government is a direct participant and has interests and objectives of its own that may conflict with those of the investor. As a result of the dynamic inconsistency problem, when an investor enters into an agreement with a host nation, the two typically will not be able reach an optimal agreement. Notice that there need not be an intent to deceive on the part of the host. Even if the host wants to reach an efficient agreement and is willing to commit itself to a certain treatment of the foreign investment, it is unable to do so credibly because the host has the ability to later change its domestic laws to suit its own purposes. Because the foreign investor cannot rely on domestic laws to protect its interests, the only alternative legal structure is international law. International law, however, does not provide a way for a host country to make credible and binding commitments to an investor. The mechanisms for the enforcement of a contract between a country and a private firm is at best extremely weak and at worst completely non-existent. The precise status of such contracts is the subject of ongoing debate in the field of public international law and is far from being settled. For the present purposes, it is sufficient to note that there is no consensus that a contract with a host, by itself, offers a firm any additional protections under international law. Furthermore, even if protections exist in theory, the investor cannot be sure that they will be enforced by an arbitral tribunal or that the host will accept the decision of a tribunal if the firm obtains a favorable ruling. The fact that such agreements cannot be relied upon with any confidence implies that it is not possible for a country to make its commitment fully credible even if it enters into an agreement. The protections afforded to contract rights are so uncertain under international law that it is reasonable to model investor behavior on the assumption that these rights are of little or no value to the investor. More importantly, because these protections are unreliable, international law does not allow the host to make credible contractual commitments. This uncertainty explains why the debate over the protections afforded by customary international law was so important. Until the rise of BITs, there were few legal constraints, beyond those provided by customary law, on the behavior of host countries toward foreign investors. Thus, if the international community concluded that customary law did not require prompt, adequate, and effective compensation upon the taking of property, there would be no way for investors to achieve these protections. On the other hand, if the international community accepted the Hull Rule as international law, there would be no way for developing countries to except their own behavior from the rule.

80 andrew t. guzman

2. The behavior of host countries and investors To understand foreign investment in developing countries, one must consider how the lack of a credible contracting mechanism affects the incentives of a government in its dealings with a particular foreign investor. During negotiations, the government of a potential host country wishes to encourage the investor to invest. The firm, on the other hand, wants to achieve the greatest possible return and will invest in the host country only if that country offers conditions that will produce the greatest anticipated profit. If investor and host had the ability to credibly bind themselves to a particular set of conditions governing the investment, we would expect, subject only to transaction costs, the investor to select the most efficient location for its investment and to write a binding contract with that country. The agreement would spell out the conditions on which the investment would take place and would provide for some division of the “surplus” (i.e., profit) from the investment between the investor and the host. This division of surplus need not be stated explicitly, but it could take the form of concessions and commitments on the part of each party. For example, the host might agree to offer certain tax advantages to the investor, agree to allow the repatriation of profits, and waive certain import restrictions. The firm, on the other hand, might bind itself to providing a certain level of employment, certain transfers of technology, and so on. The absence of a credible contractual mechanism, however, makes the investment problem much more difficult. Even if an investment is valuable enough to make it worthwhile for the country to commit to certain concessions that benefit the investor—favorable tax treatment, for example—it cannot do so in a credible fashion. Once the investment is made, the host country no longer needs to offer benefits sufficient to attract the investment; it only has to treat the investor well enough to keep the investment. The difference between the two time periods (before and after investment) comes about because both the host and the investor know that once the firm has made its investment, it typically cannot disinvest fully.13 In other words, once it has invested, withdrawal would impose a cost on the firm. The host country can take advantage of this situation, and extract additional value from the firm by, for example, increasing the tax rate beyond the level that was agreed upon when the investment took place. Had the firm known that the tax rate would be higher than the agreed upon level, it may have chosen to invest elsewhere, or not to invest at all. Once the investment is made, however, 13. By this I mean that if the firm chooses to leave the country immediately after investing, it will not be able to recoup all of its investment. Indeed, it may be able to recoup only a very small fraction of its investment. This is, in part, because some portion of the investment (and perhaps a very large portion) is made in capital that can only be used in this one project. This can include specialized machinery, training of employees, and so on. For an empirical examination of the effect of irreversible investment on foreign direct investment, see Andrew Guzman & Aart Kraay, “Uncertainty, Irreversibility, and Foreign Direct Investment” (1996) (ch. 2 of unpublished Ph.D. dissertation, Harvard University) (on file with author).

explaining the popularity of bilateral investment treaties 81

it may be cheaper for the firm simply to pay the higher tax rather than to disinvest and reinvest elsewhere. The problem may even be worse because the host can impose any level of tax (or other policy to take value from the firm) it chooses, as long as the tax is less than the amount that it would cost the firm to disinvest. Most importantly, the host can assess the firm’s situation and select the maximum possible transfer of value that the government can demand without driving the firm out of the country altogether. Extracting value from the firm by increasing the tax rate or otherwise changing the conditions under which the firm will operate represents only one of the options available to the host once the investment is made. The other two options are to abide by its original promises to the investor or to expropriate the investor’s property. There are two categories of costs facing a country that chooses to expropriate outright. The first is that the government (or the private parties to whom the government gives or sells the enterprise) may be far less competent to run the facility than the original firm. After expropriation, the firm’s managers are likely to leave the country, taking a substantial amount of human capital with them and making it difficult for the host government to run the business as well as the investors who built the facility. This will impact both the profits that the enterprise will generate (if any) and the level of spillover benefits provided (employment, technology transfer, etc.). A second cost is to the country’s reputation. The firm whose assets have been taken will undoubtedly complain to its home country, and that country may take action. Indeed, if the expropriation is severe enough, even countries whose nationals have not been affected may sanction the host country. In addition, the expropriation will be noticed by other firms which, as a result, may be hesitant to invest in that country in the future. Because the costs of outright expropriation are likely to be high, the more moderate course of extracting value from the firm without forcing divestment, as discussed above, may be attractive. This can be done in a variety of ways, including changing the tax rate, restricting the repatriation of profits, imposing new labor or local content requirements, and so on. This approach allows the country to take advantage of the existing management and their skills, thus avoiding the major costs of an outright expropriation, while still extracting value from the enterprise. This more moderate approach may also be preferred because it is less likely to provoke significant sanctions by the home country of the investor. After all, the firm’s assets have not been seized and it is often difficult to identify where the right of a government to set policy crosses over into unreasonable conduct. For any irreversible investment,14 then, the host country will be able to demand a higher payment after the investment takes place than it could have demanded before investment of the capital. This is so because the investor will invest only

14. An investment is irreversible for our purposes if withdrawal from the investment yields less than the full value that was invested.

82 andrew t. guzman

if it expects to receive revenues that are greater in present value than its total costs. Before the investment, total costs include all expected costs of the investment, including irretrievable capital costs. Once the original investment has been made, however, the investor will not include the sunk (irretrievable) costs in its calculation because those funds are lost regardless of its actions. Once the investment is made, therefore, hosts may extract at least up to the value of the sunk costs without making it profitable for the firm to withdraw. Of course, the firm understands the impact of the dynamic inconsistency problem before it invests. It may, as a result, choose not to invest. The potential host, on the other hand, wants the investment to take place, and to get the investment it would be willing to bind itself to a set of conditions on which the investment would take place. Because there is no binding commitment mechanism available, however, the host cannot make a credible commitment and the investment—desired by both parties—may not take place. In light of the above discussion, one might ask why there is any direct foreign investment at all and why the investment that does takes place is often treated well. The reason is that the above description is based on a single investment decision. In actual fact, countries want to maximize their returns over longer horizons. Thus, they may resist the temptation to seize assets today in order to create or maintain a reputation that will attract future investment. Countries may also resist the temptation to extract value from foreign firms if they fear that sanctions will be imposed by the home country of the investor. The long-term effects of individual investment decisions, in other words, change the incentives of the host country. These effects, however, do not completely remove the dynamic inconsistency problem. When the host country considers the reputational effect of its actions, it will weigh the gains from breaching its agreement with the firm against any lost benefits caused by reduced efficiency within the firm, sanctions imposed by other countries, and the effect of the action on its reputation. A priori, there is no reason to think that this balancing establishes an efficient set of incentives for the host. Indeed, the fact that host countries sign agreements with investors, even when those agreements are not enforceable, indicates that the parties do not believe that reputation, by itself, is sufficient to give the host the proper incentives. The contracts represent an attempt to increase the cost to the host of violating the terms under which investment takes place. The foregoing discussion has considered the impact of the dynamic inconsistency problem on the host country. The effect of the dynamic inconsistency problem must also be considered from a global perspective. In global terms, the most efficient outcome is achieved if investment takes place where it will earn the greatest total return. Absent transaction costs, this outcome is achieved when the parties are able to contract with one another and when a breach of contract is accompanied by expectation damages. The dynamic inconsistency problem, however, undermines efficiency because it discourages investment that would

explaining the popularity of bilateral investment treaties 83

be desirable. Firms realize that host countries have incentives to squeeze additional value from their operations after the investment is made, and this causes firms to avoid some investments altogether. Rather than facing expectation damages if it breaches, the host faces a penalty in the form of lost future investment and sanctions from foreign countries. It would be mere coincidence if these sanctions were equivalent to expectation damages, implying that the decision to breach the contract with the investor will not be based on appropriate incentives. C. Explaining the Paradoxical Behavior of LDCs As discussed above, the behavior of developing countries presents an apparent inconsistency. On the one hand, they have repeatedly sought to establish a norm that leaves significant power in the hands of the sovereign country in its relations with investors, makes it difficult for countries to enter into binding contracts with foreign investors and, therefore, leaves the dynamic inconsistency problem unresolved. On the other hand, developing countries have willingly and, indeed, enthusiastically, signed BITs with developed countries. These bilateral treaties undermine precisely the independence and control that the countries have fought so hard to protect. This part of the chapter considers and rejects the possibility that developing countries have simply changed their views on the subject or that, in exchange for signing BITs, LDCs have received concessions that they did not receive under the traditional standard of full compensation. Another explanation is then advanced, namely, that LDCs face a prisoner’s dilemma in which it is optimal for them, as a group, to reject the Hull Rule, but in which each individual LDC is better off “defecting” from the group by signing a BIT that gives it an advantage over other LDCs in the competition to attract foreign investors. 1. Existing explanations of LDC behavior One possible explanation of the behavior of LDCs is that the developing countries themselves have come to conclude that they are better off if they allow themselves to be bound through a contractual mechanism with investors. This might be termed the “LDC enlightenment theory.” For a period of time after World War II developing countries fought to defeat the Hull Rule and reduce the protections provided to international investment. More recently, however, these same countries have come to realize that it is in their interest to encourage foreign investment in their country and that one way to do this is to provide strong protections for foreign investment. Developing countries may also have developed a better understanding of the dynamic inconsistency problem and its importance in the foreign investment realm. As a result, they are now prepared to accept BITs because they appreciate the need to overcome this problem. This theory is unsatisfactory for at least two reasons. First, the period in which BITs have been signed has overlapped considerably with the period in which LDCs sought to discredit the Hull Rule. The first BIT was signed in 1959, when West Germany established a treaty with Pakistan, and by the mid-1970s, West Germany

84 andrew t. guzman

had concluded more than forty BITs with other countries. The major efforts to undermine the Hull Rule at the multilateral level took place during the 1960s and 1970s. In other words, during the very period when the General Assembly was denouncing the Hull Rule, large numbers of developing countries were signing bilateral treaties. If developing countries truly had changed their views on the value of commitment mechanisms and binding agreements, we would not expect to see a significant number of BITs in force and more being negotiated at the same time that the General Assembly voted 108 to 1 in favor of the 1973 Resolution on Permanent Sovereignty over Natural Resources and adopted the New International Economic Order. Nor can it be argued that the countries signing BITs and those fighting against the Hull Rule are different subsets of countries because both the BIT movement and the movement against the Hull Rule have included a majority of developing countries. Furthermore, had developing countries decided, as a group, that providing greater protections for foreign investors served their interest, one would expect them to express that view at the General Assembly. The choice of the General Assembly as the forum for previous resolutions concerning investment demonstrates that developing countries have found it to be a useful and accessible forum. Moreover, if LDCs had truly changed their minds, the best way to demonstrate that they no longer held the views on investment expressed in the General Assembly resolutions would be to announce their new views in the same forum. One would also expect developing countries to have signed multilateral investment treaties rather than bilateral treaties. An alternative theory is presented by M. Sornarajah, who argues that developing countries, after successfully tearing down the Hull Rule, adopted BITs in reaction to the confused status of foreign investment and the uncertain protections afforded to it by international law: “knowing the confused state of the law, [countries] entered into such treaties so that they could clarify the rules that they would apply in case of any disputes which may arise between them.”15 If the goal of BITs was to clarify existing rules, however, there is no reason for them to provide so much protection to investors. If LDCs believed that international law offered relatively weak protections for foreign investment—as they indicated in the General Assembly resolutions—they could “clarify” such a rule. BITs, however, offer much more protection for investment than any rule to which developing countries have publicly subscribed. Furthermore, it is difficult to understand why LDCs would undermine the Hull Rule—which provided a clear rule regarding the protection of foreign investment—only to adopt BITs to avoid the legal ambiguity generated by the demise of the Hull Rule. Rudolf Dolzer advances a different explanation. He claims that developing countries are prepared to accept the Hull Rule in the context of BITs because of

15. Sornarajah, The International Law of Foreign Investment, supra note 6, at 233.

explaining the popularity of bilateral investment treaties 85

the “special benefits that developing countries enjoy under such treaties.”16 Sornarajah appears to hold this view as well, arguing that although a developing “state subscribes to a particular norm of international law, it is prepared to treat the nationals of a state with which it has entered into a bilateral treaty in accordance with the norm which has been agreed to in the treaty.”17 This view is difficult to reconcile with the content of most bilateral treaties. There is little in such treaties that inures to the benefit of the host countries apart from the benefits that those countries enjoy from a regime of investor protection. If the benefits of investor protection are sufficient to make these treaties desirable to LDCs, of course, it becomes impossible to explain why the countries sought to undermine investor protection when they dismantled the Hull Rule. 2. A strategic analysis of LDC behavior To understand the apparent paradox of the LDC struggle against the Hull Rule as customary international law and the simultaneous embracing of BITs that mandate even stricter investment protections, one must realize that developing countries have different interests when they behave as a group than they do when they behave individually. In other words, the decision of individual countries to sign bilateral agreements is not a sign that these agreements are in the interest of LDCs as a group, and the efforts of LDCs as a group to defeat the Hull Rule do not imply that an individual country would not want to embrace the rule for its own purposes. a. The interests of an individual LDC. Consider first the incentives facing an individual developing country. In its negotiations with investors, the country would like to have the ability to make binding commitments to potential investors. If it is able to make credible commitments, it will be able to attract more investors. Specifically, the country may be able to attract investors who, absent a commitment to lower taxes, for example, would choose a different country for their investment. Furthermore, for those investors who would invest in the country even in the absence of a commitment mechanism, the country can simply decline to offer more favorable conditions, thus getting those investors to invest on the same terms as would exist without the ability to commit. The developing country, therefore, has a strong incentive to enter into BITs to increase the investment it receives and, thereby, increase the benefits enjoyed by the country. Put simply, by entering into a BIT, a country is better able to compete for investment. In practical terms, this implies that if a single LDC is offered the opportunity to enter into a treaty that will allow it to make binding commitments to investors without affecting the ability of other LDCs to do so, it will have a strong incentive to sign such a treaty. The amount of extra investment that can be attracted to a country that is able to enter binding contracts depends on the sensitivity of the demand for the

16. Dolzer, supra note 6, at 567. 17. Sornarajah, State Responsibility and Bilateral Investment Treaties, supra note 6, at 90.

86 andrew t. guzman

resources of that country (raw material, labor, government regulations, location, etc.). If the market for those resources of the country is highly competitive (i.e., characterized by many buyers and many sellers), even a relatively small improvement in the conditions offered to potential investors will lead to a large increase in investment. As in any competitive market, a small change in the price of the goods being sold will lead to a large increase in demand. In the foreign investment context, the goods being sold are the resources of the LDC, and the “price” at which investors can get access to those resources will fall as investors are offered more attractive conditions by the potential host. If the market is competitive, therefore, the ability to commit to a binding contract allows a country to increase dramatically the amount of investment it receives. It is important to note, however, that much of this increase in investment will come at the expense of other developing countries. If other LDCs have not signed such treaties, a country that does sign one will gain an important advantage, and if other countries have already signed BITs, a country that signs one will eliminate the advantage those other countries had in the competition for foreign investment. Thus, regardless of what other countries are doing, a developing country has a strong incentive to be enthusiastic about signing a BIT. Based on the above discussion, one might conclude that a regime that allows developing countries to contract with investors is preferred to the system advocated by developing countries in their UN resolutions. A contracting regime is preferred, the argument would go, because it increases the number of efficient investments, which in turn leads to greater global wealth. Furthermore, a contracting regime is in the interest of developing countries because it allows them to offer incentives that will increase investment and well-being in their countries. The contention is that because an individual country is able to attract investment more successfully when it can make binding commitments, the ability to make such commitments must be good for LDCs as a group. As argued below, however, the conclusion that LDCs are better off as a group simply does not follow from the fact that individual LDCs benefit from a contracting regime. b. The interests of LDCs as a group. To understand the incentives and interests of LDCs when they act as a group, imagine a scenario in which two countries are competing against each other to attract a potential investor. Assume that both countries have signed a BIT with the home country of the investor. To attract the investment, each country is willing to make concessions to the potential investor. A country whose initial offer is insufficient to attract the investment has an incentive to increase the concessions it offers as long as the benefits of the investment to the country exceed the costs of the concessions. The result, therefore, is a bidding up of the concessions made to the investor. Ultimately, if the market for the resources of the developing countries is competitive, the potential hosts will continue to bid against one another until the benefit enjoyed by the host from the investment is zero. Only then will the country that stands to lose the investment find it impossible to offer the firm a more attractive package. Once the offers to

explaining the popularity of bilateral investment treaties 87

the firm have been bid up to the point where the winning country stands to make no net gain from the investment, the firm does not have to share any of the surplus with the host and can, therefore, simply choose the location that offers the highest overall return. This is the efficient result because all investments that offer a positive expected net present value can be made and all investments will be made in the country where they offer the highest return. The impact of this bidding contest on the distribution of the gains from an investment project is dramatic. The country that receives the investment will have won the competition to attract the capital, but will gain little or nothing from its victory. The benefits to the country generated by the investment (in the form of employment, technology transfers, tax revenues, and so on) will be offset by the incentives and concessions that were needed to attract the firm (tax breaks, reduced pollution controls, relaxed employment regulations, and so on). In other words, as in any competitive market, the seller—here the host country—will receive no economic profit. The entire profit will be enjoyed by the investor. In the presence of BITs, then, potential host countries will bid down the conditions on which they allow investment in an attempt to attract as much investment as possible. Ultimately, the concessions extended to investors may be so great that countries will be indifferent between having and not having the investment. The competitive nature of the market means that the benefits of investment will all go to the investor, leaving no surplus for the host. Contrast this result with the result under the Charter of Economic Rights and Duties of States (CERDS) regime. Imagine two countries competing for a potential investor, but without any way for either country to make a binding commitment. In this situation, the investor cannot obtain any credible guarantees regarding the treatment of its investment. The investor may still decide to invest, however, because the countries in question have reputational concerns that encourage them to treat investment well. Moreover, the investor can also take steps to protect itself by, for example, entering into a joint partnership with the host (so that the host has a strong incentive to let the investor maximize profits), placing a few critical operations abroad (so that the host will gain little by expropriation), or demanding a signed agreement which, although not binding under international law, may cause the host international embarrassment if it treats the investment poorly. Most importantly, the investor may choose to invest without any binding commitments from the host country because LDCs offer advantages that are unavailable in the investor’s home country (e.g., low labor costs, favorable environmental or labor laws, locational advantages, natural resources, and so on). Even though the investor lacks the protections of a BIT, it may still be worthwhile to invest. If the investment takes place in the absence of a BIT, the country that receives the investment will be able to extract value from the investor because the host has the power to unilaterally change the conditions under which the firm operates. The firm’s only defenses are the ability to pull its operations out of the country

88 andrew t. guzman

and the reputational concerns of the host. The value extracted will depend on the reputational concerns of the country, the value that is available, and the success of the investor’s efforts to make such value extraction difficult. In any event, the host will gain more from each dollar invested than it would in a world of BITs because once the investment is made, the host can extract value without losing the investment. Thus, under the CERDS regime, hosts get more value from each investment. The disadvantage of CERDS, however, is that there will be fewer investments because the inefficiencies of the regime make it more costly to invest. Some investments that would be profitable under a BIT regime will no longer be profitable under CERDS. These investments will never be made, and developing countries will lose the benefits associated with them. Obviously, if the level of investment dropped below a certain point, LDCs would be worse off as a group under the CERDS regime than they would be under a BIT regime. On the other hand, if there is only a small reduction in the overall level of investment, LDCs may be better off under CERDS because they can receive a larger share of the return from investments. In determining which regime benefits LDCs most, it is crucial to determine how much investment will be lost under CERDS. The critical issue is the sensitivity of investment to the cost of investing. If the investment into LDCs taken as a group is sufficiently insensitive to the cost of investment, then LDCs as a group would be better off under the CERDS regime than under a BIT regime. This analysis explains the efforts of LDCs, acting as a group in the General Assembly and elsewhere, to undermine the Hull Rule. In the debate over the status of the Hull Rule as customary international law, developing countries were working to change a rule that applied to them all. Attempts to undermine the Hull Rule, therefore, were attempts to change the rules that applied to LDCs as a group. Even though individual countries have been eager to sign BITs, LDCs as a group may be better off in a regime that leaves them unable to enter binding contracts with investors. Thus, the incentives for an individual country and for LDCs as a group are different. This difference arises because developing countries compete among themselves for a limited pool of investment. As they compete, they bid away some of the surplus they would otherwise enjoy, and this lost surplus may in fact exceed the gains from new investment (i.e., from investment that would not otherwise have been made in any developing country). The net result is that whereas individual LDCs may be better off vis-à-vis other LDCs in a BIT regime, LDCs as a group may suffer an overall welfare loss. 3. The outstanding empirical question The above discussion offers a novel explanation of LDC conduct. For the theoretical explanation to be correct, however, an empirical claim about foreign investment must be made. Specifically, the above theoretical claims are true only if the flow of investment into LDCs as a group is relatively insensitive to the terms on which that investment is made as compared to the flow of investment into a single developing country.

explaining the popularity of bilateral investment treaties 89

In economic terms, the demand for the resources of LDCs as a group must be relatively inelastic while the demand for the resources of a single country must be relatively elastic. Ultimately, this is an empirical question that cannot be answered without further research. Although it is not possible to demonstrate that the empirical conditions assumed by this theory exist without a formal empirical study, it is possible to show that it is reasonable to assume the existence of these conditions. For investment flows into developing countries as a group to be less sensitive than flows into a single developing country, it must be that a developing country is more likely than a developed country to be a substitute for another developing country. In other words, at the margin, more investors will switch from one developing country to another in response to a change in costs in one developing country than will switch from developing countries to developed countries in response to a change in cost in all developing countries. Although developing countries and developed countries do share certain characteristics, there are enough distinct traits of developing countries to support this assumption. For example, labor in developing countries is often extremely inexpensive relative to developed countries. Thus, the threat of an increase in the wage rate in an LDC may not deter an investor because even if there were a substantial increase in the cost of labor, it is likely to remain below that of the developed world. Similarly, developing countries have natural resources that do not exist in developed countries, or that are not as abundant and inexpensive. In addition, the legal and regulatory climate of many developing countries may be more advantageous for investors. For a single country, it is reasonable to assume that the foreign investment decisions regarding investment in that country are relatively sensitive to the cost of investing (i.e., the elasticity of demand for the resources of the LDC is high). This is because developing countries must compete against one another for investment and, as the cost of investing changes, so may the choices of investors. If the cost of investing increases, for example, the potential investor can simply invest in a different developing country. Similarly, if the cost of investing is reduced, investment that would have gone to a different country may be attracted. For a single country, therefore, the increased cost of investing prompted by the dynamic inconsistency problem (as compared to a BIT regime)—holding conditions in other LDCs constant—will cause a relatively large reduction in the total amount of foreign investment. Investment will simply move to a developing country that can make a binding agreement. For developing countries as a group, however, the sensitivity of investment demand is likely to be much lower. If the cost of investment rises in all developing countries, an investor must either invest despite the increased cost or abandon its intention to invest in a developing country. The advantages offered by one developing country are much more likely to be found in another developing country than in a developed country. Put differently, it is reasonable to assume

90 andrew t. guzman

that, in the eyes of investors, developing countries are more like one another than they are like developed countries. Thus, investment will be much less sensitive to the cost of investing (i.e., the elasticity of investment will be lower) when we consider LDCs as a group rather than individually. In addition to the fact that the empirical assumptions necessary to support the theory advanced in this chapter are plausible, those assumptions are also supported by the observed behavior of developing countries. No other theory has been advanced that is capable of explaining why developing countries simultaneously opposed the Hull Rule and embraced BITs. If the empirical assumptions underlying the theory of this chapter are incorrect, these actions by LDCs are irreconcilable. 4. An economic interpretation. A fundamental insight that drives the results found in this chapter is the recognition that the presence or absence of a credible mechanism for contracting changes the competitiveness of the market for foreign investment. If it is possible to make credible commitments through contracts, every potential host country must compete for the investment, leading to a competitive market for the resources of those countries and, therefore, zero economic profit (or, at least, low profits) for the “seller” (i.e., the host). As in any competitive market, the seller must compete for business, and the buyer—here the investor—receives the entire surplus from the transaction. If LDCs can act as a group, however, there is less competition. Imagine, for example, explicit collusion among all developing countries that is aimed at increasing the rents earned from the sale of their resources to investors. If that collusion was successful, one would expect LDCs as a group to have some market power and, therefore, to be able to increase the price at which investment takes place and to extract some of the surplus of the transaction. The collusion, of course, will cause a reduction of the overall level of investment, but the gains from colluding would outweigh the losses to LDCs. The host is able to extract rents because once the investment is made, the host is in the position of a monopolist. It can choose to set the price of its resources at the level that maximizes its own return. The basic theory of monopoly pricing teaches that a monopolist will set a price that is above the competitive price in order to extract monopoly rents. The result, of course, is a reduction in the demand for the resources, a loss to the buyer (here, the investor), and increased profits. Overall, there is a net loss, referred to as a “deadweight loss.” In the context here, the host will demand more value from the investor than it would in a competitive environment. Thus, to the extent potential hosts compete against one another for investment, and to the extent that this leads to a competitive market for the resources offered by potential hosts, it should be expected that hosts will seek to extract value from the firm after the investment takes place. Because the investor has made an irreversible investment, it cannot easily withdraw from the country, making its demand for those resources very inelastic indeed. Collusion among LDCs would still leave LDCs as a group in competition with developed countries. Developing countries would only be able to extract rents

explaining the popularity of bilateral investment treaties 91

from investors if they were able to obtain some market power despite the presence of developed countries. Whether or not they are able to do so is the empirical question discussed in section D.3 above, where it is suggested that such market power is in fact plausible. All that remains, then, is to explain the relationship between explicit collusion and the demise of the Hull Rule. By ending the Hull Rule, developing countries eliminated the rule of customary international law that required them to pay full compensation for “takings.” In the absence of any other international law or treaty, no mechanism existed for a host to commit to an investor in a credible fashion. Thus, the legal regime ensured that no country could bind itself to a certain standard of treatment before investment; only reputational constraints controlled the behavior of LDCs. Without the ability to obtain a credible commitment from host countries and without a rule of customary international law protecting investment, investors faced a higher expected cost of investment because of the dynamic inconsistency problem. The effect, therefore, was the same as an explicit agreement among all developing countries that they would not bid against one another for investment. In fact, without BITs, the regime was even better for LDCs than a collusive agreement because it was impossible for any single country to “defect” from the agreement; thus, the “cartel” of LDCs was extremely stable. The above explains, in economic terms, why developing countries fought to undermine the Hull Rule. Their willingness to sign BITs, as explained above in section C.2, was caused by the fact that BITs introduced a mechanism through which developing countries could compete for investment. Just as members of a cartel may seek to defect from the cartel to increase their sales, individual LDCs embraced BITs as a way to compete for foreign investment. D. Efficiency, welfare, and international law 1. Efficiency implications Under most BIT arrangements, contracts between investors and host countries are binding at the international level. This binding contractual mechanism of BITs is made possible by the dispute settlement procedures. Failure to respect the terms of a contract with a foreign investor is a violation of the BIT and gives the investor the right to pursue a remedy through a dispute settlement procedure, which in most circumstances is binding arbitration. To the extent that the damage scheme under a BIT is interpreted as expectation damages (as opposed to, say, restitution damages), the efficient outcome is achieved. Moreover, even if the measure of damages is not expectation damages, the BIT regime is more efficient than either CERDS or the Hull Rule because it governs a wider range of potential host-investor disputes (i.e., it applies to more than direct expropriation).18 18. Although rarely clear on the question, most commentators appear to view the “prompt, adequate, and effective” standard—present in both BITs and the Hull Rule— as a form of expectation damages. See, e.g., Brice M. Clagett, “Protection of Foreign

92 andrew t. guzman

More important than the measurement of damages, however, is the fact that the BIT framework, by providing a binding and credible contractual mechanism, allows the parties to avoid the dynamic inconsistency problem. The presence of an impartial dispute settlement mechanism that is capable of ensuring compliance by the host helps ensure that host governments will honor their agreements.19 Subject only to transaction costs, a BIT regime will cause capital to be invested where it stands to earn the greatest return. Thus, the cost of investing is reduced, more investment will take place, and the investment that does occur will be allocated in an efficient manner. BITs, therefore, yield an efficient allocation of capital. In contrast to a BIT regime, the rules of CERDS introduce a significant degree of inefficiency. Under CERDS, the security of an investment is dependent on the goodwill of the host country. Outside the discipline provided by the market for foreign investment, investors enjoy little protection against actions by the host. Although these reputational concerns may provide nontrivial protection, investors still have cause to be concerned about expropriation and other less dramatic actions by the host country because there is no reason to think that reputational concerns are enough to cause hosts to honor all commitments. The lack of a credible commitment mechanism, in turn, drives up the cost of investment and causes profitable investments, which both the host and the investor desire, to be foregone because they are rendered unprofitable by the dynamic inconsistency problem. In addition to reducing the amount of investment, the CERDS regime may distort decisions regarding where to invest. Imagine, for example, that two countries wish to attract a particular investment. Country A may be the better location for the investment because of, say, its geographic location and the available social infrastructure. Nevertheless, the investor may decide to invest in country B because country B is considered more likely to honor the agreement under which the investment takes place. This is an inefficient result because the investment would be, by assumption, more valuable in country A. If it were possible for country A to write a binding contract with the investor, the distortionary effect of reputation would be eliminated and the investment could be made efficiently. A BIT arrangement, therefore, is better on efficiency grounds than a regime based on CERDS. Investment Under the Revised Restatement,” 25 Va. J. Int’l L. 73 n.8 (1984) (“The standard method of establishing [adequate compensation] is called discounted-cash-flow analysis.”); Schachter, supra note 9, at 124–25 (stating that the Hull standard was “full value,” i.e., fair market value); World Bank Guidelines on the Treatment of Foreign Direct Investment, reprinted in Ibrahim F.I. Shihata, Legal Treatment of Foreign Investment: The World Bank Guidelines 155, 161, 163 (1993) (defining “adequate” to mean fair market value, and “effective” to mean in a convertible currency). 19. More precisely, the agreement will be honored or the investor will be able to recover its losses.

explaining the popularity of bilateral investment treaties 93

2. Welfare implications The efficiency of BITs, however, is not the whole story. As compared to CERDS, BITs sharpen the competition for investment among potential hosts. This forces LDCs to offer greater and greater concessions to potential investors, bidding away the gains the host would otherwise enjoy. In effect, BITs make the market for foreign investment much more competitive by allowing competition in the price of investment, that is, the terms under which investment takes place. In the absence of BITs, international law currently yields the same economic result as would an agreement among developing countries to never negotiate with potential investors before the investment. Such collusion would force investors to either invest without knowing the final terms under which they have to operate or to refrain from investing. The practical effect is to increase the price at which the resources of developing countries are sold. This in turn reduces the amount sold and, assuming investment into LDCs as a group is not overly sensitive to changes in the terms of investment, the result is monopolistic gains for developing countries. The CERDS regime, in other words, makes the market for foreign investment and LDC resources imperfectly competitive, allowing developing countries (the sellers) to capture a larger share of the rents. Capital importers, therefore, are better off as a group under the CERDS regime than under the BIT regime. Just as a monopolist (or an oligopolist) enjoys an increase in welfare when it is able to reduce the competitiveness in a market, so LDCs enjoy greater returns under CERDS because this regime makes the market for foreign investment less competitive. From the point of view of the welfare of developing countries as a group, the best of the three possible regimes is CERDS, followed by the Hull Rule (which only covers direct expropriation), followed by the BIT regime. The BIT regime is the least beneficial to LDCs because it includes the most expansive definition of investment and thereby allows greater competition among developing countries. The purpose of this chapter is not to advocate one of these systems over another. Rather, it seeks simply to show the relevance of the distributional issues. Without a mechanism to redistribute wealth between countries, the distributional consequences of a particular policy should be considered. The rise of BITs has reduced the market power held by developing countries, which, in turn, has reduced the benefit these countries can capture from any particular investment. For this reason, the BIT regime may actually reduce the overall welfare of developing countries and therefore should not be uncritically embraced by those who seek the interests of LDCs. On the other hand, there can be no serious doubt that, from a global efficiency perspective, a regime that allows for contracting between host governments and investors is more efficient than a regime in which potential hosts cannot effectively commit to any particular behavior or agreement. 3. The impact of BITs on customary international law This chapter has sought to explain the paradox of LDC objections to the Hull Rule and the widespread

94 andrew t. guzman

adoption of BITs. In addition, the chapter has pointed out the ambiguous effect of these treaties on the welfare of LDCs. The explanation of BITs presented here also allows the analysis of another issue that is attracting some attention, namely, the role of BITs in the establishment of customary international law. Do these treaties codify an agreed upon set of principles that applies to all, or do they merely represent lex specialis as between the parties?20 The debate is of some importance because if the BITs establish a rule of customary law, that law will apply to all countries even in situations where the host has not signed a BIT with the home country of the investor. Those who argue that the BITs represent the codification and entrenchment of customary principles of international law point to the large number of such treaties and the fact that many of the nations that rejected the traditional Hull Rule standard of compensation have signed BITs. The prevalence of BITs, the argument goes, demonstrates that both developed and developing countries now consider the traditional compensation standard to be the relevant standard of international law. As one commentator has put it, “Is it possible for a State to reject the rule according to which alien property may be expropriated only on certain terms long believed to be required by customary international law, yet to accept it for the purpose of these treaties?”21 The analysis presented in this chapter shows that this question can and must be answered in the affirmative. The arguments of those who view BITs as evidence of customary law are flawed for two reasons. First, as the above quote illustrates, these arguments overlook the fact that the Hull Rule ceased to be a rule of customary law sometime before 1975. Second, the arguments fail to take into account that customary law requires not only practice, but also a sense of legal obligation. As the International Court of Justice has made clear, Not only must the acts concerned amount to a settled practice, but they must also be such, or be carried out in such a way, as to be evidence of a belief that 20. For a detailed argument that BITs do not contribute to the formation of customary law, see Bernard Kishoiyian, “The Utility of Bilateral Investment Treaties in the Formulation of Customary International Law,” 14 Nw. J. Int’l L. & Bus. 327, 329 (1994) (“[E]ach BIT is nothing but a lex specialis between parties designed to create a mutual regime of investment protection.”). For the opposing view, see Asoka de Z. Gunawardana, “The Inception and Growth of Bilateral Investment Promotion and Protection Treaties,” 86 Am. Soc’y Int’l L. Proc. 544, 550 (1992) (“Although the provisions of the bilateral investment promotion and protection treaties may not have attained the status of customary international law, they have an undoubted part to play in that regard.”); David R. Robinson, “Expropriation in the Restatement (Revised),” 78 Am. J. Int’l L. 176, 177 (1984) (“[M]any of the same developing nations that supported these [United Nations] declarations as political statements have, in their actual practice, signed bilateral investment treaties reaffirming their support for the traditional standard as a legal rule.”). 21. F.A. Mann, “British Treaties for the Promotion and Protection of Investments,” 52 Brit. Y.B. Int’l L. 241, 249 (1982).

explaining the popularity of bilateral investment treaties 95

this practice is rendered obligatory by the existence of a rule of law requiring it. . . . The States concerned must therefore feel that they are conforming to what amounts to a legal obligation. The frequency, or even habitual character of the acts is not in itself enough.22 Therefore, in the words of another scholar, “The repetition of common clauses in bilateral treaties does not create or support an inference that those clauses express customary law. . . . To sustain such a claim of custom one would have to show that apart from the treaty itself, the rules in the clauses are considered obligatory.”23 The relevant inquiry concerning whether BITs establish a rule of customary international law, therefore, is whether the presence of BITs establishes a sense of legal obligation or at least serves as evidence of such an obligation. Clearly, the BITs’ effect on international law would be simplified if these treaties included an explicit acknowledgment that the treaty merely codified rules of customary law; unfortunately, BITs do not contain such language. Similarly, if the treaties explicitly stated that they did not represent a codification of a legal obligation, it would be clear that BITs should not be taken as evidence of customary law. Again, the treaties themselves are silent on this point. To determine whether BITs evidence a sense of legal obligation on the part of signatories, therefore, requires an inquiry into the reasons countries sign BITs. If BITs are signed out of a sense of obligation or to clarify a legal obligation, they must be considered evidence of customary international law. On the other hand, if BITs are signed for reasons unrelated, or even contrary, to a country’s sense of legal obligation, BITs are not evidence of customary international law. This chapter has provided an explanation for the popularity of BITs among developing nations that is based on the economic interests of those nations. As discussed earlier, signing a BIT offers an LDC an advantage in the competition for foreign investment. That BITs have been signed in large numbers merely demonstrates the magnitude of the perceived benefits associated with the ability to avoid the dynamic inconsistency problem. Thus, if countries have signed BITs out of economic motives, the treaties should not be interpreted as evidence of customary international law. It is equally plausible that BITs represent a permissible derogation from the existing rules of customary law and that countries have pursued the treaties because it is in their economic interest to do so. This means that BITs offer no evidence concerning the rules of customary international law that govern compensation for appropriations. The absence of a sense of legal obligation is further demonstrated by the vigorous opposition by developing countries to the Hull Rule. By the mid-1970s,

22. North Sea Continental Shelf (F.R.G. v. Den.; F.R.G. v. Neth.), 1969 I.C.J. 3, 44 (Feb. 20). 23. Schachter, supra note 9, at 126.

96 andrew t. guzman

the LDCs’ consistent objections to the Hull Rule successfully undermined its status as customary law, and there is no evidence that LDCs have since developed a greater sense of legal obligation toward the protection of foreign investment. It is simply not possible to explain the paradoxical behavior of LDCs toward foreign investment based on a view that BITs reflect opinio juris. The demise of the Hull Rule and the rise of BITs represent a struggle between developed and developing countries over the international protections to be provided for foreign investment. In the first round of this fight, developing countries successfully dismantled the Hull Rule. In the second round, developed countries responded with treaties that offered each individual LDC an opportunity to improve its position in the competition for investment. Although it appears that the developing world has lost the battle over investment protection, it must be recognized that the international legal structures were changed along the way. Developing countries have demonstrated that they do not feel an international legal obligation to provide full compensation for expropriation or to honor their contractual commitments to investors. On the other hand, they have, in pursuit of their economic self-interest, committed themselves to such behavior through BITs. BITs, therefore, do not reflect a sense of legal obligation but instead are the result of countries using the international tools at their disposal to pursue their economic interests.

conclusion Bilateral investment treaties have become the dominant international vehicle through which North-South investment is protected from host country behavior. Because these treaties allow investors and hosts to establish binding and enforceable contracts, there is little doubt that BITs increase the efficiency and reduce the cost of foreign investment. In particular, the treaties solve the dynamic inconsistency problem by permitting the host country to bind itself to a particular course of action before the investment takes place. This chapter has shown, however, that there is more to the story. Although BITs improve the efficiency of foreign investment, they may not increase the welfare of developing countries. BITs give an individual country the ability to make credible promises to potential foreign investors. As a result, the country is more attractive to foreign investors and will receive a larger volume of investment than it would without the ability to make such promises. The increase in investment, however, is likely to come in large part at the expense of other developing countries. Developing countries as a group, therefore, will enjoy relatively modest gains from an increase in total investment. It is probable that these gains will be outweighed by the losses those countries will suffer as they bid against one another to attract investment. Developing countries would be better off if, rather than competing against one another to attract investment, they could require potential investors to

explaining the popularity of bilateral investment treaties 97

commit their investments to a particular country without a binding investor-host agreement. In this situation, which exists if neither the Hull Rule nor a BIT governs the investment, the host country can extract rents from a foreign investor because it can wait until an investment is made before increasing the costs to the investor. Just as a monopolist increases the price and reduces the quantity of goods sold to maximize profits, host countries under CERDS can increase the costs to investors and maximize the gains to the host country. This strategic analysis of the behavior of developing countries explains why developing countries support CERDS–a collective action that allows LDCs to maximize their profits as a group—and, contemporaneously, sign BITs—an individual action that gives a signatory an advantage relative to other developing countries. It also makes it possible to assess the welfare implications of BITs. There is little doubt that BITs increase the overall efficiency of foreign investment, but they appear to do so at the cost of reducing the gains to developing countries. Whether this is a desirable trade off is, perhaps, a matter of debate. Finally, the analysis herein argues against viewing BITs as evidence of customary law. Developing countries sign these treaties to gain an advantage in the competition for investment rather than from a sense of legal obligation, as is required to establish a rule of customary international law.

This page intentionally left blank

4. double tax treaties: an introduction reuven s. avi-yonah

The existing network of more than 2,500 bilateral double tax treaties (DTTs) represents an important part of international law. The current DTTs are all based on two models, the Organisation for Economic Co-operation and Development (OECD) and United Nations (UN) model DTTs, which in turn are based on models developed by the League of Nations between 1927 and 1946. Despite some differences that will be discussed below, all DTTs are remarkably similar in the topics covered (even the order of articles are always the same) and in their language. About 75% of the actual words of any given DTT are identical with the words of any other DTT. Thus, the DTT network is the most important element of the international tax regime, that is, the generally applicable rules governing income taxation of cross-border transactions. Indeed, I have argued that given the similarities among all DTTs, certain rules embodied in them (such as the requirement to prevent double taxation by granting an exemption or a foreign tax credit) have become part of customary international law, and therefore may be binding even in the absence of a DTT. DTTs are generally titled “Convention Between [Country X] and [Country Y] for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion With Respect to Taxes on Income.” This title provides us with quite a bit of information. First, DTTs are bilateral: They represent a bargain between two countries, like BITs but unlike other economic law treaties, such as the General Agreement on Tariffs and Trade (GATT). Moreover, unlike the BITs, DTTs generally do not contain a Most Favored Nation (MFN) article, which means that their provisions cannot be transferred to third countries. Second, the title states that the DTT, like all DTTs, is for the “Avoidance of Double Taxation and the Prevention of Fiscal Evasion.” In truth, DTTs are generally not necessary to prevent double taxation, although they may help in borderline situations, such as cases where the source of income is disputed. This is because almost all countries prevent double taxation (i.e., taxation by both the residence and source country) unilaterally by having the residence country either grant an exemption to foreign source income or grant a foreign tax credit for source country tax on that income. Since these provisions apply unilaterally without a DTT, DTTs are generally not needed to prevent double taxation. If DTTs do not address double taxation, what do they do? DTTs shift tax revenue from source countries to residence countries, because under the generally

100 reuven s. avi-yonah

accepted rules, the source country is allowed to impose the first tax on any revenue deriving from sources within it. In the absence of a DTT, source countries can tax both active and passive income within the country. In addition, source countries are not bound by a permanent establishment or DTT sourcing rule defining what income originates within the country. DTTs shift the burden of taxation from source to residence country in two ways. The main mechanism for active income is the definition of permanent establishment. DTTs generally bar source-based taxation unless an enterprise of the other state has a permanent establishment, i.e., some kind of fixed base of operations directly or through a dependent agent, in the source country. The main mechanism for passive income is a reduction in withholding at its source. The U.S. model DTT, for example, reduces taxation on interest and royalties to zero; the only category of passive income that is eligible for source-based taxation is dividends, which are taxed at a reduced rate. The OECD model DTT, which is the main model for developed countries, reduces tax on royalties to zero but has a positive rate on interest and dividends. The UN model DTT, which is the main model for developing countries, has higher rates of source-based taxation on passive income (and a lower permanent establishment threshold for active income), but even it shifts tax revenues from the source to the residence country. This DTT structure works well if the flows of income are reciprocal, but it creates a problem for developing countries. In the reciprocal situation, residents of country A derive income from sources from country B, and residents of country B derive income from sources of income from country A. In the absence of the DTT, country A will tax the country B residents’ source income, and country B will tax the country A residents’ source income; both countries A and B will probably grant a tax credit or exemption to alleviate double taxation and encourage cross-border investment. The DTT shifts the taxation of some categories of income, particularly passive income, from the source to the residence country. Under the DTT, country B will not tax passive income that goes to country A residents, and country A will not tax passive income that goes to country B residents. As long as the capital flows are more or less reciprocal, the DTT reduces the administrative burden of imposing withholding taxes, and the net revenue is more or less the same. The amount that country A loses by not imposing its withholding tax is regained by not having to give credit for the taxes imposed by country B on income its own residents earn overseas. For example, suppose that investors from B are taxed on $100 at a rate of 30% and this falls to 0%, then A loses $30 in revenue. This however would be offset if A also saw the tax on its $100 in investment on B fall from 30% to 0%, allowing it to tax this income at its standard rate. However, if A has no investment in B, then this is just a loss of $30. If the investment flow only goes one way, and investment always flows from country B into country A, then it is much harder to get into a DTT because a DTT will always transfer revenue from country A to country B. Thus, developing

double tax treaties: an introduction 101

countries have traditionally not chosen to enter into DTTs with developed countries because the DTTs lead to a loss of tax revenue. Some developed countries, such as Germany, Sweden and Japan, have historically had extensive DTT networks with developing countries because they were willing to provide tax-sparing credits (credits for taxes that would have been collected at source but for a tax holiday), but the United States, which refuses to grant tax sparing in its DTTs, had few DTTs with developing countries until the 1990s, although the situation has changed somewhat in recent years. One reason for the recent expansion in U.S. DTTs with developing countries is that the DTT provides certainty for U.S. investors regarding the tax law of the other country, and most developing countries consider it to their benefit to encourage U.S. investment. Another reason is that DTTs generally include an exchange-of-information provision that allows the developing country to obtain information exchanged from the United States, and developing countries have increasingly been interested in trying to tax capital invested by their rich residents overseas. Since the OECD model is the source of most DTTs, we shall focus on it and compare it to the UN and U.S. models. One such difference is that the U.S. model DTT, but not the OECD or UN models, “shall not restrict in any manner any benefits now or hereafter accorded by the laws of either contracting state.” In other words, from a U.S. perspective, DTTs may never increase taxation, but they may only reduce the taxation that would otherwise apply. One reason for this is that tax laws are passed by Congress as a whole, whereas DTTs are ratified only by the Senate. A tax increase through a DTT would be unconstitutional because it would never have been ratified by the House of Representatives. A related point is that the United States only allows DTTs to reduce foreign taxation of U.S. citizens and U.S. taxation of foreigners; DTTs cannot affect the way the United States taxes U.S. residents. This provision is written into the savings clause, appearing in most U.S. tax DTTs at the end of the first article. The savings clause also states that people who have lost their citizenship for tax -motivated reasons should be treated as if there were no DTT because the United States argues that DTTs are not designed to protect U.S. citizens from U.S. taxes. Article 2 of the DTT states that the taxes covered in the DTT are only income taxes. For example, in U.S. DTTs, the Social Security tax, which is a payroll tax, is generally excluded and sometimes is covered by other agreements. Estate and gift taxes are also covered by other agreements. Thus, the DTT has the largest effect on the imposition of the income tax. Importantly, U.S. DTTs do not generally protect against any type of state tax (although in the OECD and UN models, sub-federal income taxes are covered). Most states have corporate and individual income taxes, which may impose a high burden. For example, a foreign company might wish to open an office in New York City to engage in preparatory and auxiliary activities, which are exempt from federal taxation by the permanent establishment article; however, the combined New York state and New York City

102 reuven s. avi-yonah

corporate income tax can be as high as 20%, imposing a significant tax burden on the company. Article 4 covers residence and is important because it defines who is covered by the DTT. In general, groups covered by the DTT are tax residents (i.e., people considered residents for tax purposes, which generally requires physical presence greater than 183 days); tie-breaking rules are included to prevent dual-residency situations. Corporations are deemed to be residents in the country in which they are incorporated (the U.S. position) or in the country in which they are managed and controlled (the UK position). Next we turn to a discussion of the substantive provisions. Article 5 covers permanent establishment. This provision is quite narrow in scope; in the OECD and U.S. models, the permanent establishment threshold is set high because they are for developed countries interested in reducing source-based taxation of capital-exporting enterprises. Thus, in the OECD and U.S. models, a construction facility or an oil drilling facility must be in the country for more than twelve months to be taxed, but in the UN model, it is only six months. The OECD and U.S. article also includes a long list of exceptions and a specific bar against force-of-attraction rules (such as found in the UN model) in which income is taxed when there is a permanent establishment, even though it is not attributable to the permanent establishment. Article 6 covers real property; taxation of real property at source is allowed, including, as under the U.S. rule, corporations most of whose assets are real property. Article 7 is the business profit article, which talks about taxation of business profits only if they are connected to a permanent establishment; Article 9 is the associated enterprise article, which says that if there is a transfer pricing adjustment and the other country agrees to it, then the other country shall make a corresponding adjustment to prevent double taxation – but notice that the other country must agree. Many transfer pricing adjustments, unfortunately are not agreed to by the other country and result in double taxation (source-source double taxation), so this article is only of a little help. The subsequent articles reduce source-based taxation on passive income, dividends, interest, royalties, and capital gains. These articles are the heart of the DTT, whose main function (as explained above) is to reduce source-based taxation of passive income. Under the U.S. model, the only source-based taxation that is allowed is the tax on dividends; there is no tax at all on interest, royalties, or capital gains. The OECD model, by contrast, permits tax on interest. Under the U.S. model DTT, a tax of 5% is allowed on direct dividends (dividends to corporations that own a high percentage of the shares) and 15 % for portfolio dividends, but recent U.S. DTTs reduce the tax on direct dividends to zero. The next articles address independent services and other special topics. Article 22 of the U.S. model covers limitation of benefits and is a major element in modern U.S. DTTs. The limitation-of-benefits article is designed to prevent treaty-shopping (i.e., the use of treaties between two countries by parties from a third, nontreaty country to obtain withholding tax reductions) such as the

double tax treaties: an introduction 103

SDI case, in which copyrighted software located in Bermuda was licensed to the Netherlands and from there licensed to the United States, and royalties were paid from the United States to the Netherlands and from the Netherlands to Bermuda. This was beneficial to the company because the Netherlands has a DTT with the United States that reduces taxation of the royalties to zero and Bermuda (a tax haven) does not. In that case, the IRS argued that although the royalties from the Netherlands to the United States were protected by the DTT, the royalties from the Netherlands to Bermuda were also U.S.-source royalties because of the software’s use in the United States; however, the court rejected that argument, saying that if it permitted the taxation of the NetherlandsBermuda royalties it would be allowing double taxation if the U.S.-Netherlands royalties were not protected by DTT. This seems like a strange argument because the treaty-shopping occurred only because of the DTT, so that there would be no U.S.-Netherlands royalties if the DTT did not exist. Most U.S. DTTs did not have elaborate anti-treaty-shopping mechanisms before the 1980s, and other countries were able to use those DTTs to get reduced withholding taxation. Consider the example of the Netherlands Antilles, a Caribbean tax haven that used to belong to the Netherlands. Before the enactment of the portfolio interest exemption in 1984, U.S. companies established Netherlands Antilles subsidiaries and were subject to a zero withholding tax on interest from those subsidiaries through the U.S.-Netherlands DTT. In 1984, the United States terminated the extension of the DTT to the Netherlands Antilles and enacted a portfolio interest exemption; at the same time, it instituted a limitation of benefits. These limitation-of-benefits articles in U.S. DTTs are often much more complicated than the model version because other countries want to create loopholes to allow for treaty-shopping. In addition, the OECD and UN models do not contain limitation on benefits, although the commentary on Article 1 of the OECD model has a model limitation on benefits article. Limitation-of-benefits provisions state that the DTT confers benefits only on individuals who are physically present in the other DTT country and companies that either are publicly traded on a stock exchange of the other country or are privately owned companies that do not pay half or more of their income to a resident of a nontreaty country. Thus, the U.S. model DTT takes the view that reductions in source income taxation should be accompanied by increases in residence income taxation. For example, the residence article states that if an entity is fiscally transparent in the residence country and is a partnership not subject to residence-based taxation, the entity really belongs to a group of people in another country and is therefore not considered a resident. Thus, the DTT attempts to reduce taxation at the source only if taxation increases on a residence basis, although it is unclear that the limitation-on-benefits provisions really achieve this purpose. The U.S. model DTT also includes an earning stripping provision that prevents the deduction of too much interest from the United States. Note, however, that the provision applies to interest but not to royalties; companies may therefore

104 reuven s. avi-yonah

strip earnings through royalties without penalty. The OECD and UN models do not include this provision. The United States is probably correct in insisting on limitation of benefits, although other countries certainly do not agree that they need to abide by the U.S. position. Without limitation of benefits, nontreaty countries have less incentive to negotiate because they have a so-called “treaty with the world,” meaning that they can always benefit from other countries’ DTTs by entering the DTT network through another country. Article 24 contains a nondiscrimination provision stating that countries may not discriminate against residents of the other country. This provision is weak compared with similar provisions in the GATT and the BITs, and is hard to enforce. Article 25 is the competent authority procedure for mutual agreement, which provides for some (generally nonbinding) arbitration in cases where the DTT lacks binding force. The OECD model now contains a binding arbitration provision designed to prevent double taxation, and some new U.S. DTTs include a similar provision. Note, however, that no DTT provides for binding arbitration at the request of the taxpayer, like the BITs and NAFTA. The last important component of the model DTT is the exchange-of-information provision, which is found in Article 26. The United States and most other OECD countries believe that this provision is essential if it is to enforce residence-based taxation on its own residents. The United States has been willing to forgo the ratification of a DTT rather than ignore this provision. For example, the U.S.-Israel DTT was delayed for almost twenty years because the Israelis were not willing to give sufficient written assurance of cooperation in exchange-of-information requests. The exchange-of-information provision raises important privacy questions: Is it necessary to make a specific request for specific information about a specific resident, or is it possible to request information about a group of residents without including names? In addition, bank secrecy provisions mean that often a government might not have the information requested. In addition, no worldwide system of tax identification numbers exists, so it is not necessarily true that information provided by a country could be linked to specific taxpayers. Now we turn to two topics that are important in DTT negotiation, although they are not included in the model DTTs. The first topic is tax sparing, reflected in the provisions of Article 23, which requires that foreign tax credit be given only if foreign taxes are actually paid. A number of countries provide for taxsparing credits; Germany and Japan, for example, give credit for taxes that would have been collected at source from a permanent establishment or a subsidiary except for a tax holiday. The rationale is that a tax holiday in the absence of tax sparing may simply lead to a taxpayer owing more tax to the residence country; however, this may be an overly narrow interpretation because the availability of deferral and averaging (cross-crediting between high and low-tax jurisdictions) mean that tax holidays usually benefit taxpayers even without tax sparing. Tax sparing has been an especially contentious issue in DTT ratification: The U.S.

double tax treaties: an introduction 105

Senate has been insistent that it will never ratify a DTT that provides for tax sparing because it can result in double nontaxation. The second issue involves DTT overrides, which are a mostly U.S. provision with relatively limited scope. The U.S. Congress takes the position that treaties do not have a status above domestic laws; the treaty is superior if it is implemented after a law, but the law is superior if it is implemented after the treaty. Treaty overrides are based on the supremacy clause in the U.S. Constitution that says that laws and treaties shall be the supreme law of the land. Because laws and treaties are used in the same sentence, without a specific priority given to one or the other, the clause has been interpreted to mean that laws and treaties have the same status. Unlike laws, however, treaties are negotiated with another party; the other party may feel that it is entitled to the benefits of the treaty without the risk that the treaty will be unilaterally changed by the U.S. Congress, and international law supports this argument. The U.S. position is based on the argument that DTTs are only ratified by the Senate whereas tax laws have to be passed by both houses of Congress, and therefore Congress must to be able to supersede DTTs. Although DTT overrides are a contentious issue, the actual number of DTT overrides are relatively small. One explicit DTT override was the branch profit tax provision in 1986, which added a limitation-of-benefits rule to preexisting U.S. DTTs that did not have such a provision. This override is generally obsolete now because almost all of the DTTs that it affected have since been renegotiated. The most recent example of a DTT override occurred in the context of tax arbitrage. This was a situation in which a Canadian company had an U.S. subsidiary with U.S.-source income, which it repatriated to Canada. For U.S. tax purposes, the subsidiary was treated as a branch, and the payments to Canada were treated as interest that was deductible and subject to a reduced withholding tax under the U.S-Canada DTT. For Canadian tax purposes, the U.S. subsidiary was treated as a corporation, and the payments were treated as dividends, which are exempt in Canada. The result was double nontaxation. The United States took the position that a reduction in tax on source income should be contingent on an increase in tax on residence income. The United States, therefore, passed a DTT override stating that taxation would not be reduced for hybrid entities that are treated inconsistently. Almost immediately, Canada agreed with the interpretation of the DTT and negotiated a protocol to change the DTT. Although this may imply that the DTT override was unnecessary because the Canadians were willing to renegotiate, it takes a lot of time to negotiate a protocol, so it was not necessarily inappropriate for the United States to use a DTT override. Because of the contention surrounding DTT overrides, the United States hesitates to make use of the provision and has been very careful not to override DTTs since 1986. In general, the United States attempts to avoid overrides and the appearance of overrides. When the U.S. earning stripping rule (restricting interest

106 reuven s. avi-yonah

deductions to tax-exempt related parties) was enacted in 1989, the United States went to great lengths to avoid the appearance of a DTT override by extending the provision to domestic tax-exempt entities and not just to foreigners, thus preventing an apparent violation of the nondiscrimination article. In general, DTTs between developing countries and developed countries benefit the developing countries despite the absence of tax sparing because DTTs ensure developed country investors a certain level of institutional stability in the developing country. However, empirical economic studies have failed to show that the existence of a DTT materially affects foreign direct investment. This is unsurprising for DTTs between developed countries, which mostly affect the distribution of revenue between the governments of the two countries, but similar studies of DTTs between developing and developed countries also fail to show that the existence of a DTT has a significant positive effect on the flows of foreign direct investment into the developing country. Nevertheless, most scholars believe that DTTs increase investor confidence in the stability of investing in developing countries, and therefore, although the developing country might forgo some tax revenue from the DTT, it probably benefits in the long run from the increased foreign direct investment. One final important issue regarding DTTs is why they are bilateral, rather than multilateral, such as the GATT. The usual explanation is that DTTs depend too much on the specific investment flows between countries, and therefore cannot be multilateral; however, the fact that developing countries are willing to enter into DTTs even if they lose revenue suggests that this is not a complete explanation. In my view, the fact that DTTs are bilateral is mostly due to the fact that the models were developed before World War II, when bilateral treaties were the norm, and when differences between the tax laws of different countries were larger than they are today. If that is true, it suggests that the time may be at hand to try to negotiate a multilateral DTT, especially given the difficulty DTTs face when dealing with “triangular cases” involving third countries. Tax laws have converged a lot since the 1920s, and multilateral treaties are now the norm, so that a renewed effort to negotiate such a multilateral DTT (perhaps in the World Trade Organization context) seems to be called for.

part two exploring the impact of bilateral investment treaties on foreign direct investment flows

This page intentionally left blank

5. do bit s really work?: an evaluation of bilateral investment treaties and their grand bargain∗ jeswald w. salacuse and nicholas p. sullivan introduction International investment law has undergone a remarkable transformation in a relatively short time. The fundamental tool for effecting that transformation has been the bilateral investment treaty (BIT), an international legal instrument through which two countries set down rules that will govern investments by their respective nationals in the other’s territory.1 From 1959 to 2002, nearly 2,200 individual BITs were formed,2 making the BIT one of the most widely used types of international agreement for protecting and influencing foreign investment.3 As the twenty-first century begins, the time has come to evaluate whether BITs have achieved their objectives.4 To answer this question, Part A examines ∗ This chapter was reprinted with permission from the Harvard International Law Journal. The chapter was originally published as “Do BITs really work: an evaluation of bilateral investment treaties and their grand bargain,” 46 Harv. Int’l L.J. 67 (2005). 1. The literature and doctrinal commentary on Bilateral Investment Treaties (BITs) are abundant and have expanded over the years as the number of BITs has grown. See generally, e.g., Rudolf Dolzer & Margrete Stevens, Bilateral Investment Treaties (1995); M. Sornarajah, The International Law on Foreign Investment 225–76 (1994); UN Conf. on Trade and Dev. (UNCTAD), Bilateral Investment Treaties, 1959–1999, UN Doc UNCTAD/ ITE/IIA/2 (2000), at http://www.unctad.org/en/docs//poiteiiad2.en.pdf (last visited Nov. 29, 2004); UNCTAD, Bilateral Investment Treaties in the Mid-1990s, UN Doc. UNCTAD/ITE/IIT/7 (1998); K. J. Vandevelde, United States Investment Treaties: Policy and Practice (1992); Antonio R. Parra, The Scope of New Investment Laws and International Instruments in Economic Development, in Economic Development, Foreign Investment and Law 27 (R. Pritchard ed., 1996). In addition, see the Web site of the International Centre for Settlement of Investment Disputes (ICSID) for materials on BITs, including the texts of many BITs, arbitration awards that have interpreted and applied them, and a bibliography of books and articles commenting on BITs. See ICSID, ICSID Bilateral Investment Treaties, at http://www.worldbank.org/icsid/treaties/treaties.htm (last visited Nov. 30, 2004). 2. See UNCTAD, World Investment Report 2003: FDI Policies For Development: National and International Perspectives 89, UN Doc. UNCTAD/WIR/2003 (Sept. 4, 2003) (stating that 2,181 BITs were in effect as of the end of 2002). 3. See id. 4. For earlier speculation on this question, see Jeswald W. Salacuse, BIT by BIT: The Growth of Bilateral Investment Treaties and Their Impact on Foreign Investment in Developing Countries, 24 Int’l Law. 655, 656–61 (1990).

110 jeswald w. salacuse and nicholas p. sullivan

the historical movement to form BITs. Part B explores the goals motivating BITs, namely foreign investment protection, market liberalization, and foreign investment promotion. The three succeeding parts assess the success of BITs in achieving each of these goals. Finally, we conclude by considering the implications of the BIT movement for the further development of international investment law. A. History of the BIT movement 1. Impetus for the BIT movement As recently as 1970, the International Court of Justice in the Barcelona Traction case found it “surprising” that the evolution of international investment law had not gone further and that no generally accepted rules had yet crystallized in light of the growth of foreign investments and the expansion of international activities by corporations in the previous half-century.5 International law, as stated in Article 38(1) of the Statute of the International Court of Justice, consists of three primary sources: treaties, customs that the international community considers binding, and general principles of law common to the world’s legal systems.6 In the period after World War II, as foreign investment gained momentum as an increasingly important international economic activity, foreign investors who sought the protection of international investment law encountered an ephemeral structure consisting largely of scattered treaty provisions, a few questionable customs, and contested general principles of law. This legal structure was seriously deficient in several respects. First, applicable international law failed to take account of contemporary investment practices and address important issues of concern to foreign investors.7 For example, customary international law had virtually nothing to say about the right of foreign investors to make monetary transfers from the

5. See Barcelona Traction, Light and Power Co., Ltd. (Belg. v. Spain), 1970 I.C.J. 3, 46–47 (Feb. 5). 6. Article 38 (1) states that: 1. The Court, whose function is to decide in accordance with international law such disputes as are submitted to it, shall apply: a. international conventions, whether general or particular, establishing rules expressly recognized by the contesting states; b. international custom, as evidence of a general practice accepted as law; c. the general principles of law recognized by civilized nations; d. subject to the provisions of Article 59, judicial decisions and the teachings of the most highly qualified publicists of the various nations, as subsidiary means for the determination of rules of law. Statute of the International Court of Justice, art. 38(1), June 26, 1945, 59 Stat. 1055, 33 U.N.T.S. 993, reprinted in 55 Yale L.J. 1318, 1326 (1946). 7. Indeed, as late as 1994, a leading commentator on international investment law stated: “There are few customs in this sense in the field of foreign investment.” Sornarajah, supra note 1, at 74.

do bit s really work?: an evaluation of bilateral investment treaties 111

host country. Second, the principles that did exist were often vague and subject to varying interpretations. Thus, although there was strong evidence that customary international law required the payment of compensation upon nationalization of an investor’s property, no principles had crystallized on how that compensation was to be calculated.8 Third, this existing framework prompted disagreement between industrialized countries and newly decolonized developing nations. For example, capitalexporting states claimed that international law imposed an obligation on host countries to accord foreign investors a minimum standard of protection and required states expropriating property of foreign investors to provide compensation.9 Many developing countries, believing that the existing international rules served only to maintain their poverty, rejected this view and, beginning in the 1970s, demanded that their particular needs and circumstances be taken into account.10 Their position on foreign investment was incorporated11 into Article 2 of the 1974 United Nations Charter of Economic Rights and Duties of States, adopted by the United Nations General Assembly.12 Finally, existing international law offered foreign investors no effective enforcement mechanism to pursue their claims against host countries that had injured or seized their investments or refused to respect their contractual obligations. As a result, investors had no assurance that investment contracts and arrangements made with host country governments would not be subject to

8. See id. 9. See Restatement (Third) of the Foreign Relations Law of the United States § 712 (1987). 10. Inspired by the success of the oil-producing countries in raising petroleum prices in 1973–74, developing countries had hoped that by building a numerically strong coalition amongst themselves, they would be able to bring about desired change in various international fora. As a result of the debt crisis in the early 1980s, the internal economic restructuring demanded by international financial institutions, such as the International Monetary Fund (IMF) and the World Bank, and the abandonment of command economy models by developing countries, the movement for a “New International Economic Order” lost steam and was virtually dead by 1990. See Thomas Waelde, Requiem for the “New International Economic Order,” in Festschrift Fuer Ignaz Seidl-Hohenveldern 771 (Gerhard Hafner et al. eds., 1998). See generally Jeffrey Hart, The New International Economic Order (1983); The New International Economic Order: The North-South Debate (Jagdish N. Bhagwati ed., 1977). 11. See Robert F. Meagher, An International Redistribution of Wealth and Power: A Study of the Charter of Economic Rights and Duties of States (1979). 12. UN Charter of Economic Rights and Duties of States, G.A. Res. 3281, UN GAOR, 29th Sess., Supp. No.31, at 50, UN Doc. A/9946 (Jan. 15, 1974); UN Doc. A/9631 (1974), reprinted in 14 I.L.M. 251, 255 (1975). Article 2 provides that each state has the right to expropriate foreign property, subject to the duty to pay “appropriate” compensation, is not required to give foreign companies preferential treatment, and has the right to revise and renegotiate contracts it has made with foreign companies.

112 jeswald w. salacuse and nicholas p. sullivan

unilateral change by those governments at some later time. Although the International Centre for Settlement of Investment Disputes (ICSID) had been formally established in 1965 as an affiliate of the World Bank to resolve disputes between host countries and foreign private investors,13 the Centre did not hear its first case until 1972.14 Injured foreign investors who were unable to negotiate a satisfactory settlement, secure an arbitration agreement with a host government, or find satisfaction in the local courts had few options other than to seek espousal of their claims by their source country governments, a process that by its very nature was more political than legal. A short three decades later, the legal architecture for the protection of foreign investment has changed dramatically. From the point of view of the foreign investor, this structure has become a far more protective shelter than it was in the 1970s. In most cases, a foreign investor benefiting from a BIT may now look to a comprehensive, specific, and largely uncontested set of international legal rules, with recourse to international tribunals for enforcement.15 Equally important to the change in the content of international investment law has been the change in its sources. Today, unlike the situation that prevailed in the early 1970s, foreign investors are protected primarily by international treaties, rather than by customary international law alone.16 For all practical purposes, BIT law has become the fundamental source of international law in the area of foreign investment.17 In addition to largely displacing customary law as a source of controlling legal principles in specific investment cases, and perhaps even beginning to influence the formation of a new customary international law of investment, BITs have also displaced, and in some cases replaced, the relevant domestic law of the host country in many important respects. It is this latter aspect of the development of international investment law that has generated, and is continuing to generate, significant controversy within both developed and developing countries, as it

13. Convention on the Settlement of Investment Disputes Between States and Nationals of Other States, Mar. 18, 1965, 17 U.S.T. 1270, 575 U.N.T.S. 159. 14. See ICSID, List of Concluded Cases, at http://www.worldbank.org/icsid/cases/ conclude.htm (last visited Nov. 30, 2004) (listing concluded cases in chronological order). 15. As of October 2004, for example, ICSID’s docket consisted of eighty-two pending cases brought by foreign investors against host countries. Since its creation, ICSID has handled 170 foreign investment cases, including 83 pending and 87 completed cases. See ICSID, ICSID Cases, at http://www.worldbank.org/icsid/cases/cases.htm (last visited Nov. 30, 2004). In 2003 alone, a record number of 26 cases were registered with ICSID. For a listing of these disputes before the Centre, see ICSID, ICSID 2003 Annual Report 9–31 (2003). See also ICSID, at http://www.worldbank.org/icsid (last visited Nov. 30, 2004) (providing information and awards relating to many of these cases). 16. See UNCTAD, supra note 2, at 89. 17. See Patrick Juillard, L’Evolution des Sources du Droit des Investissements, 250 Recueil des Cours de L’Académie de Droit International 74 (1994).

do bit s really work?: an evaluation of bilateral investment treaties 113

places host country concerns about national sovereignty and the right to control the activities of foreign investors in opposition to investor concerns about protection from unjustified interference in their investments.18 BITs may also have an impact in creating or shaping law outside of their specific texts. In ratifying a BIT, a country makes the treaty part of its legal system. In order to meet the demands of both the letter and intent of a BIT, a country may modify internal legislation affecting investment. With the proliferation of BITs and the great expansion of the number of countries that have signed them, one may consider whether the BITs themselves are evidence of a new international consensus on investment law—an evolving set of customs that the international community is coming to consider obligatory. Even if they cannot yet be seen as custom, perhaps at least they provide evidence of another source of international law—general principles of law common to the world’s legal systems. We shall attempt to address both possibilities in the conclusion of this chapter. 2. Evolution of the BIT movement The movement to create an international law of investment began in the 1950s with the rapid expansion of international investment in the post–World War II era. Over the past three decades in particular, BITs have proliferated as foreign direct investment (FDI) has experienced phenomenal growth. Total annual FDI reached $1.1 trillion in 2000, a drastic increase from $25 billion in 1973.19 For much of this period, FDI grew faster than international trade. During the period between 1973–1995, the estimated value of FDI outflows increased twelvefold while the value of merchandise exports increased less than ninefold.20 This international flow of capital has both driven and been driven by the development of international investment law. Investors seeing profitable economic opportunities for their capital and technology abroad have pushed their governments to enter into arrangements with other countries to facilitate and protect their investments.21 At the same time, the development of new

18. See, e.g., Aaron Cosbey et. al., Investment and Sustainable Development: A Guide to the Use and Potential of International Investment Agreements 12–15 (2004) (discussing expropriation and the balancing of host, public, and foreign investor interests), available at http://www.iisd.org/pdf/2004/investment_invest_and_sd.pdf (last visited Nov. 21, 2004); Konrad von Moltke & Howard Mann, Towards a Southern Agenda on International Investment: Discussion Paper on the Role of International Investment Agreements 29–30 (2004) (discussing “regulatory chill” that BITs can induce), available at http:// www.iisd.org/pdf/2004/investment_sai.pdf (last visited Nov. 19, 2004). 19. See UNCTAD, supra note 2, at 9. 20. See World Trade Org. (WTO), 1 Annual Report 1996, 46 (1996). 21. See, e.g., Statement, International Chamber of Commerce (ICC), ICC Statement on Behalf of World Business to the Heads of State and Government attending the Evian Summit (June 1–3, 2003), available at http://www.iccwbo.org/home/statements_rules/ statements/2003/G8.asp (last visited Nov. 19, 2004).

114 jeswald w. salacuse and nicholas p. sullivan

international investment rules may also have had the effect of encouraging new capital flows to the countries concerned, a proposition that this chapter specifically seeks to address. Due to the inadequacy of customary international law, capital-exporting nations since World War II have made efforts to create international rules to facilitate and protect the investments of their nationals and companies abroad. These efforts have taken place at both the bilateral and multilateral levels, which, though separate, have tended to inform and reinforce each other.22 Early attempts to create an international investment law for the post–World War II era sought to create multilateral treaties. The first such effort was the Havana Charter of 1948, which would have created the International Trade Organization with powers to promulgate rules on international investment.23 Due partly to opposition from the business community, the Havana Charter failed to gain support from a sufficient number of states.24 Subsequent efforts included the International Chamber of Commerce’s International Code of Fair Treatment of Foreign Investment (1949), the International Convention for the Mutual Protection of Private Property Rights in Foreign Countries (1957), a private effort known as the Abs-Shawcross Convention, and the Organization for Economic Cooperation and Development (OECD) Draft Convention on the Protection of Foreign Property (1967).25 Although none of these proposals was ever adopted, they did inform and influence the development of the BIT movement that was to come.26 The most immediate and practical impact of such bilateral efforts was the creation of enforceable rules to govern international investment. Bilateral commercial treaties had existed for centuries, but the primary purpose of these

22. See, e.g., Thomas W. Walde, Introductory Note, European Energy Conference: Final Act, Energy Charter Treaty, Decisions, and Energy Charter Protocol on Energy Efficiency and Related Environmental Aspects, 34 I.L.M. 360, 360 (1995) (noting the strong influence of BITs on the trade provisions of a multilateral energy treaty); Patrick Juillard, Le Réseau Français des Conventions Bilatérales d’Investissement: á la Recherche d’un Droit Perdu?, 13 Droit et Pratique du Commerce Internationale 9, 16 (1987) (noting that France based its model BIT on the 1967 Organization for Economic Cooperation and Development (OECD) Draft Convention on the Protection of Foreign Property). 23. Havana Charter for an International Trade Organization, March 24, 1948, UN Doc. E/ Conf. 2/78. 24. See William Diebold, Jr., The End of the ITO 9 (Princeton Essays in International Finance No. 16, 1952), cited in Todd S. Shenkin, Trade-Related Investment Measures in Bilateral Investment Treaties and the GATT: Moving toward a Multilateral Investment Treaty, 55 U. Pitt. L. Rev. 541, 555 n.68 (1994). 25. See generally Franziska Tschofen, Multilateral Approaches to the Treatment of Foreign Investment, 7 ICSID Rev.-Foreign Inv. L.J. 384, 385–86 (1992) (surveying various efforts to prepare multilateral treaties on foreign investment). 26. See Juillard, supra note 22, at 16.

do bit s really work?: an evaluation of bilateral investment treaties 115

earlier agreements was to facilitate trade, rather than investment.27 In its early history, the United States made large numbers of such agreements—known as treaties of friendship, commerce, and navigation28—and their geographic spread reflected the expansion of U.S. foreign trade.29 Though these treaties were intended to facilitate trade and shipping, they occasionally contained provisions affecting the ability of one country’s nationals to own property or do business in the territory of the other contracting state.30 Responding to the increase of U.S. foreign investment after World War II, the United States undertook a program to conclude a network of bilateral treaties of friendship, commerce, and navigation that, in addition to other commercial matters, specifically sought to facilitate and protect U.S. direct investment abroad.31 Although the United States signed twenty-three such treaties between 1946 and 1966,32 this effort soon lost momentum as developing countries, increasingly skeptical of the benefits of foreign investment, grew reluctant to agree to the guarantees that the United States requested to protect American investments abroad.33 A new and important phase in the historical development of modern international investment law began in the late 1950s, as individual European countries negotiated bilateral treaties that, unlike the previous commercial agreements, dealt exclusively with foreign investment. These countries sought to create a basic international legal framework to govern investments by nationals of one country in the territory of another. The modern BIT was thus born. Germany, which had lost all of its foreign investments as a result of its defeat in World War II, took the lead in this new phase of bilateral treaty-making. Starting with an

27. See, e.g., Shenkin, supra note 24, at 570 (“The early [U.S. treaties of friendship, commerce, and navigation] were concerned primarily with the trade and shipping rights of individuals.”). 28. See Robert R. Wilson, United States Commercial Treaties and International Law (1960). For a history of U.S. treaties of friendship, commerce, and navigation, see Herman Walker Jr., Modern Treaties of Friendship, Commerce and Navigation, 42 Minn. L. Rev. 805 (1958). 29. See Kenneth J. Vandevelde, The Bilateral Investment Treaty Program of the United States, 21 Cornell Int’l L.J. 201, 204 n.29 (1988) (recounting how the United States made bilateral commercial treaties first with Western Europe, then with Latin America, later with Asia, and still later with Africa). 30. See Shenkin, supra note 24, at 570. 31. See Sumitomo Shoji America v. Avagliano, 457 U.S. 176, 185–90 (1982) (discussing history of U.S. BITs). See also Salacuse, supra note 4, at 656–61. 32. See Shenkin, supra note 24, at 573. 33. See K. Scott Gudgeon, United States Bilateral Investment Treaties: Comments on their Origin, Purposes, and General Treatment Standards, 4 Int’l Tax & Bus. L. 105, 111 (1986), cited in Shenkin, supra note 24, at 573.

116 jeswald w. salacuse and nicholas p. sullivan

agreement with Pakistan in 1959,34 Germany proceeded to negotiate similar investment treaties with countries throughout the developing world.35 Switzerland, France, the United Kingdom, the Netherlands, and Belgium followed in relatively short order.36 By 1977, European countries had concluded approximately 130 BITs with a broad array of developing countries.37 Several factors may have contributed to the relative success of the European programs. Compared to the United States, European countries generally were less demanding with respect to guarantees on such matters as free conversion of local currency, abolition of performance requirements, and protection against expropriation. Moreover, specific foreign aid relationships between certain European countries and the European Community, on the one hand, and individual developing countries, on the other, may have predisposed some of the latter to look favorably on concluding BITs with European states.38 Spurred in part by the experience of the Europeans, the United States launched its own BIT program in 1981.39 By September 2004, it had signed forty-five BITs with developing countries and emerging markets.40 As non-Western countries began 34. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 177–79 (reporting historical data on the BITs entered into by Germany). 35. Germany remains one of the leaders in BIT formation, having concluded 119 treaties as of June 2004. See UNCTAD, Investment Instruments Online, at http://www.unctadxi. org/templates/DocSearch.aspx?id=779 (last updated June 15, 2004) (last visited Nov. 19, 2004). 36. See generally id. (providing current statistics on BITs). 37. See International Chamber of Commerce, Bilateral Investment Treaties for International Investment 13–16 (1977). 38. Foreign aid to developing countries has been a function of numerous factors, including strategic, commercial, political, and humanitarian considerations. Peter Hjertholm and Howard White, Foreign Aid in Historical Perspective: Background and Trends, in Foreign Aid and Development: Lessons Learnt and Directions for the Future 99–100 (Finn Tarp ed., 2000). As a result, it is difficult to know precisely the impact of a donor country’s aid policies on its BIT negotiations with a specific aid recipient. From the point of view of a recipient country, one indicator of the quality of aid is the percentage that is “untied,” i.e. not required to be spent on acquiring goods and services from the donor country. It is interesting to note that European countries whose aid was the least “tied” were among the countries that had concluded the largest number of BITs in 1981. In that year, when the percentage of untied aid given by the United States was only 33%, Germany, with untied aid of 74%, had signed 49 BITs; Switzerland with untied aid of 50% had signed 33 BITs; the Netherlands with untied aid of 57% had signed 16 BITs; and Sweden with untied aid of 84% had signed 8 BITs. Hjertholm and White, supra at 96. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 159–217. 39. For additional background on U.S. BITs, see Vandevelde, supra note 1; Pamela B. Gann, The U.S. Bilateral Investment Treaty Program, 21 Stan. J. Int’l L. 373 (1985); Gudgeon, supra note 33, at 107–11 Salacuse, supra note 4. 40. See Fact Sheet, Bureau of Economic and Business Affairs, U.S. Dep’t of State, U.S. Bilateral Investment Treaty Program (Sept. 15, 2004) (listing all U.S. BITs as of Sept. 15, 2004), available at http://www.state.gov/e/eb/rls/fs/22422.htm (last visited Nov. 19, 2004).

do bit s really work?: an evaluation of bilateral investment treaties 117

to export capital, they too negotiated BITs to create a legal framework for their nationals’ investments in specific countries. Thus, by 1997, Japan had signed four BITs, and Kuwait had signed twenty-two.41 While BITs are usually between capital-exporting states and developing countries, on occasion, two developing countries or two industrialized countries have formed such agreements. Examples of the former include BITs between Thailand and China and between Egypt and Morocco.42 The most notable example of the latter is the 1988 agreement between the United States and Canada that created a free trade area between the two countries.43 This agreement included a special chapter that in effect functions as a BIT, considering how closely its provisions parallels those of BITs that the United States has negotiated with other countries.44 By 1994, the agreement evolved into the North American Free Trade Agreement (NAFTA) between Canada, Mexico, and the United States.45 For all intents and purposes, NAFTA’s section on investment, Chapter Eleven, constitutes a BIT between the three countries. The late 1980s witnessed a new phase in the history of the BIT movement with the end of the Communist era and the abandonment of command economies in many parts of the world. The emerging economies of Eastern and Central Europe, as well as of certain Latin American, African, and Asian countries that had previously been hostile to foreign investment, now actively sought foreign capital to finance their development. This dramatic transformation entailed sweeping changes in law and policy.46 Reflecting this policy shift, countries with emerging markets entered into BITs with industrialized states from which they hoped to receive capital and technology to advance their development, and they did so at an accelerating pace. Whereas some 309 BITs had been concluded by the end of 1988,47 2,181 BITs were concluded by 2002.48 This dramatic change in so short a period of time represents a substantial feat of international lawmaking. In 2001 alone, a total of ninety-seven countries concluded some 158 BITs,

41. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 185–86. 42. Agreement for the Promotion and Protection of Investments, P.R.C.-Thail., Mar. 12, 1985, available at http://www.unctad.org/sections/dite/iia/docs/bits/china_thailand.pdf (last visited Nov. 19, 2004); Agreement Regarding the Encouragement and Protectionof Investment, Egypt-Morocco, June 6, 1976, available at http://www.unctad. org/sections/dite/iia/docs/bits/egypt_morocco_arb.pdf (last visited Nov. 19, 2004). 43. See Free Trade Agreement, U.S.-Can., Jan. 2, 1988, 27 I.L.M. 281 (1988). 44. See id. at 373–80. 45. North American Free Trade Agreement, U.S.-Can.-Mex., Dec. 17, 1992, 32 I.L.M. 289 (1993). 46. See Jeswald W. Salacuse, From Developing Countries to Emerging Markets: A Changing Role for Law in the Third World, 33 Int’l Law. 875, 875–77 (1999). 47. See Athena J. Pappas, References on Bilateral Investment Treaties, 4 ICSID Rev.-Foreign. Inv. L.J. 189, 194–203 (1989). 48. See UNCTAD, supra note 2, at 89.

118 jeswald w. salacuse and nicholas p. sullivan

a numerical record for any single year since the BIT movement began in 1959.49 The result of this effort has been the creation of an increasingly dense BIT network linking more than 170 different countries.50 Meanwhile, the number of BITs involving two developing countries has been increasing steadily. China, for example, had concluded ninety-nine BITs with both developed and developing countries by June 2004.51 Overall, however, developed countries have signed relatively few BITs with one another.52 B. The Goals of the BIT Movement The impetus behind the rapid expansion of BITs rests in the desire of companies of industrialized states to invest safely and securely in developing countries, as well as the consequent need to create a stable international legal framework to facilitate and protect those investments. Without a BIT, international investors are forced to rely on host country law alone for protection, which entails a variety of risks to their investments. Host governments can easily change their own domestic law after a foreign investment is made, and host country officials may not always act fairly or impartially toward foreign investors and their enterprises. Investor recourse to local courts for protection may prove to be of little value in the face of prejudice against foreigners or governmental interference in the judicial process.53 Indeed, these fears were realized in the 1960s and 1970s when numerous instances of interference and expropriation of foreign investments by host country governments occurred. The number of expropriations of foreign-owned property grew steadily each year from 1960 and reached its peak in the mid-1970s.54 The lack of consensus on the customary international law applicable to foreign investments also created uncertainty in the minds of investors as to the degree of protection they could expect under international law. To gain greater certainty and to counter the threat of adverse national law and regulation, the host countries of these investors sought to conclude a series of BITs that would

49. See UNCTAD, World Investment Report 2002: Transnational Corporations & Export Competitiveness 8, UN Doc. UNCTAD/WIR/2002 (2002). 50. See, e.g., Press Release, UNCTAD, Bilateral Investment Treaties Signed in Bangkok, UN Doc. UNCTAD/INF/PR/025X (Feb. 18, 2000) (indicating that 174 countries had entered into BITs by the end of 1998), available at http://www.unctad.org/Templates/ Webflyer.asp?docID=308&intItemID=2527&lang=1. 51. See UNCTAD, supra note 35. 52. See UNCTAD, Lessons from the MAI at 22 n.9, UN Doc. UNCTAD/ITE/IIT/ Misc.22 (1999) (reporting that, as of 1999, fewer than 10 percent of BITs were between OECD countries). 53. See UNCTAD, supra note 2, at 114–18. 54. The United Nations identified 875 distinct acts of governmental taking of foreign property in sixty-two countries during the period between 1960–1974. Don L. Piper, New Directions in the Protection of American-Owned Property Abroad, 4 Int’l Trade L.J. 315, 330 (1979).

do bit s really work?: an evaluation of bilateral investment treaties 119

provide clear rules and effective enforcement mechanisms, at least with regard to their treaty partners. Their primary goal, therefore, was protection of investments made by their nationals and companies in foreign countries.55 In addition to protecting the investments of their nationals, some countries, especially the United States, have had another objective in negotiating BITs: to facilitate the entry and operation of these investments by inducing host countries to remove various impediments in their regulatory systems. They have sought to encourage or induce investment and market liberalization within their negotiating partners.56 Moreover, in the view of certain developed countries, BITs will have the effect of liberalizing the developing country’s economy as a whole, by facilitating the entry of a treaty partner’s investment and creating conditions favoring their operations. In the process of reforming their economies to foster private enterprise, some developing countries have realized that creating favorable conditions for foreign investment can be integral to their success.57 Although the BITs themselves do not specifically enunciate the goal of investment and market liberalization, that goal has clearly been in the minds of developed country negotiators and is sometimes reflected in background documents.58

55. Virtually all BITs are titled: “A Treaty Concerning . . . the Protection of Investment.” See, e.g., Treaty Concerning the Reciprocal Encouragement and Protection of Investment, U.S.-Arm., Sept. 23, 1992, S. Treaty Doc. No. 103-11 (1993); Treaty Concerning the Promotion and Reciprocal Protection of Investments, F.R.G.-Pol., Nov. 10, 1989, 29 I.L.M. 333 (1990). 56. The Deputy United States Trade Representative stated the U.S. goals in negotiating BITs as follows: The BIT program’s basic aims are to: (1) protect U.S. investment abroad in those countries where U.S. investors’ rights are not protected through existing agreements; (2) encourage adoption in foreign countries of market-oriented domestic policies that treat private investment fairly; and (3) support the development of international law standards consistent with these objectives. Jeffrey Lang, Keynote Address, 31 Cornell Int’l L.J. 455, 457 (1998). See also United States Trade Representative, USTR Focus on Investment, at http://ustr.gov/Trade_Sectors/ Investment/Section_Index.html. (last visited Nov. 6, 2004). 57. See generally Salacuse, supra note 46, at 875–77. 58. See, e.g., Investment Treaty With Albania, U.S.-Alb., Jan. 11, 1995, S. Treaty Doc. No. 104-19 (1995). (In the Message from the President of the United States Transmitting The Treaty Between the Government of the United States of America and the Government of the Republic of Albania Concerning the Encouragement and Reciprocal Protection of Investment With Annex and Protocol Signed at Washington on January 11, 1995, President Clinton stated: “The bilateral investment treaty (BIT) with Albania will protect U.S. investment and assist the Republic of Albania in its efforts to develop its economy by creating conditions more favorable for U.S. private investment and thus strengthen the development of its private sector.”). See also United States Trade Representative Web site, supra note 56.

120 jeswald w. salacuse and nicholas p. sullivan

Concluding and maintaining a treaty requires a bargain from which both parties believe they will derive benefits. An investment treaty between two developed countries, both of whose nationals expect to invest in the territory of the other, would be based on the notion of reciprocity and mutual protection; however, this bargain would not seem applicable in the context of a treaty between a developed, capital-exporting state and a poor, developing country whose nationals are unlikely to invest abroad. One might therefore ask, Why would developing countries enter into such agreements? Why would they constrain their sovereignty by entering into treaties that specifically limit their ability to take necessary legislative and administrative actions to advance and protect their national interests?59 The answer to these questions is that developing countries sign BITs to promote foreign investment, thereby increasing the amount of capital and associated technology that flows to their territories. The basic assumption behind BITs is that a bilateral treaty with clear and enforceable rules to protect and facilitate foreign investment reduces risks that the investor would otherwise face, and that such reductions in risks, all things being equal, encourage investment. In the 1980s and 1990s, as other forms of financial assistance became less available from commercial banks and official aid institutions, developing countries increasingly felt the need to promote foreign investment in order to foster economic development; they saw BITs as one means of pursuing a campaign of investment promotion and therefore signed them in increasing numbers.60 Thus, a BIT between a developed and a developing country is founded on a grand bargain: a promise of protection of capital in return for the prospect of more capital in the future. An interesting question is why the nations of the world have been willing to conclude BITs in growing numbers over the last fifty years but have steadfastly refused to join multilateral agreements on investment.61 One technical

59. This question assumes that the developing country is not expecting other benefits from its developed country treaty partner, such as increased foreign aid or enhanced security guarantees, which are extraneous to a BIT relationship. 60. See UNCTAD, supra note 2, at 85. 61. One partial exception is the European Energy Charter Treaty, art. 16, opened for signature Feb. 1, 1995, 34 I.L.M. 360 (1995)(ratified by forty-five states as of 2002). The basic aim of the treaty is to create a legal framework that will encourage the development of a secure international energy supply through liberalized trade and investment among the member states. It includes an investment chapter that has undeniably been influenced by the BIT movement. See Jeswald W. Salacuse, The Energy Charter Treaty and Bilateral Investment Treaty Regimes, in The Energy Charter Treaty: An East-West Gateway for Investment & Trade 321–48 (Thomas Walde ed., 1996). One of the unique features of this treaty, which distinguishes it from other international investment agreements, is that it is a sectoral agreement, that is, it only applies to investments in a particular economic sector. Its scope is limited to investments associated with economic activities concerning

do bit s really work?: an evaluation of bilateral investment treaties 121

explanation is that a bilateral treaty must accommodate the interests of only two parties and is therefore far less complicated to negotiate than a multilateral global treaty, which must accommodate the interests of many countries.62 Politically, given the asymmetric nature of BIT negotiations between a strong developed country and a usually much weaker developing country, the bilateral setting allows the developed country to use its power more effectively than does a multilateral setting, where that power may be much diluted. For example, in multilateral settings, developing countries have the opportunity to form blocking coalitions with like-minded countries to enhance their power in the negotiations, something that is impossible in bilateral negotiations. On the other hand, the prospects of investment capital from specific developed countries, along with other political and economic benefits arising from a definite bilateral relationship, may make a developing country more willing to enter into a BIT with a specific developed country than it would a multilateral agreement, where those benefits may seem more tenuous and theoretical. Moreover, whereas developed countries would be willing to enter into bilateral treaties with developing countries for investment liberalization, knowing full well that little if any enterprises from the developing country would ever invest in the developed country, they have been unwilling to enter into treaties that would grant such liberalization to investors from other developed countries, which could become strong competitors to the host countries’ own enterprises.63 It should be noted that investment promotion, a fundamental objective of developing countries, and investment and market liberalization, a subsidiary aim of developed countries, are separate and distinct goals.64 Within the context the exploration, extraction, refining, production, storage, land transport, transmission, distribution, trade, marketing, or sale of energy materials and products. Although the United States participated in the negotiations, it chose not to sign the treaty, apparently because it believed that its provisions on investment did not meet the strong international standards that the United States had obtained in its own BITs. William Fox, The United States and the Energy Charter Treaty: Misgivings and Misperceptions, in The Energy Charter Treaty: An East-West Gateway for Investment & Trade 194, 196 (Thomas Walde ed., 1996). For information on the current status of the treaty, see Energy Charter Treaty, at http://www.encharter.org (last visited Nov. 19, 2004). 62. For a discussion of the differences between bilateral and multilateral negotiations, see Fen Osler Hampson, Multilateral Negotiations: Lessons from Arms Control, Trade and the Environment 1–51, 345–60 (1995); International Multilateral Negotiation: Approaches to the Management of Complexity 1–10, 213–22 (I. William Zartman ed., 1994). 63. Such a problem arose during the negotiation of the failed OECD Multilateral Agreement on Investment, conducted between 1995 and 1998. See Glen Kelley, Note, Multilateral Investment Treaties: A Balanced Approach to Multinational Corporations, 39 Colum. J. Transnat’l. L. 483, 494–98 (2001). 64. Investment protection and investment liberalization are also distinct concepts. Investment liberalization refers to facilitating the entry and operation of foreign investments in the host country. Investment protection refers to protecting the investment,

122 jeswald w. salacuse and nicholas p. sullivan

of BITs, investment promotion for host countries means the attraction of investment projects that the host country determines are in its best interests. Investment liberalization, on the other hand, is a favorite term of capital-exporting countries and generally means creating a climate in which investors may undertake investments that investors judge to be in their interests. For example, a host country government might actively seek to promote investments in the electronics industry, which it judged would foster the future development of its economy but which were not yet present in the country. At the same time, it may desire to impede investment in the retail industry, which is already served by politically powerful local entrepreneurs who fear foreign competition. In such a situation, the developing country, through its treaty relationships and its internal legislation, would be following a policy of investment promotion but not one of investment liberalization. If the three fundamental goals of the BIT movement are investment protection, promotion, and liberalization, one may well ask whether the 2,200 treaties negotiated over the past five decades have achieved these goals. To what extent have BITs actually protected, liberalized, and promoted foreign investment? The answer to this question is important. The continued vitality of the BIT movement and the prospects for creating a global, multilateral legal structure for foreign investment similar to the one that exists for international trade will be influenced by how the concerned countries evaluate the benefits and costs of the BIT process. The following three parts of this chapter will examine each goal separately to determine the extent to which BITs have attained these three objectives. C. An Evaluation of Bit Goal No. 1: Investment Protection Most BITs pursue the objective of investment protection by establishing rules about the host country’s treatment of foreign investment and processes for enforcing these rules. The rules restrain the ability of host governments in dealing with foreign investors and investments. The enforcement process provides an international mechanism outside the jurisdiction of the host country to enforce the rules in cases of dispute. Although the precise provisions of BITs are not uniform and some BITs restrict host country governmental action more than others, virtually all BITs address the same issues. One of the functions of the BIT movement since its inception has been to define in some detail what an

once it has entered the country, from actions by governments and others that would interfere with investor property rights and the functioning of the investment in general. For example, in launching negotiations for a Multilateral Agreement on Investment in September 1995, the OECD mandate called for “a broad multilateral framework for international investment with high standards of liberalisation of investment regimes and investment protection.” OECD, Multilateral Agreement on Investment: Launch of the Negotiations:1995 CMIT/CIME Report and Mandate, available at http://www1.oecd.org/ daf/mai/htm/cmitcime95.htm (last visited Nov. 30, 2004).

do bit s really work?: an evaluation of bilateral investment treaties 123

investment treaty should be in order to create an agreed-upon legal framework for the protection of foreign investment, despite variations of that framework from BIT to BIT. The basic structure of any BIT encompasses eight topics: 1. Scope of application 2. Conditions for the entry of foreign investment 3. General standards of treatment of foreign investments 4. Monetary transfers 5. Operational conditions of the investment 6. Protection against expropriation and dispossession 7. Compensation for losses 8. Investment dispute settlement These topics will be examined below through the lens of investment protection in order to illustrate whether BITs have been successful in achieving this goal. 1. Scope of application Key elements in any BIT are its provisions on the scope of application, that is, the definition of the investors and the investments that may benefit from the treaty. The rules on scope of application are generally found at the beginning of the BIT in sections defining “investors,” “companies,” “nationals,” “investments,” and “territory.”65 As a result of entering a BIT, a contracting country owes obligations only to investors of other contracting countries that make investments in its territory. A contracting country, therefore, owes no obligations under a BIT to people or investments that do not come within the definitions of these terms as defined in the treaty document. In defining the nature of covered investments, most BITs take four basic definitional dimensions into consideration: (1) the form of the investment; (2) the area of the investment’s economic activity; (3) the time when the investment is made; and (4) the investor’s connection with the other contracting country. Most BITs define the concept of investment broadly so as to include various investment forms: tangible and intangible assets, property, and rights. Their approach is to give the term “investment” a broad, nonexclusive definition, recognizing that investment forms are constantly evolving in response to the creativity of investors and the rapidly changing world of international finance. The effect is 65. See, e.g., Agreement for the Liberalization, Promotion and Protection of Investment, Japan-Vietnam, art. 1, Nov. 14, 2003 (defining “investor,” “investments,” and “Area”), available at http://www.unctad.org/sections/dite/iia/docs/bits/japan_vietnam.pdf (last visited Nov. 19, 2004); Treaty Concerning the Reciprocal Encouragement and Protection of Investment, U.S.-Czech Rep., arts. 1(a)–(b), Oct. 22, 1991 (defining “investment” and “company of a Party”), available at http://www.unctad.org/sections/dite/iia/docs/bits/ czech_us.pdf (last visited Nov. 19, 2004); Treaty concerning the Reciprocal Encouragement and Protection of Investments, U.S.-Turk., arts. 1(a), (c), (e), Dec. 3, 1985 (defining “company,” “Investment,” and “national”), available at http://www.unctad.org/sections/dite/ iia/docs/bits/us_turkey.pdf (last visited Nov. 19, 2004).

124 jeswald w. salacuse and nicholas p. sullivan

to provide an expanding umbrella of protection to investors and investments. Despite the breadth of language, the issue of what is and is not an investment is important. In a recent case, the ICSID tribunal found that expenditures made in Sri Lanka by a U.S. firm in furtherance of a contemplated investment that did not materialize were not investments for purposes of either the United States-Sri Lanka BIT or the ICSID Convention.66 A further issue faced in BIT negotiations is whether investments made prior to the treaty will benefit from its provisions. Developing countries have sometimes sought to limit a BIT’s application to future investment only or at least to those investments made in the relatively recent past.67 Viewing the BIT primarily as an investment promotion mechanism, they have claimed to see little purpose in granting additional protection to investments already in the host country. Moreover, they argue that their governmental authorities might not have approved such investments had they realized that the investor’s rights and privileges would later be expanded by a BIT.68 Capital-exporting countries, on the other hand, have generally sought to protect all investments made by their nationals and companies, regardless of when they were made. For example, Article XII of the model BIT used by the United States in its negotiations provides: “[This Treaty] shall apply to investments existing at the time of entry into force as well as to investments made or acquired thereafter.”69 Most BITs seek to continue to provide protection to an investor once a host country has terminated or withdrawn from the treaty. This continuing effects provision protects investors who have made investments based on the expectation

66. See Mihaly Int’l Corp. v. Sri Lanka, ICSID Case No. ARB/00/2 (Mar. 5, 2002), 17 ICSID Rev.—Foreign Inv. L.J. 142, 159 (2002). 67. See, e.g., Agreement for the Promotion and Protection of Investments, U.K.-Indon., art. 2(3), Apr. 27, 1976, 1074 U.N.T.S. 195 (“The rights and obligations of both Contracting Parties with respect to investments made before 10 January 1967 shall be in no way affected by the provisions of this Agreement.”), available at http://www.unctad.org/sections/dite/ iia/docs/bits/uk_indonesia.pdf (last visited Nov. 19, 2004). 68. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 42. 69. Model Bilateral Investment Treaty (BIT) and Sample Provisions From Negotiated BITs, in 1 Basic Documents in International Economic Law 655, 662 (Stephen Zamora & Ronald Brand eds., 1990) [hereinafter Model BIT]. In February 2004, the U.S. State Department released a new and more detailed model BIT. See Press Release, U.S. State Dep’t, Update of U.S. Model Bilateral Investment Treaty (Feb. 5, 2004), available at http:// www.state.gov/e/eb/rls/prsrl/2004/28923.htm (last visited Nov. 19, 2004). The new model BIT follows the approach of its predecessor, providing that, “‘covered investment’ means, with respect to a Party, an investment in its territory of an investor of the other Party in existence as of the date of entry into force of this Treaty or established, acquired, or expanded thereafter.” See U.S. State Department, 2004 Model BIT (DRAFT), § A, art. 1 (2004), available at http://www.state.gov/documents/organization/29030.doc (last visited Nov. 6, 2004) [hereinafter New Model BIT].

do bit s really work?: an evaluation of bilateral investment treaties 125

of treaty protection. The usual period of continued protection is between fifteen and twenty years. Defining which investors can benefit from the treaty is an important issue, since the goal of the contracting country is to secure benefits for its own nationals, companies, and investors, rather than those of other countries. The problem is essentially one of determining what link needs to exist between an investor and a party to a treaty in order for the investor to benefit from the treaty’s provisions. In the case of physical persons, the task is not difficult because virtually all BITs rely on nationality or citizenship, a status that generally is easily determined. For investors that are companies or other legal entities, the problem of determining an appropriate link with a contracting country is more complex. Such legal forms may be created and owned by people who have no real connection with either country that is a party to the treaty. In particular, three types of cases raise problems in this respect: (1) companies organized in a treaty country by nationals of a nontreaty country; (2) companies organized in a nontreaty country by nationals of a treaty country; and (3) companies in which nationals of a nontreaty country hold a substantial interest. For a company to be covered by the treaty, most BITs require that a treaty partner is at least one of the following: (1) country of the company’s incorporation;70 (2) country of the company’s seat, registered office, or principal place of business;71 or (3) country whose nationals have control over, or a substantial interest in, the company making the investment.72 Sometimes these requirements are combined so that an investing company must satisfy two or more to qualify for coverage under the BIT. 70. See, e.g., Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Sri Lanka, art. 1(b), Sept. 20, 1991 (“‘[C]ompany’ of a Party means any kind of corporation, company, association, partnership or other organization, legally constituted under the laws and regulations of a Party or a political subdivision thereof.”), available at http://www.unctad.org/sections/dite/iia/docs/bits/us_srilanka.pdf (last visited Nov. 19, 2004). BITs concluded by Denmark, the Netherlands, the United Kingdom and the United States are frequently of this type. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 39. 71. See, e.g., Treaty Concerning the Encouragement and Reciprocal Protection of Investments, F.R.G.-Swaz., art. 1(4)(a), Apr. 5, 1990 (“The term ‘companies’ means . . . in respect of the Federal Republic of Germany: any juridical person as well as any commercial or other company or association with or without legal personality having its seat in the German area of application of this Treaty, irrespective of whether or not its activities are directed at profit.”), available at http://www.unctad.org/sections/dite/iia/docs/bits/ germany_swaziland.pdf (last visited Nov. 19, 2004). BITs concluded by Belgium, Germany and Sweden are frequently of this type. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 40. 72. See, e.g., Agreement on Encouragement and Reciprocal Protection of Investments, Lith.-Neth., art. 1(b)(iii), Jan. 26, 1994: The term ‘investor’ shall comprise with regard to either contracting party: . . . (iii). legal persons not constituted under the law of that Contracting Party but controlled, directly

126 jeswald w. salacuse and nicholas p. sullivan

2. Conditions for the entry of foreign investment The conditions of entry are matters more closely related to investment promotion and investment liberalization than to investment protection. Conse-quently, this aspect of BITs will be considered below in the context of investment liberalization. 3. General standards of treatment of foreign investments The totality of obligations that a host country owes a foreign investor or investment after the investment is made is generally referred to in BITs as the treatment owed to the investor or the investment. BITs stipulate the standard of treatment that a host country must accord to a foreign investment in two respects. They define certain general standards of treatment and also state specific standards for particular matters, such as monetary transfers, the employment of foreign personnel, and the resolution of disputes with the host government. This section of the chapter will examine the general treatment standards, while succeeding sections will discuss treatment with regard to specific matters. One may divide the general standards of treatment of most BITs into six component parts: a) fair and equitable treatment; b) the provision of full protection and security; c) protection from unreasonable or discriminatory measures; d) treatment no less than that accorded by international law; e) requirement to respect obligations made to investors and investments; and f) national and/or most-favored-nation treatment. An individual BIT may provide for some or all of these treatment standards. a. Fair and equitable treatment. One of the most common standards of treatment found in BITs is an obligation that the host country accord foreign investment “fair and equitable treatment.”73 This is a classic formulation of international law and has been the subject of much commentary and country practice.74 Nonetheless, its precise meaning in a specific situation has been open to varying interpretations. b. Full protection and security. Another general standard of treatment found in most BITs is the obligation of the host country to accord “full protection and security” or “constant protection and security” to investments made by nationals and companies of its treaty partners. Two cases interpreting BIT

or indirectly, by natural persons as defined in (i) [of the Contracting Party’s nationality] or by legal persons as defined in (ii) [legal persons constituted under the law of the Contracting Party] above, who invest in the territory of either Contracting Party. available at http://www.unctad.org/sections/dite/iia/docs/bits/netherlands_lithuania. pdf (last visited Nov. 13, 2004). “Ownership or control,” as these provisions are called, are used in BITs concluded by the Netherlands, Sweden and Switzerland. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 39. 73. Mohamed I. Khalil, Treatment of Foreign Investment in Bilateral Investment Treaties, 7 ICSID Rev.-Foreign Inv. L.J. 339, 351 (1992). 74. UN Ctr. on Transnational Corp. (“UNCTC”), Bilateral Investment Treaties 41–45, UN Doc. ST/CTC/65 (1988).

do bit s really work?: an evaluation of bilateral investment treaties 127

provisions on this point held that this standard does not make the host country responsible for all injuries that befall the investment.75 Thus, although the host country is not a guarantor, it is liable when it fails to show due diligence in protecting the investor from harm. A definition of due diligence that was cited favorably by an ICSID arbitral tribunal defines it as “reasonable measures of prevention which a well-administered government could be expected to exercise under similar circumstances.”76 Consequently, the failure by a host country to take reasonable measures to protect the investment against threats, such as those from brigands or from violence by police and security officers, might render the country liable for compensation of an injured investor under a BIT. c. Unreasonable or discriminatory measures. Many BITs contain language to the effect that “no Contracting Party shall in any way impair by unreasonable or discriminatory measures the management, maintenance, use, enjoyment or disposal” of an investment.77 The specific application of this provision to the individual case will depend on the facts involved; however, it is worth noting that the use of the term “unreasonable” may give host countries grounds to defend actions that they may take against foreign investors. d. International law. Many BITs provide that in no case should foreign investments be given less favorable treatment than that required by international law. Thus, this constitutes the very minimum standard of treatment. Its application in individual cases will also be subject to a variety of interpretations, particularly on those issues where there is significant dispute by developing countries, such as those that occurred during the efforts to secure a New International Economic Order.78 A further question is whether the reference to “international law” is limited only to customary international law or whether treaty provisions and general principles on investments are also to be considered. e. Contractual obligations. To the extent that a contracting party has entered into obligations with an investor or investment, many BITs require a signatory country to respect those obligations. This provision then acts as counter to the claim, advanced during the era of the New International Economic Order,79 that host 75. Compare Asian Agric. Prod. Ltd. v. Sri Lanka, ICSID Case No. ARB/87/3 (June 27, 1990), 6 ICSID Rev.–Foreign Inv. L.J. 526 (1991)(interpreting the words “full protection and security” in the U.K.-Sri Lanka BIT), with Elettronica Sicula S.P.A.(Elsi) (U.S. v. Italy), 1989 I.C.J. 15 (July 20) (interpreting the words “constant protection and security” in the United States-Italy Treaty of Friendship, Commerce and Navigation). 76. Asian Agric. Prod. Ltd., 6 ICSID Rev.–Foreign Inv. L.J., at 558. (citing Alwyn V. Freeman, Responsibility of States for Unlawful Acts of Their Armed Forces 15–16 (1957)). 77. See, e.g., Treaty concerning the Reciprocal Encouragement and Protection of Investment, U.S.-Turk., supra note 65, art. 2(3) (“Neither Party shall in any way impair by arbitrary or discriminatory measures the management, operation, maintenance, use, enjoyment, acquisition, expansion, or disposal of investments.”). 78. See Waelde, supra note 10, at 771; see also supra text accompanying note 10. 79. See id.

128 jeswald w. salacuse and nicholas p. sullivan

countries should be able to unilaterally revise contracts that they have made with foreign investors. It may also mean that, as a result of the BIT, contracts between the foreign investor and the host government, which are normally subject only to host country law, are also to be governed by international law.80 f. National and/or most-favored-nation treatment. In addition to these general standards, many BITs also contain a further refinement—nondiscrimination in relation to other investors, both foreign and national. Thus, they provide for national treatment, which requires that a host country treat an investor or an investment once made, no less favorably than they treat their own national investors or investments made by their own nationals. As a result, a BIT may also provide for most-favored-nation treatment, which means that a host country may not treat an investor or investment from a BIT partner any less favorably than it treats investors or investments from any other country. One consequence of such a provision is that it allows the foreign investor to take advantage of the highest standard of treatment provided to a country in any BIT to which the host country is a party.81 Certain BITs, like those contracted by the United States, combine both of these standards and require host countries to grant investors national treatment or most-favored-nation treatment, whichever is the more favorable. Some developing countries, recognizing the disparity in financial and technological resources between their national enterprises and those of foreign companies, have resisted or sought to limit the scope of the national treatment guarantee in BITs. In particular, they have tried to avoid giving foreign investors the benefits and subsidies designed to strengthen national industries.82 4. Monetary transfers For any foreign investment project, the ability to repatriate income and capital, to pay foreign obligations in another currency, and to purchase raw materials and spare parts from abroad is crucial to a project’s success. For this reason, capital-exporting countries in BIT negotiations have pressed for unrestricted freedom for investors to undertake these monetary operations. Such operations are collectively referred to as “transfers.”83 The monetary transfer provisions of most BITs deal with five basic issues: (1) the general nature of the investor’s rights to make monetary transfers; (2) the types of payments that are covered by the right to make transfers; (3) the currency with which the payment

80. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 56–57. 81. See, e.g., Emilio Agustín Maffezini v. Spain, ICSID Case No. ARB/97/7 (Nov. 13, 2000), 16 ICSID Rev.—Foreign Inv. L.J. 248 (2001) (permitting Argentine national bringing claim against Spain under Argentina-Spain BIT to have benefit of less stringent procedural requirements of Chile-Spain BIT). 82. See, e.g., UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 64–65. 83. Khalil, supra note 73, at 360.

do bit s really work?: an evaluation of bilateral investment treaties 129

may be made; (4) the applicable exchange rate; and (5) the time within which the host country must allow the investor to make transfers. However, developing countries that face chronic balance-of-payments difficulties and the need to conserve foreign exchange to pay for essential goods and services have considerably reduced their ability and willingness to grant foreign investors the unrestricted right to make such monetary transfers.84 Accordingly, many developing countries have exchange-control laws to regulate the conversion and transfers of currency abroad.85 As a result of this fundamental conflict in goals, the negotiation of BIT provisions on monetary transfers has often been among the most difficult to conclude. Capital-exporting countries seek broad, unrestricted guarantees on monetary transfers, while developing countries press for limited guarantees subject to a variety of exceptions. 5. Operational conditions BITs sometimes provide treatment standards with respect to certain operational conditions, such as the investor’s right to enter the country, employ foreign nationals, and be free of performance requirements. One of the most important conditions, of course, is for the investor’s employees to be able to enter the host country and to manage and operate the investment. Most BITs do not grant the investor an automatic right to enter and stay in the host country. German BITs, for example, provide that each contracting party will give “sympathetic consideration” to applications for entry,86 and U.S. BITs give “nationals” of contracting parties the right to enter the other contracting country for purposes of establishing or operating investments subject to the laws of the host country.87 6. Compensation for losses from armed conflict or internal disorder Many BITs also deal with losses to an investment due to armed conflict or internal disorder; however, they do not normally establish an absolute right to compensation. Instead, many promise that foreign investors will be treated in the same manner as nationals of the host country with respect to compensation.88 Some treaties may also provide for most-favored-nation treatment on this question. The ICSID case of Asian Agricultural Products Ltd. v. Sri Lanka89 is one of the few 84. See, e.g., Time to Turn off the Tap?, Economist, Sept. 12, 1998, at 83. 85. Jeswald W. Salacuse, Host Country Regulation and Promotion of Joint Ventures and Foreign Investment, in International Joint Ventures: A Practical Approach to Working with Foreign Investors in the U.S. and Abroad 107, 122–23 (David N. Goldsweig & Roger H. Cummings eds., 1990). 86. Treaty Concerning the Encouragement and Reciprocal Protection of Investments, F.R.G.-Swaz., supra note 71, ad art. (3)(c). 87. See, e.g., Treaty Concerning the Encouragement and Reciprocal Protection of Investment, U.S.-Sri Lanka, supra note 70, art. 2(3). 88. See, e.g., Agreement for the Promotion and Reciprocal Protection of Investments, U.K.-Ukr., art. 5, Feb. 10, 1993, available at http://www.unctad.org/sections/dite/iia/docs/ bits/uk_ukraine.pdf (last visited Nov. 19, 2004). 89. Asian Agric. Prod. Ltd., supra note 75, at 526.

130 jeswald w. salacuse and nicholas p. sullivan

cases that have considered this provision in detail in connection with a dispute between an injured investor and a host country government. The tribunal concluded that in addition to any specific compensatory actions taken for the benefit of other investors, this provision would also make applicable to an injured investor any promised higher standard, for example in another BIT, granted to investors from other countries. 7. Protection against dispossession One of the primary functions of any BIT is to protect foreign investments against nationalization, expropriation, or other forms of interference with property rights by host country governmental authorities. Despite opposition by some developing nations in various multilateral forums, virtually all BITs with developing countries adopt some variation of the traditional Western view of international law that a country may not expropriate property of an alien except: (1) for a public purpose; (2) in a nondiscriminatory manner; (3) upon payment of just compensation, and in most instances; (4) with provision for some form of judicial review. The various elements of the traditional rule have taken different formulations in different treaties, some more and some less protective of investor interests. Perhaps the greatest variations in treaty provisions and some of the most difficult negotiations arise with respect to the standard of compensation. Nonetheless, many, if not most, BITs have adopted the traditional rule, expressed in the socalled “Hull Formula”90 that such compensation must be “prompt, adequate and effective.”91 They then proceed to define the meaning of each of these words.92 8. Investment dispute settlement The seven issues discussed above form the protective architecture of the BIT. In theory at least, the scope of protection seems broad in that these seven issues govern most, if not all, of the foreign investor’s principal areas of concern with respect to the political risks associated with a foreign investment. A fundamental, practical question, of course, is whether BIT countries actually respect their treaty commitments and, if not, whether an injured investor has effective legal redress against a host country’s violations of an applicable BIT. Unfortunately, substantial, systematic evidence on whether proposed or contemplated government actions against foreign investment have actually been constrained or prevented by BIT provisions is not easily accessible. The limited evidence that is available on this point is anecdotal and far from comprehensive, given the vast numbers of governmental actions that one would have to examine. Therefore, the only available information on the

90. See 3 Green H. Hackworth, Digest of International Law 655–64 (1942). 91. See, e.g., UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 69. 92. See, e.g., Agreement for the Promotion and Reciprocal Protection of Investments, U.K.-Costa Rica, art. 5, Sept. 7, 1982, available at http://www.unctad.org/sections/dite/ iia/docs/bits/costarica_uk_sp.pdf (last visited Nov. 13, 2004).

do bit s really work?: an evaluation of bilateral investment treaties 131

protective effect of BITs lies in the actual cases brought by investors under BIT provisions against host countries. For foreign investors and their governments, one of the great deficiencies of customary international law has been its lack of effective and binding mechanisms for the resolution of investment disputes. One aim of the BIT movement has been to remedy this situation. Most BITs provide for two distinct dispute settlement mechanisms: one for disputes between the two contracting countries, and another for disputes between a host country and an aggrieved foreign investor. With respect to the former, BITs stipulate that in the event of a dispute over the interpretation or application of the treaty, the two countries concerned will first seek to resolve their differences through negotiation and then, if that fails, through ad hoc arbitration. With respect to the latter, the trend among more recent BITs is also to provide a separate international arbitration procedure, often under the auspices of ICSID, for the settlement of disputes between an aggrieved foreign investor and the host country government. By concluding a BIT, the two countries, in most cases, give the required consent needed to establish ICSID or other arbitral jurisdiction in the event of a future dispute between one signatory and a national of the other signatory. Although the investor must first try to resolve the conflict through negotiation and may also have to exhaust remedies available locally, it ultimately has the power to invoke compulsory arbitration to secure a binding award.93 According a private party the right to bring an action in an international tribunal against a sovereign country with respect to an investment dispute is a revolutionary innovation that now seems to be taken for granted. Yet its uniqueness and power should not be overlooked. The field of international trade law, for example, contains no similar procedure. Violations of trade law, even though they strike at the economic interests of private parties, are matters resolved directly and solely by countries. The World Trade Organization (WTO) does not give a remedy to private persons injured by trade law violations.94 It should also be noted that BITs grant aggrieved investors the right to prosecute their claims autonomously, without regard to the concerns and interests of their source countries. It is this mechanism that gives important, practical significance to BITs, a mechanism that truly enables these bilateral treaties to afford protection to foreign investment. As a result, foreign investors are bringing increasing numbers of claims in arbitration when they believe that host countries have denied them promised protection under a BIT. BIT cases accounted for five of twelve ICSID arbitrations in 2000, twelve of fourteen the following year, and a full fifteen of

93. See, e.g., Agreement for the Liberalization, Promotion and Protection of Investment, Japan-Vietnam, supra note 65, art. 13. 94. See, e.g., Glen T. Schleyer, Power to the People: Allowing Private Parties to Raise Claims Before the WTO Dispute Resolution System, 65 Fordham L. Rev. 2275, 2277 (1997).

132 jeswald w. salacuse and nicholas p. sullivan

nineteen in 2002.95 As of February 2003, ICSID had registered a total of eightyseven BIT cases and an additional ten under NAFTA.96 Of this number, ICSID tribunals had rendered eighteen final awards in BIT cases, of which the investor prevailed in ten, and six final awards in NAFTA cases of which the investor prevailed in two.97 An additional three BIT cases were concluded through settlement.98 To understand the protective force of BITs, one must also take into account numerous non-ICSID institutional and ad hoc arbitrations brought by injured investors claiming a violation of legal rights conferred by BITs. Accurate, systematic data on the number and nature of such non-ICSID arbitrations is unfortunately not available, primarily because of the confidentiality rules of the institutions concerned. Press accounts and anecdotal evidence of individual cases suggest that the number of claims based on BITs is growing.99 One notable recent example is the case of CME Czech Republic B.V. v. Czech Republic,100 a United Nations Commission on International Trade Law arbitration under the Netherlands-Czech Republic BIT, which resulted in an award and payment of $355 million to an injured investor, one of the largest awards ever made in an arbitration proceeding.101 One effect of that award, along with others rendered against sovereign countries in favor of individual private investors, is to cause host countries to take their BIT responsibilities more seriously. The BIT treaty provisions, their enforcement mechanisms, and the fact that arbitral tribunals hold host countries accountable, constitute an external discipline upon governments’ behavior in their relations with foreign investors. This results in a relatively effective system of foreign investment protection. It is also to be noted that decisions of arbitral tribunals, although not systematically made public, tend to take the form of lengthy, reasoned, and scholarly decisions that form part of the

95. See Luke Eric Peterson, Int’l Inst. for Sustainable Dev., Research Note: Emerging Bilateral Investment Treaty Arbitration and Sustainable Development, 3 (2003), available at http://www.iisd.org/pdf/2003/trade_bits_disputes.pdf (last visited Nov. 19, 2004). 96. Int’l Bank for Reconstruction and Dev., World Development Report 2005: A Better Investment Climate for Everyone 181 (2004). 97. Id. 98. Id. 99. See generally Int’l Inst. for Sustainable Dev., Invest-SD: Investment and Sustainable Development News Bulletin, (providing information on pending and recent investment arbitration cases), available at http://www.iisd.org/investment/invest-sd (last visited Nov. 22, 2004). 100. CME Czech Republic B.V. v. Czech Rep., Mar. 14, 2003, available at http://www. cetv-net.com/iFiles/1439-Final_Award_Quantum.pdf (last visited Nov. 19, 2004). The other opinions and awards in this case are available at CME Central European Media Enterprises: Arbitration Awards, at http://www.cetv-net.com/arbitration.asp (last visited Nov. 19, 2004). 101. Peter S. Green, Czech Republic Pays $355 Million to Media Concern, N.Y. Times, May 16, 2003, at W1.

do bit s really work?: an evaluation of bilateral investment treaties 133

jurisprudence of this emerging international investment law and serve to solidify and give force to BIT provisions. Although the more than 2,200 BITs concluded since 1959 tend to cover the same issues, they differ in how they treat those issues. Some are more protective than others. For example, the BITs negotiated by the United States generally exhibit higher standards of protection than the BITs of many other countries.102 Nonetheless, despite divergences among individual treaties, BITs as a group also demonstrate many commonalities, including their coverage of similar issues and their use of equivalent or comparable legal concepts and vocabulary. It is these commonalities that are contributing to the creation of an international framework for investment. Moreover among more recent BITs, one detects increasing consensus on certain points; for example, all BITs now require the payment of compensation for expropriation; however, the formulas used to determine compensation in recent treaties vary from country to country.103 After reviewing the nature and scope of BIT provisions, the strength of related enforcement mechanisms, and the actual cases brought against host countries by aggrieved investors, one may conclude that BITs have achieved their first goal of fostering investment protection. While that protection is not absolute (no legal device provides absolute protection), investors and investments covered by a BIT certainly enjoy a higher degree of protection from the political risks of governmental intervention than those that are not. D. An evaluation of BIT goal no. 2: investment and market liberalization The ideal of economic liberalism holds that the market, not governmental laws and regulations, should determine economic decisions.104 Beginning in the post–World War II period, virtually all developing countries rejected the liberal economic model and believed that their governments had the primary responsibility for bringing about national economic development.105 As a result, their systems were characterized by: (1) state planning and public ordering of their economies and societies; (2) reliance on state enterprises as economic actors; (3) restriction and regulation of the private sector; and (4) governmental limitation and control of international economic transactions, especially foreign investment.106 By the mid-1980s, this approach to development began to lose its hold on the minds and actions of policy makers, aid agencies, and international

102. Juillard, supra note 17, at 211 (asserting that the level of protection achieved by U.S. BITs is superior to the level of protection achieved by European BITs). 103. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 69. 104. See generally George T. Crane & Abla Amawi, The Theoretical Evolution of International Political Economy (2d ed. 1997) (describing the history and development of the concept of economic liberalism) 105. See Int’l Bank for Reconstruction and Dev., World Development Report: The State in a Changing World, 1–2 (1997). 106. Salacuse, supra note 46, at 877–80.

134 jeswald w. salacuse and nicholas p. sullivan

financial institutions. Developing countries increasingly privatized their state enterprises, engaged in deregulation, and opened their economies.107 In short, they embarked on a process of economic liberalization. As indicated earlier in this chapter, one of the goals of many countries, particularly the United States, in negotiating BITs with developing countries was to encourage market and investment liberalization.108 An evaluation of the extent to which BITs have achieved the goal of market and investment liberalization depends on one’s definition of the concept of liberalization. Two definitional approaches to liberalization present themselves. The first, which could be called the “absolutist approach,” seeks to determine how well the actual situation meets the liberal economic model; the second, which could be referred to as “the relativist approach,” aims primarily to determine the extent to which the existing situation in a country has moved away from the preexisting command economy system toward the liberal model. 1. An absolutist evaluation of liberalization With respect to foreign investment, applying an absolutist approach to economic liberalization would mean that foreign investors would not be subject to legal or regulatory constraints in undertaking investments in the country concerned. In fact, probably no country, either in the developed or the developing world, has taken the absolutist position,109 and BITs have not served to deny countries the right to control the entry of foreign investment. Many BITs make a distinction between the treatment to be accorded an investor in making an investment (preestablishment) and the treatment to be given after the investment is made (postestablishment). With respect to the former, BITs generally contain a provision to the effect that “each Contracting Party shall encourage and create favorable conditions for investors of the other Contracting party to make investments in its territory.”110 Despite the inclusion of such provisions, no BIT requires a host country to admit any and all investments proposed by an investor from the other treaty country. Most countries have special laws governing the entry of foreign capital,111 and BITs generally provide that host 107. Id. at 882–86. 108. See supra text accompanying notes 56–58. 109. Even the United States, which strongly supports the liberal economic model, restricts or limits the ability of foreigners to invest in certain areas, including commercial aviation, telecommunications, and maritime industries. Moreover, several states restrict the ability of foreigners to own real estate. Roger H. Cummings, United States Regulation of Foreign Joint Ventures and Investment, in International Joint Ventures: A Practical Approach to Working with Foreign Investors in the U.S. and Abroad 137, 139 (David N. Goldsweig & Roger H. Cummings eds., 1990). 110. E.g., Agreement for the Promotion and Protection of Investments, India-U.K., art. 3(1), Mar. 14, 1994, 34 I.L.M. 935, 940 (1995) 111. See generally Salacuse, supra note 85, at 107–36 (explaining how host countries regulate joint ventures and the effect of such regulation on their operation and formation).

do bit s really work?: an evaluation of bilateral investment treaties 135

countries may admit investment in their territories in accordance with their laws.112 In effect, no BIT ever guarantees investors of a contracting country access to the other contracting country’s markets.113 A common provision is that the host country “shall admit investments in conformity with its laws.”114 Consequently, one must conclude that the BIT movement has not been effective in attaining the goal of absolute investment liberalization, if by that term one means establishing by treaty a completely open door to investment from a BIT treaty partner.115 This is unsurprising when one considers that no BIT has expressly adopted such an objective. Investment and market liberalization are better characterized as consequences the developed country treaty partner hopes for when it enters a BIT. From an absolutist point of view, while the use of BITs “affirm[s] liberal economic theory” and supports the adoption of liberalizing policies, the treaties are not actually designed to create a liberal investment regime; rather, BITs are driven more by motives of economic nationalism than they are economic liberalism.116 Indeed “[t]he interventionist measures permitted by the BITs are antithetical to economic liberalism.”117

112. E.g., Agreement for the Promotion and Reciprocal Protection of Investments, Hung.-U.K., art. 2.1, Mar. 9, 1987, 1990 U.K.T.S. 44 (Cm. 1103), reprinted in 4 ICSID Rev.Foreign Inv. L.J. 159, 160 (1989) (“Each Contracting Party . . . , subject to its right to exercise powers conferred by its laws, shall admit . . . capital [of the other contracting party].”). 113. Kenneth J. Vandevelde, Investment Liberalization and Economic Development: The Role of Bilateral Investment Treaties, 36 Colum. J. Transnat’l L. 501, 511 (1998). 114. See, e.g., Free Trade Area of the Americas, Investment Agreements in the Western Hemisphere: A Compendium (Oct. 14, 1999) (“The most representative clause reads as follows: Each Contracting Party shall promote, in its territory, investments of investors of the other Contracting Party and shall admit such investments in accordance with its laws and regulations.”), available at http://www.ftaa-alca.org/ngroups/ngin/publications/ english99/compinv1.asp. (last visited Nov. 22, 2004) 115. Professor Vandevelde has concluded: BITs are very limited tools for liberalization. Access provisions are subordinate to local law; nondiscrimination provisions apply only post establishment of investment and are subject to exceptions; security is afforded against certain types of state interference, but generally not against private interference; dispute provisions apply only to public, not private, disputes; and transparency provisions are rare. Vandevelde, supra note 113, at 514. 116. Kenneth J. Vandevelde, The Political Economy of a Bilateral Investment Treaty, 92 Am. J. Int’l L. 621, 633 (1998). According to Vandevelde, a liberal economic model of a BIT would do a better job reflecting “investment neutrality” (i.e., state nonintervention in cross-border investment flows) and “market facilitation” (enabling the state to current market failures). Id. at 633–35. As they stand now, however, BITs are more about “protecting the interests of home state investors and preserving the political prerogatives of the host state” than they are about improving economic efficiency. Id. at 634. 117. Id. at 634.

136 jeswald w. salacuse and nicholas p. sullivan

2. A relativist evaluation of liberalization Viewing the significant changes in many developing countries over the last twenty years as they have sought to transform themselves into emerging markets,118 it is clear that their economies have experienced significant liberalization, even though they have not attained the ideal liberal model. Their laws and regulations governing foreign direct investment, in particular, have been liberalized as a general phenomenon. For example, a study by the U.N. Conference on Trade and Development (UNCTAD) found that during the period between 1991 and 2002, “1,551 (95%) out of 1,641 changes introduced by 165 countries in their FDI laws were in the direction of greater liberalization.”119 At this point, it is difficult to determine the precise role that BITs have played in this liberalization process. A study seeking to correlate the timing and number of BITs signed by individual countries with the timing and number of their liberalizing reforms would shed some important light on this question. In general terms, it is interesting to note, however, that during the period measured by the UNCTAD study, BITs experienced their most significant expansion in both number and geographic coverage.120 The link between BITs and liberalization may be both direct and indirect. The direct link may be found in some BIT provisions that have may have a liberalizing effect. For example, in the negotiation of some BITs, capital-exporting countries, with varying degrees of success, have sought to protect their nationals and companies from unfavorable discrimination by securing treatment on admission that is no less favorable than the treatment given investments made by host country nationals or nationals of a third country. For example, Article II(1) of the U.S. BIT Prototype provides: Each party shall permit and treat investment, and activities associated therewith, on a basis no less favorable than that accorded in like situations to investment and associated activities of its own nationals or companies, or of nationals and companies of any third party, whichever is the most favorable.121 Accordingly, Article II(1) of the United States-Albania BIT grants national treatment or most-favored-nation treatment, whichever is the more favorable (with specified exceptions), “to the establishment, acquisition, [and] expansion . . . of covered investments.”122 The implication of this provision is clear. In deciding on admission of a foreign investment project, the host country must treat

118. See Salacuse, supra note 46, at 875–90. 119. See UNCTAD, supra note 2, at 20; see also World Bank, World Development Report 2005: A Better Investment Climate for Everyone 111–12 (2004), available at http://siteresources.worldbank.org/INTWDR2005/Resources/complete_report.pdf (last visited Nov. 19, 2004). 120. See UNCTAD, supra note 2, at 159–217. 121. Model BIT, supra note 69, at 656. 122. See Investment Treaty with Albania, supra note 58, art. II(1).

do bit s really work?: an evaluation of bilateral investment treaties 137

applications by investors of its treaty partner in the same way it treats applications by its own national investors or investors from other countries. For countries seeking to encourage investments by their own nationals, such a provision may raise problems. For example, the host country may have closed certain sectors to foreign investment for strategic or political reasons. Additionally, many developing countries give special preference to national investors because of their belief that national investors cannot compete on equal footing with foreign firms.123 They, therefore, would probably find it easier to grant mostfavored-nation treatment on the entry of foreign investment than they would national treatment. However, applying the concepts of national treatment and most-favorednation treatment to foreign investment projects, no two of which are exactly alike, is far more difficult than applying them to international trade in fungible goods, where these concepts were first developed. The qualifying words “in like situations” contained in the clause quoted above may allow different treatment with respect to the entry of investments if the projects themselves or the surrounding circumstances are sufficiently dissimilar.124 Moreover, treaties including this type of entry provision also contain a specific list of areas and sectors where foreign investment may be prohibited.125 Nonetheless, to the extent that exceptions to these provisions are relatively limited, it can serve to have a liberalizing effect on the foreign investment regime of the host country in that it gives greater scope for market factors to determine investment decisions and proportionately less scope for governmental decisions. One specific type of discriminatory treatment that host countries often impose in the making and operation of foreign investments is a “performance requirement” or “trade-related investment measure” (TRIM), such as those that require an investment project, as a condition for entry, to export a certain proportion of its production, restrict its imports to a certain level, or purchase a minimum quantity of local goods and services. Although most BITs have not dealt with the question of performance requirements,126 the United States, with some success recently, has sought to protect its investors from them through its BIT negotiations.127 123. World Bank, supra note 119, at 159. See also UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 64–65. 124. See Model BIT, supra note 69, at 656. 125. See, e.g., Treaty Concerning the Reciprocal Encouragement and Protection of Investment, U.S.-Gren., art. II(1), May 2, 1986, available at http://www.unctad.org/sections/dite/iia/docs/bits/us_grenada.pdf (last visited November 13, 2004). 126. UNCTC, supra note 74, at 69. 127. See Model BIT, supra note 69, art. II(5), at 657 (“Neither Party shall impose performance requirement as a condition of establishment, expansion or maintenance of investments, which require or enforce commitments to export goods produced, or which specify that goods or services must be purchased locally, or which impose any other

138 jeswald w. salacuse and nicholas p. sullivan

The concern with performance requirements as measures that unjustifiably burden trade and investment was further developed in the Uruguay Round of the GATT.128 That Round produced an Agreement on TRIMs, forbidding the imposition of measures that are inconsistent with GATT’s Article III on national treatment129 and Article XI on the elimination of quantitative restrictions.130 Its purpose is to prevent WTO members from imposing local content and trade balancing requirements as a condition for the creation or operation of foreign investment projects. BITs may also have an indirect positive effect on liberalization of host country economies. Under certain circumstances, the introduction of FDI can contribute to that liberalization. The demonstrated economic success of particular foreign enterprises, competitive pressures caused by their presence, the governmental desire to attract even more FDI, and the demands by national entrepreneurs to secure treatment equal to the privileges often given to foreigners may create strong pressures for change in host country regulatory systems.131 So to the extent that BITs have encouraged foreign investment in developing countries, they have also contributed indirectly and modestly to market liberalization. Economic liberalization is a complex process that cannot be brought about by any single magic bullet. It requires a host of sound policies, laws, and institutions across a wide domain of human activity.132 BITs are just one policy instrument among many others that may facilitate the process. E. An evaluation of BIT goal no. 3: investment promotion The third declared goal of BITs is investment promotion. A BIT purports to create a symmetrical relationship between the two contracting countries by similar requirements.”) See also Investment Treaty with Albania, supra note 58, art. VI (prohibiting four specified types of performance requirements). 128. See Final Act Embodying the Results of the Uruguay Round of Multilateral Trade Negotiations, Apr. 15, 1994, Legal Instruments—Results of the Uruguay Round, 33 I.L.M. 1125 (1994). 129. General Agreement on Tariffs and Trade, art. III, Oct. 30, 1947, 61 Stat. A-11, T.I.A.S. 1700, 55 U.N.T.S. 194. 130. Id. art. XI. 131. Egypt is an excellent example of this phenomenon. During the time of President Gamal Abdel Nasser, the country was virtually closed to foreign investment. After his death, the Sadat government took a first tentative step toward liberalization by seeking only Arab capital in 1971 and then foreign investment in 1974. Gradually, both policy and law evolved to the point that Egypt was encouraging all private investment, both foreign and domestic. See Jeswald W. Salacuse, Back to Contract: Implications of Peace and Openness for Egypt’s Legal System, 28 Am. J. Comp. L. 315 (1980). See also Jeswald W. Salacuse, Foreign Investment and Legislative Exemptions in Egypt: Needed Stimulus or New Capitulations? in Social Legislation in the Contemporary Middle East 241 (L. Michalak & J. Salacuse eds., 1986). 132. See generally World Bank, supra note 119 (discussing complexities of developing a liberal investment climate).

do bit s really work?: an evaluation of bilateral investment treaties 139

providing that the nationals and companies of either party to the treaty may invest under the same conditions and be treated in the same way in the territory of the other. In reality, of course, in a BIT between an industrialized country and a developing nation, an asymmetry exists between the parties since one country (the industrialized country) will typically be the source and the other country (the developing country) the recipient of that capital. Some developing countries have assumed that their industrialized treaty partners would take affirmative action to encourage the industrialized country’s nationals to invest in the developing country—an expectation no doubt raised by the words “encouragement” and “promotion” in the treaty.133 Capital-exporting countries have steadfastly refused to agree to any provision in a BIT obligating them to encourage or induce their nationals to invest in the territories of their BIT partners. The general premise of BITs is that the goal of investment promotion is to be achieved by the host country’s creation of a stable legal environment that favors foreign investment. The basic working assumption upon which BITs rest is that clear and enforceable rules that protect foreign investors reduce risk, and a reduction in risk promotes investment. No language in a BIT binds a source country to encourage its investors and companies to invest abroad. That said, it is vital to examine the effects BITs have actually had in promoting investment. In light of the many variables that influence investment decisions, it is probably impossible to pinpoint the precise effect of a BIT on an investor’s decision to invest in a given country.134 Local economic conditions and government policies are probably more important than BITs in influencing the investment decision. Indeed, industrialized countries probably sign BITs only with those developing countries whose policies and laws are sufficiently protective of and favorable to foreign investment.135 Thus, the BIT is often a codification, and not a source, of pro-foreign investment policies. On the other hand, by entering into a BIT, an instrument of international law, a signatory country is raising those policies to the level of international law and thereby limiting its ability to change policy easily. BITs therefore have the effect of stabilizing a county’s investment policy and its legal and contractual commitments to individual foreign investors.136 Developing countries have signed BITs on the assumption that these treaties would result in increased flows of FDI. Policy analysts and scholars have accepted that assumption, despite the lack of compelling evidence.137 Given how questionable

133. See supra Treaty Concerning the Reciprocal Encouragement and Protection of Investment, U.S.-Arm.; Treaty concerning the Promotion and reciprocal protection of investments, F.R.G.-Pol., note 55. 134. See Vandevelde, supra note 113, at 524–25. 135. Id. at 523. 136. Id. at 522–25. 137. See, e.g., Andrew T. Guzman, Why LDCs Sign Treaties That Hurt Them: Explaining the Popularity of Bilateral Investment Treaties, 38 Va. J. Int’l L. 639, 679 (1998) (“Any single

140 jeswald w. salacuse and nicholas p. sullivan

this assumption appears, at a minimum, a plausible case needs to exist that there is some relationship between a BIT and the promotion of investment if the momentum of the BIT movement is to continue and perhaps lead to more global forms of investment protection. To evaluate the BITs’ impact on promoting investment, one must turn to empirical econometric research. Without econometric tools, it is impossible to know whether an increase in FDI is a function of a BIT, an improvement in a country’s investment opportunities, or a global increase in capital flows. The econometric approach,138 assuming enough good data, observations, and a logical model that includes all key determinants, allows for comparisons before and after the conclusion of a BIT, or between countries with and without a BIT—while holding all other contributing factors, such as gross domestic product (GDP), exchange rates, and market size, constant. This chapter presents the authors’ own econometric research on this topic, as well as a review of other relevant econometric studies. Before presenting results, we briefly review the basic methods of reading and understanding econometric results. 1. Econometric research Typical econometric analysis begins with the following premise: y and x are two variables, representing some parameters of interest (such as GDP and FDI). We are interested in “explaining y in terms of x,” or in “studying how y changes with changes in x.”139 The changes in both x and y can be modeled to reflect linear one-to-one relationships (for example, if gross domestic product increases by $100 million, then FDI increases by $100,000), or percentage changes (if GDP increases by 3%, FDI increases by 1%). To determine the relationship, one collects as much data as possible (putting a premium on accuracy, consistency, and reliability) and plots it on a graph. Such a graph might indicate, for example, that Argentina, with a certain GDP, attracted a certain FDI in 2000, while Brazil, with a different GDP, attracted a greater or lesser amount of FDI in 2000, and so on. (Another approach, known as “fixed effects” or “time series,” which the authors also use and describe in more detail later, is to chart one set of data—say, from Brazil and the United States—across a number of years). The graph will show multiple data points, some bunched together, some outlying. One can see the larger relationship capital importing country has an incentive to sign a BIT because such a treaty helps that country attract foreign investment.”). 138. Econometric analysis selects a dependent variable (in this case, FDI flows) and a host of explanatory variables, using as many observations of empirical data as possible, to determine which variables have a positive, negative, or neutral effect and with what degree of magnitude. A positive coefficient on an explanatory variable at the 95%–99% confidence level does not prove a positive correlation with the dependent variable. Conversely, it states with a high degree of confidence that the lack of a correlation (known as the “null hypothesis”) is statistically highly unlikely. 139. See, e.g., Jeffrey M. Wooldridge, Introductory Econometrics: A Modern Approach 22 (1999).

do bit s really work?: an evaluation of bilateral investment treaties 141

between GDP and FDI by drawing a line or curve on the graph that comes as close as possible to the mean of all data points. Determining how to achieve the best “fit” for that line is the essence of econometrics. The slope of the line indicates the correlation—positive, negative or zero, and to what degree—between the two variables. It is important to note that no such correlative relationship constitutes a hard-and-fast empirical truth about any one country pair, but is merely the best prediction of how GDP would affect FDI for any two countries selected at random. It is understood that any specific relationship (between, for example, Argentina and the United States) will be unique. In our study of BITs, we first formulate a hypothesis: The presence of a BIT between the United States and a developing country has a positive impact on FDI outflows from the United States to that developing country. In econometric terms, we are looking at the effect of an explanatory variable x (i.e., the presence of a BIT) on a dependent variable y (FDI inflows). We postulate the simple econometric equation or “model” as follows: FDIi = B0 + B1usbit + u

When graphed, the coefficient of the usbit parameter (B1) measures the change in annual FDI inflows (to a given country i) when that country has signed a BIT with the United States. Since we expect the effect to be positive, we expect B1 to be greater than 0 (i.e., B1 > 0). In this and other econometric equations, B0 is a slope intercept that indicates where on the y axis the “fitted line” begins, and has no analytical significance for the present analysis; u is a catch-all error term that includes the effects of all unobserved parameters. The latter is important because it indicates that we can never fully describe in mathematical terms what happens in the real world. Econometric analysis cannot prove this or any other hypothesis because it is working with a random sampling of data that is not all inclusive. But it can reject the null hypothesis—i.e., that there is no relationship between a BIT and FDI flows, given enough good data to provide strong significance levels.140 Once the equation or model is written out, one collects as much empirical data as possible, since any statistical study requires a large body of data to produce significant results. In this simple hypothetical example, one would collect data for the GDP and FDI inflows for more than 100 countries. Then one would input the data and run the regressions (using a computerized software package, such as STATA), which returns a coefficient on each of the measured variables. For example, in our sample equation (FDIi = B0 + B1usbit + u), where FDI is 140. Economists generally reject the null hypothesis at a five percent statistical significance level. This means that they are willing to mistakenly reject the null hypothesis when it is true 5 percent of the time. In some cases, where data are hard to gather, a 10 percent significance level is considered acceptable, although it is a minimum threshold. When large amounts of data are available, a 1 percent significance is ideal and gives economists a “high degree of confidence” in rejecting the null hypothesis. In short, “statistically significant” results are those in the 1%–10% significance range. See generally id. at 113–52.

142 jeswald w. salacuse and nicholas p. sullivan

measured in millions of U.S. dollars, a coefficient of 5.25 for B1 would be interpreted as follows: The presence of a BIT between the United States and another country could be correlated with an additional $5.25 million of FDI per year. The simple equation above, however, is not an accurate reflection of the real world. Numerous other factors besides BITs affect investment flows, and because it would be misleading to ignore them, one includes these variables in the equations to assess their relative impact.141 We are thus studying the effect of various x’s on y, in an effort to achieve a situation in which all other relevant factors are considered, which is known as a ceteris paribus (“other things equal”) effect. These other factors likely include GDP, market size, exports as a percentage of GDP, rule of law, inflation and exchange rates, treaties with other countries, distance between major ports, and cultural or linguistic ties. For example, if the simple equation (FDIi = B0 + B1usbit + u) showed a strong correlation between a BIT and FDI flows, much of that correlation might be a function of a country’s GDP. But what if the United States only signed BITs with countries with large GDPs? Only if GDP is included in the equation would we learn whether the size of a country’s GDP or a BIT had more influence on FDI flows; without GDP in the equation, the BIT variable would pick up all the influence. Similarly, education or other labor force factors might drive FDI. So, if the United States only signs BITs with countries that have strong labor forces, and those factors are not measured, it would appear that FDI was merely a function of the BIT, rather than the labor force. If we could collect data for a large number of countries that all had the same GDP and labor force attributes (and all other important factors), we could then run a simple regression on the BIT variable. Everything else would be the same, except the BIT, and we could measure that impact in a vacuum, as it were. But no two countries are exactly alike. Multiple regression, by employing multiple variables, effectively allows us to mimic this situation.142 By including a host of factors that might affect FDI flows, we “control” for the possibility that they might exert as much or more influence on FDI as our variable of interest, a BIT. Only in this fashion do we simulate a real-world environment. The selection of explanatory variables is derived from a wide range of FDI literature. The literature on FDI falls into a number of camps, focusing on countries, firms, location within countries, macro economic effects—and on bilateral versus aggregate studies. The key explanatory variables that show up repeatedly, in our reading of the literature, are host country GDP, population, real exchange rate, infrastructure, human capital, and openness to trade/global integration. 141. One of the classic examples in econometrics is assessing the impact of education on income. Were we to examine the impact of years of education alone, ignoring native ability, parental occupation, quality of schools and so on, we would not get a true reading of education’s impact. 142. See Wooldridge, supra note 139, at 66.

do bit s really work?: an evaluation of bilateral investment treaties 143

Often these variables are summed (GDP) or differentiated (wages, for example) to show positive or negative correlation about country differences. Another important factor in attracting FDI is the creation of Special Economic Zones.143 Literature evaluating the impact of BITs on FDI is scant, consisting of three UN studies, the last of which is based on econometric analysis, and a recent paper by World Bank Senior Economist Mary Hallward-Driemeier,144 which is similarly based on econometrics. We review these studies before presenting our own results. 2. Prior UN studies The first UN study145 was published in 1988, when a mere 265 BITs had been concluded, and thus its value lies more in highlighting the significant transformation of the BIT process in the last fifteen years than in shedding light on BITs’ impact. It notes that only half of the Group of 77 were parties to BITs, many had not signed new treaties in more than ten years, and that while some countries had signed one or two BITs, others had concluded many more.146 “If all the members of the Group of 77 concluded [BITs] with all the members of OECD, one would arrive at a figure of 2,500,”147 says the report, with no anticipation that the number would grow to 2,200 BITs today.148 In fact, the report speculates that the 1985 development of investment insurance through the Multilateral Insurance Guarantee Agency (MIGA) might “reduce the attractiveness of bilateral investment treaties to developing countries.”149 In all, the report strikes a number of pessimistic notes about the value of BITs as a protection or promotion device, and it shares the view of a 1985 report about the inconclusiveness of quantitative results.150 It found “no apparent relationship” between the number of bilateral agreements and the volume of FDI flows, and concludes that the “reasons for the increase or fall in foreign direct investment can be explored in a meaningful way only if each case is examined separately, taking into account all relevant factors.”151 This conclusion would appear to be a clear call for an econometric study.

143. See K.C. Fung et al., Determinants of U.S. and Japanese Direct Investment in China, 30 J. Comparative Econ. 567, 573 (2002). 144. Mary Hallward-Driemeier, Do Bilateral Investment Treaties Attract FDI? Only a Bit . . . and They Could Bite (World Bank, Working Paper No. 3121, June 2003), available at http://econ.worldbank.org/files/29143_wps3121.pdf (last visited Nov. 29, 2004). 145. See generally UNCTC, supra note 74 (examining FDI flows from OECD countries to thirty-three developing countries). 146. Id. at 72, ¶¶ 331–32. 147. Id. at 72, ¶ 333. 148. See UNCTAD, supra note 2, at 89. 149. UNCTC, supra note 74, at 72, ¶ 334. 150. Id. at 11, ¶ 38. 151. Id.

144 jeswald w. salacuse and nicholas p. sullivan

The second UN report,152 which mainly catalogued BITs through 1991, expressly made no effort “to discuss or assess the merits of bilateral investment treaties as policy instruments on foreign investment.”153 The third study,154 published by UNCTAD in 1998, is an econometric study. It takes a two-stage approach—a multi-year study of FDI inflows into 133 host countries at one point in time (using 1995 FDI data), and a study of bilateral FDI flows from fourteen source countries into seventy-two host countries (using data between 1971 and 1994).155 The multi-year or “time series” approach, in theory, allows for a more precise identification of FDI determinants, as it compares specific bilateral partners over a period of years, so that anomalies from any one year are discounted and smoothed, and comparisons can be made before-and-after BIT signings. By contrast, the one-year (cross-section) study analyzes aggregate FDI flows to a given country based on the total number of BITs signed, but does not examine bilateral relationships. While sacrificing precision, this data is easier to collect and more consistent. We examine in detail both stages of the 1998 UNCTAD study. a. Bilateral FDI flows covering 23 years (1971–1994). In the time-series study covering twenty-three years, the coefficients on BITs as a determinant of FDI flows are positive, indicating that a BIT could increase FDI flows between treaty partners, holding constant other factors such as GDP or population growth. The magnitude of the effect varies depending on the specific before-and-after signing time period, with which the authors experiment in attempting to describe a lag period between signing and impact.156 However, the results are not statistically robust, showing 5%–10% significance levels; 1%–5% significance would suggest much higher confidence in the results. In addition, the data collected for a twenty-three-year period comes from a variety of sources, suggesting likely discrepancies in collection methods. Furthermore, the bulk of the data comes from the Cold War era (1970s and 1980s), when FDI flows were much lower and BITs far fewer than in the post-Soviet era of the 1990s.157 On a more definite note, the authors observe that the most consistently positive dependent variables (y) were FDI/inflows (share of host country in source country’s total FDI outflows) and FDI/outflows (share of source country in host country’s total FDI inflows), both of which are a better measure of the role of BITs than FDI alone or FDI/GDP.158 The authors conclude that “BITs may serve,

152. UNCTC, Bilateral Investment Treaties, 1959–1991, UN Doc. ST/CTC/136 (1992). 153. Id. at iv. 154. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1. 155. Id. ch. IV. 156. Id. at 110. 157. Id. at 108. 158. Id. at 109.

do bit s really work?: an evaluation of bilateral investment treaties 145

at the margin, to redirect the share of FDI from/to BIT signatories.”159 Further, BITs signed by African countries appeared to have more effect than BITs in other regions, likely because BITs are more important when host countries are less developed (though this is an untested hypothesis).160 The authors conclude that the influence of BITs on FDI is weak; however, they do say: [F]ollowing the signing of a BIT, it is more likely than not that the host country will marginally increase its share in the outward FDI of the source country; the same applies to the share of the source country in the FDI inflows of the host country. The effect, however, is usually small.161 b. 133 host countries using 1995 FDI data. In the cross-country study, which uses FDI flows in 1995 as the dependent variable, BITs were found to have a positive and statistically significant effect in three of the nine reported regressions.162 In one regression, “each BIT [signed] in 1993 can be said to be associated with an incremental 162 million U.S. dollars in FDI flows in 1995,”163 report the authors; however, they note that this is a “statistical abstraction, in the sense of a fitted regression trend line for many countries, and not a policy conclusion.”164 Overall, the prime explanatory variables for FDI flows were GDP, population, and domestic investment, consistent with other economics literature on FDI flows.165 In addition, the only BITs that showed statistically significant positive coefficients were those signed one or two years before the measured FDI flows, suggesting a delayed impact. Though not mentioned, the lag between signing and actual ratification by both countries’ legislatures is typically two or three years. In its conclusion, the study finds the influence of BITs on bilateral FDI flows to be weak, saying that “BITs appear to play a minor and secondary role in influencing FDI flows.”166 One reason for the relatively weak influence of BITs is that their signaling power may have eroded during the 1990s as investors increasingly saw them as a “normal feature of the institutional structure.”167 Further, the study notes some evidence that foreign investors encourage governments to conclude BITs with

159. Id. 160. Id. at 111. 161. Id. at 122. 162. The authors ran 192 regressions overall, testing eight different measures of FDI flows and stock, looking for the variables with the most explanatory power. See id. at 118. 163. Id. at 120. 164. Id. 165. Id. at 118–19. 166. Id. at 120. 167. Id. at 122.

146 jeswald w. salacuse and nicholas p. sullivan

host countries in which they already have FDI, a phenomenon paralleled in findings about multinational corporations lobbying for Double Taxation Treaties.168 The fact that BITs play a “minor and secondary role”169 as a determinant of FDI in the cross-country study, and have a consistently positive if somewhat marginal statistical (90%–95% significance levels) impact in the time-series study, certainly implies that BITs do play a quantifiably positive role in promoting investment. c. 2003 World Bank econometric study of 537 bilateral pairs from 1980 to 2000. The Hallward-Driemeier World Bank study170 uses an econometric time-series approach to study the impact of BITs on FDI flows between 537 bilateral country pairs (from twenty OECD countries to thirty-one developing countries) from 1980 to 2000. In many ways, this is an updated version of the UNCTAD timeseries study, with the length of the time period allowing for significant analysis of pre- and post-treaty FDI flows. The various regressions control for the size of the source country, the size of the host country, the host country’s macroeconomic stability (represented by its inflation rate), its openness to trade (represented by trade as a percentage of GDP), and the gap in average years of education between source and host.171 Two other notable factors that are considered are the transition from Soviet to market economies, and the signing of NAFTA.172 Due to the complexity of cataloging different types of BITs, the World Bank paper treats all BITs equally, noting the general point that BITs strengthen property rights;173 however, the author notes that there are significant differences among BITs, and that “it is possible that there would be more of an effect if one looked only at those treaties with the strongest investor protections,”174 an observation that will be further developed in our own study of U.S. BITs. Finally, in addition to examining the impact of individual variables on FDI flows, the paper also studies the combined impact of a BIT when assessed along with the quality of the legal system, as well as with the level of corruption. Its assumption in this respect is that, while BITs should signal protection of the

168. See Bruce A. Blonigen & Ronald B. Davies, Do Bilateral Tax Treaties Promote Foreign Direct Investment? 4–7 (Nat’l Bureau of Econ. Research, Working Paper No. 8834, Mar. 2002), available at http://dsl.nber.org/papers/w8834.pdf (last visited Nov. 30, 2004). 169. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 122. 170. Hallward-Driemeier, supra note 144. 171. Id. at 12–13. 172. Id. at 13. 173. Id. at 14. While the NAFTA chapter on investment includes language similar to that of a BIT and can thus be considered a BIT for certain purposes, NAFTA as a whole was largely a trade agreement, one that made Mexico a more attractive destination for investment as an export platform to the United States and Canada. 174. Id.

do bit s really work?: an evaluation of bilateral investment treaties 147

property rights of the foreign investor, the credibility of the signal will be affected by the degree of corruption and strength of the legal system.175 The findings are more or less in line with most previous studies on BITs. “The larger the source country and the larger the host country, the larger the FDI flows. Flows are also higher to richer host countries. Macroeconomic instability discourages FDI.”176 Not surprisingly, the impact of NAFTA is quite significant on FDI flows—but “it is hard to disentangle which effect [(trade or investment)] really dominates.”177 On the key issue, the coefficient on the BIT treaty is negative and not significant in most regressions, with some minor exceptions. For example, “only in year five after the ratification is there a positive (and extremely weak) association.”178 In another case, looking at FDI going to a particular host country as a share of the total FDI a source country sends, “one gets the one significant positive result that a BIT could increase FDI.”179 Finally, the interaction of a BIT with institutional capacity (as measured by the World Bank’s Kaufmann, Kraay and Zoido-Lobatón [KKZ] indicators for rule of law, corruption, government effectiveness, and regulatory quality)180 has either no effect, or a positive interaction. That is, a country with a BIT and a stronger institutional capacity appears to attract slightly more FDI than a country with just a BIT. Thus, the author concludes, BITs complement rather than substitute for strong domestic institutions; BITs are more effective in settings of higher institutional quality.181 This conclusion, of course, undermines a central rationale for less developed countries that enter into investment agreements hoping to bypass the need to strengthen property rights and enforcement mechanisms under domestic law. 3. The authors’ studies on the impact of U.S. BITs As with the 1998 UNCTAD study,182 the authors employ two approaches to study the effect of U.S. BITs on FDI flows. The first analyzes aggregate FDI inflows to more than 100 developing countries in a given year, with separate regressions run for 1998, 1999, and 2000. The second examines a dataset of U.S. FDI flows to thirty-one countries repeated over a ten-year period. Overall, the results indicate that U.S. BITs are more likely to induce FDI inflows than those concluded by other OECD183

175. Id. at 20–21. 176. Id. at 18. 177. Id. 178. Id. at 19. 179. Id. at 20. 180. World Bank, Worldwide Governance Research Indicators Dataset (2002), at http:// www.worldbank.org/wbi/governance/data.html#dataset2001 (last visited Nov. 18, 2004). 181. Hallward-Driemeier, supra note 144, at 21. 182. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 103. 183. The OECD consists of thirty member countries that have developed economies.

148 jeswald w. salacuse and nicholas p. sullivan

countries, and that a host country with a U.S. BIT is more likely to increase its overall FDI (from all OECD countries) than a country without a U.S. BIT, holding other factors equal. To our knowledge, this is the first study to examine the impact of BITs negotiated by one specific country (i.e., the United States), rather than an aggregate collection of BITs. Furthermore, as noted above,184 the U.S. BIT, compared with those negotiated by other countries, offers the strongest investor protections; thus, it is plausible to hypothesize that the U.S. BIT might have a stronger association with FDI flows than less stringent BITs. a. Effect of U.S. BITs on aggregate FDI flows to developing countries. In this study on aggregate FDI flows (y variable), the explanatory (x) variables of interest are the presence of a U.S. BIT, the total number of BITs signed with all other OECD countries (excluding the United States), and the number of BITs signed with all other developing countries. Other explanatory variables, included because of their known impact on FDI, are host country GDP, GDP per capita, inflation, real effective exchange rate, population, and rule of law (see Data Appendix A for a list of explanatory variables and sources). The econometric model185 measures the percentage change in FDI inflows, which is more meaningful than absolute FDI inflows because it allows for more relevant comparisons between large- and small-market host countries. Data for GDP, exports and inflation are lagged one year, i.e., data for these explanatory variables is for the year previous to the measured FDI flows. The lag is consistent with other econometric FDI studies, on the theory that investors were acting on known information from the year before.186 In addition to studying the relationship between a U.S. BIT and overall FDI flows, the dataset provides an opportunity to examine the dynamic interaction of BITs—whether a U.S. BIT has more or less effect than another OECD country’s BIT, whether a greater number of OECD BITs increases flows at the margin and by how much, and the difference between a BIT with a developed country and one with a developing country. The UNCTAD study did not address any of

184. Juillard, supra note 17, at 211. See also supra note 102 and accompanying text. 185. The estimation used is an ordinary least squares (OLS) multivariate regression specified as follows: lnfdi99 = B0 + B1lngdp98 + B2exports98 + B3pop + B4rlaw + B5infl98 + B6oecdbits +B7usbit + B8nonoecd + ui The dependent variable is lnflows98, lnflows99, or lnflows00, the natural log of total FDI flowing into a given country in 1998, 1999, or 2000. The explanatory variables of interest are usbit, oecdbits, and nonoecd (BITs)—controlling for the effects of lngdp (the natural log of GDP, lagged one year), population, rule of law, exports per GDP (lagged one year), and infl (inflation rate, lagged one year). 186. See Hallward-Driemeier, supra note 144; see also UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1.

do bit s really work?: an evaluation of bilateral investment treaties 149

these issues. Instead, it treated all BITs as equal, as the movement toward “SouthSouth” BITs was just gathering momentum at the time of its 1995 study.187 The regression results indicate that the presence of a U.S. BIT has a large, positive, and significant association with a country’s overall FDI inflows. Coefficients on the U.S. BIT variable range from. 77 to. 85, which translates into increased FDI to a given country in a given year by 77% to 85% (at the 1%–5% significance levels). Because the FDI inflows measured are aggregate and not bilateral with the United States, it is not clear from these regressions whether the correlation is due to increased U.S. flows, or because a U.S. BIT induces flows from other countries. But it is clear that a U.S. BIT is more highly correlated with FDI inflows than other BITs. In each year, the addition of a new OECD BIT has a weak positive effect, and the addition of another BIT with a developing country has a weak negative effect. In both cases, the effects lack statistical significance (see Table 2). As expected, a country’s GDP (lagged one year) was overall the main determinant of FDI flows, with a 1% increase in GDP correlating with a 90%–97% increase in FDI to that country. The “Rule of Law” and “Exports” variables exhibited positive significance in a number of regressions, while inflation showed a weak negative influence, which was statistically significant in just one regression. The presence of a U.S. BIT appears to exert a huge impact on FDI inflows (correlated with a 112% to 157% increase in the three different years) when a country’s overall OECD BITs are below the mean number of OECD BITs (7.3 BITs) (see Table 2). Conversely, a U.S. BIT seems to show a very weak and statistically insignificant effect when OECD BITs are above the mean. The results could describe a certain crowding out of U.S. FDI when other OECD countries have relationships with a developing country—or indicate that flows to these countries are already high and any extra flows from the United States or other OECD countries have marginal effect. The results are in line with previous UNCTAD conclusions on Africa (and, to a lesser extent, transition countries in Central and Eastern Europe), showing that the less developed the country (thus with fewer overall BITs), the more apparent effect any one BIT will have. Our findings, however, conflict with the more recent finding that less developed countries cannot expect BITs to substitute for institutional capacity.188 187. Whereas most BITs have been concluded between a developed and a developing nation, the transformation of some successful developing countries into capital exporters has led to a huge increase in the number of BITs between two developing nations, between developing nations and countries in Eastern Europe, and between Central and Eastern European countries. These BITs have increased from 63 in 1989 to 833 in 1999. See Press Release, UNCTAD, Bilateral Investment Treaties Quintupled During the 1990s, New UNCTAD Publication Releases the Latest Data on the Universe of BITs, UN Doc. TAD/ INF/PR/077 (Dec. 15, 2000). 188. See Hallward-Driemeier, supra note 144, at 21.

150 jeswald w. salacuse and nicholas p. sullivan

When OECD BITs are dropped as an explanatory variable (not shown in Tables 1–5), the apparent impact and significance of a U.S. BIT is more extreme. When the U.S. BIT is dropped, the apparent impact of OECD BITs is large and statistically significant. Taken together, these results suggest not only that concluding an additional BIT with an OECD country is likely to increase a given country’s FDI flows, ceteris paribus (in keeping with the UNCTAD 1998 study that found a redirection of flows at the margin),189 but also that if the additional BIT is with the United States, the increase in FDI flows is likely to be substantially larger. Consequently, in the case of U.S. BITs, and to a lesser extent the BITs of other OECD countries, BITs arguably have a positive impact on promoting investment to the signatory country. Thus, it would appear that a BIT could achieve the goal of investment promotion to varying degrees. Furthermore, comparing our study with the 1998 UNCTAD study,190 one may make the argument that if a developing country truly wishes to promote foreign investment, it is better to sign a BIT with high protection standards, like those advocated by the United States, than one with weaker standards as evidenced by certain other OECD countries. The basis of this argument is that a BIT with stronger standards creates a less risky investment climate than a BIT with weaker standards of protection and that, all other things being equal, foreign investors will prefer a less risky investment climate. Signing a U.S. BIT may also tend to lead to increased FDI flows from other OECD countries because OECD investors, by virtue of the most-favored-nation clause in OECD treaties, gain the protection of the high protective standards in U.S. BITs. On the other hand, since the United States is the leading foreign direct investor in the world and is likely to sign BITs with countries where U.S. multinational enterprises (MNEs) are engaged in or lobbying for business, other OECD MNEs may match U.S. investors just to remain competitive in certain regions. That is, OECD MNEs may not be influenced as much by the presence of a U.S. BIT as they are by the presence of U.S. MNEs with which they compete. Finally, the signing of a U.S. BIT may signify a commitment to economic liberalization in line with the Washington Consensus191 principles of the early

189. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 109; supra text accompanying note 159. 190. UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1. 191. The term “Washington Consensus” was coined in a 1990 paper by John Williamson, which outlined ten policy reforms that Latin America should undertake: fiscal discipline, redirection of public expenditures toward high-yield areas such as health and education, tax reform, interest rate liberalization, competitive exchange rate, trade liberalization, liberalization of inflows of FDI, privatization, deregulation, and secure property rights. John Williamson, What Washington Means by Policy Reform, in Latin American Adjustment: How Much Has Happened? 5, 5–20 (John Williamson ed., 1990).

do bit s really work?: an evaluation of bilateral investment treaties 151

1990s, and which U.S. negotiators demand. President Clinton, in a letter asking the Senate to ratify a Free Trade Agreement with Jordan in 2001, wrote: This Agreement is a vote of confidence in Jordan’s economic reform program, which should serve as a source of growth and opportunity for Jordanians in the coming years. The United States-Jordan Free Trade Agreement achieves the highest possible commitments from Jordan on behalf of U.S. business on key trade issues, providing significant and extensive liberalization across a wide spectrum of trade issues.192 Similarly, U.S. Trade Representative Robert Zoellick, in announcing the beginning of negotiations on a United States-Pakistan BIT, said the BIT “can play an important role in strengthening Pakistan’s economy, so as to create new opportunities for exporters and investors in both economies.”193 To the extent that a U.S. BIT helps revamp the overall economic climate, or signals a move in that direction, other OECD investors may be more inclined to invest. b. Ten-year study on impact of U.S. BITs on bilateral U.S. FDI outflows. Our time-series study uses a dataset of U.S. FDI flows to thirty-one countries194 over a ten-year period (1991–2000). Data is available from most countries for all years, with the exception of transition countries in the early 1990s (see Table 4). Of the thirty-one countries, eleven, primarily from Latin America and Central Europe,195 have concluded BITs with the United States. The time-series (or panel data) model pools results from each of the thirty-one country groups to gain close to 300 observations and calculates a coefficient showing the degree to which a U.S. BIT correlates with FDI inflows. The decision to focus on U.S. FDI, rather than on a “basket” of OECD flows, was based on the robustness of U.S. data. In addition, given the switch to the euro during the covered time-series, the focus on U.S. FDI also had the simplicity of

192. Letter from the White House Office of the Press Secretary to the Congress of the United States (Jan. 6, 2001), available at http://clinton6.nara.gov/2001/01/2001-01-06letter-from-the-president-on-fta.html (last visited Nov. 19, 2004). 193. Press Release, Office of the United States Trade Representative, Executive Office of the President, Pakistan, United States to Negotiate Bilateral Investment Treaty (Sept. 28, 2004), available at http://usinfo.state.gov/ei/Archive/2004/Sep/29-36999. html (last visited Nov. 19, 2004). 194. Countries included in the dataset are those included in both the UNCTAD list of developing countries and the primary FDI data source for this study. OECD, International Direct Investment Statistics Yearbook 1980–2000 (2001 ed., 2002). 195. These countries are Argentina, Bulgaria, Czech Republic, Egypt, Morocco, Panama, Poland, Romania, Slovakia, Turkey, and Ukraine. See Fact Sheet, Bureau of Economic and Business Affairs, U.S. Department of State, U.S. Bilateral Investment Treaty Program (Sept. 15, 2004), available at http://www.state.gov/e/eb/rls/fs/22422. htm.

152 jeswald w. salacuse and nicholas p. sullivan

converting one bilateral real exchange rate for each of the thirty-one countries in ten time periods. Furthermore, we felt that focusing on the primary source country, and one whose treaties enforce strong investor protections, might yield more conclusive results than previous time-series studies, which examine a “basket” of source countries (see Table 3 for a list of all U.S. BITs). Because we are looking at a set of bilateral relationships, we can use a so-called “fixed-effects” technique, which assumes that many of the explanatory variables remain constant over a ten-year period. Thus, rather than controlling for all the explanatory variables that might affect FDI, such as telecom infrastructure, access to deep-water ports, or labor force education, these “fixed effects”196 (also referred to as “unobserved effects”) are not included in the model. As they don’t change significantly from year to year, they are by definition held constant. This time-series model allows us to observe whether U.S. FDI flows to a given country change after a BIT is signed or ratified, holding all other factors about the bilateral pair constant. The constant factors (“fixed effects”) in each bilateral pairing that drop out when “differencing” one year from the next (and thus need not be included in the regression) include population, history of trade and political engagement, linguistic ties, rule of law, cultural ties, distance between major ports, quality of the labor pool, savings rate, etc. In the articulated econometric model,197 we do include variables to measure changes in several macroeconomic indicators, such as GDP, GDP growth rate, ratio of private investment, and the bilateral real exchange rate, all of which change from year to year (see Data Appendix B for all variables and sources). Our finding is that a U.S. BIT is correlated with a major increase in (at a 1% significance level) U.S. FDI outflows to a given country, ceteris paribus, compared to U.S. flows to a country without a U.S. BIT (see Table 5). In fact, the study suggests that a U.S. BIT is correlated with an extra $1 billion (approximately)

196. Wooldridge, supra note 139, at 420. 197. The estimation used is a fixed effect model specified as follows: usfdiit = B0 + B1lngdpit-1 + B2gdppercapit-1 + B3oecdbits + B4lnRERit-1 + B5inflowsit-1 + B6Usbitit + eit + ui The dependent variable is usfdi (U.S. FDI flows to country i at time t). Ordinarily, the measure of FDI flows is logged to identify their elasticity (percentage change), but in this case there are numerous years when flows to a given country are negative, and it’s important to capture that data rather than treating it as 0 or missing data. The key economic explanatory variables include lngdp1 (the natural log of total GDP (lagged one year), gdppercap (lagged), lnrer (the natural log of bilateral real exchange rate, lagged), inflows (total FDI inflows to a given country in time t), oecdbits (the number of OECD BITs), usbit (the presence of a U.S. BIT), bitsign (the year a U.S. BIT was signed), bitforce (the year a U.S. BIT was ratified), and treatage (the number of years since the BIT was signed, up to a maximum of 10).

do bit s really work?: an evaluation of bilateral investment treaties 153

in increased FDI per year. As does the UNCTAD study,198 we would also caution that this is a statistical abstraction, factoring in data from many countries grouped together, and not necessarily a policy conclusion about any single treaty partner. In addition, we measure solely the level of FDI, and not the increase or decrease in the share of FDI that flows from the United States to the host. The timing variables (the date of signing, the date of ratification, and the number of years since the signing) have no coefficients that are statistically significant; however, the correlations indicate that ratification has a positive effect on FDI flows (and much more effect than the mere signing of a BIT), as does the passage of each year after the treaty is signed. Again, this is in keeping with both the UNCTAD study, which tentatively concludes that “the response lag after the signing of a BIT may be as little as zero but is more likely to be two years,”199 and the Hallward-Driemeier study, which finds a positive (but very minor) correlation in FDI in year five after ratification.200 GDP and GDP per capita both have a positive and statistically significant correlation, as expected. Also as expected, an appreciation of the bilateral real exchange rate in the host country (resulting in greater expense for foreigners to do business) has a large and significant negative correlation. Total FDI inflows have virtually no effect in either direction on U.S. flows, so there is no indication that FDI from other OECD countries crowds out U.S. FDI, or that FDI from other OECD countries attracts U.S. FDI. One variable that we did not control for is “policy effects,” which includes NAFTA and its impact on flows to Mexico, and WTO accession and its impact on flows to China. In the Hallward-Driemeier study, NAFTA’s impact on OECD flows was large and significant, although the policy change in the transition countries of Eastern Europe was generally negative or negligible.201 Sorting out the impact of various trade and investment pacts on FDI may be fodder for a future study. Due to data constraints, not every country with a U.S. BIT is included in the study. Although Chile, Colombia, Costa Rica, and Venezuela have BITs with the United States, the only Latin American countries included are Argentina, Mexico, and Panama. Moreover, while the United States has a number of treaties with the transition countries of Eastern and Central Europe, FDI was scarce or the data nonexistent in the early 1990s.202 There is no published OECD data on FDI inflows for some small countries that have signed BITs with the United States,

198. See UNCTAD, Bilateral Investment Treaties in the Mid-1990s, supra note 1, at 120; supra text accompanying note 164. 199. Id. at 111. 200. See Hallward-Driemeier, supra note 144, at 33. 201. Id. at 31–36. 202. The impact of a U.S. BIT on flows to Eastern and Central Europe is negative, but statistically insignificant; perhaps with more data, this uncertainty may be resolved.

154 jeswald w. salacuse and nicholas p. sullivan

such as Mongolia, Moldova, Bangladesh, Cameroon, and Mozambique; thus, these countries are not included in our dataset. Finally, the United States has no BIT with China, which has attracted an increasing level of FDI from the United States in recent years. In a future study, we may expand the analysis to compare several other sets of bilateral partners that include key source countries such as Japan (which has largely abstained from the BIT movement), the United Kingdom, Germany, Belgium, Luxembourg, and others. This approach may begin to refine the analysis of the impact of different OECD BITs on FDI flows. To better test the dynamics of other OECD BITs on U.S. BITs and FDI flows, it would also make sense to control for OECD BITs in each year, as we did in the cross-country study; however, data on signing dates for individual BITs beyond 1996 has not been published by UNCTAD or the World Bank,203 making application of such data more difficult. It is hard to make broad generalizations regarding U.S. BIT/FDI dynamics with certainty, as such a huge proportion of FDI to developing countries goes to China, Brazil, and Mexico, and of these only Mexico has a BIT (NAFTA) with the United States. Nonetheless, the two econometric studies are large enough (with about 100 and 300 observations, respectively), and tested with enough variables to assure that there were no key omissions. Nor is any variable constant or a perfect linear combination of the others (“no perfect collinearity”), a required assumption for a clean econometric model. We can be confident in concluding that: a. A U.S. BIT is more likely than not to exert a strong and positive role in promoting U.S. investment. b. A U.S. BIT is more likely than not to exert a strong and positive role in promoting overall investment. c. A U.S. BIT is likely to exert more of an impact than other OECD BITs in promoting overall investment. After reviewing both the literature, which makes note of the potential impact of BITs with strong investor protections, and our own econometric study on the promotional effects of a U.S. BIT, we find strong evidence that BITs have, to a significant extent, attained their stated goal of promoting investment.

203. ICSID Web site, supra note 1 (listing individual parties to BITs through 1996). The UNCTAD database that the authors relied on for their econometric analysis did not list individual BIT signatories, but grouped signatories as developed or developing countries, and by region. However, the most recent UNCTAD database does list individual BIT signatories. See UNCTAD, Country-Specific Lists of BITs, available at http://www.unctad. org/Templates/Page.asp?intItemID=2344&lang=1 (last visited Nov. 28, 2004).

do bit s really work?: an evaluation of bilateral investment treaties 155

conclusion: a grand bargain realized This chapter has sought to answer the basic question of whether BITs really work. Has the expanding network of 2,200 BITs concluded over the last five decades produced the effects intended by the countries that have signed them? This chapter has tried to answer that question by evaluating the impact of BITs in relation to their intended goals: foreign investment protection, investment and market liberalization, and investment protection. In this regard, it concludes that while BITs, in and of themselves, may have not directly and substantially liberalized FDI, there is strong evidence to show that they both protect and promote FDI in developing countries and the United States. BITs have a particularly strong effect on encouraging FDI in developing countries. In short, the grand bargain between developing and developed countries that underlies BITs, the bargain of investment promotion in return for investment protection, seems to have been achieved, although the effect of the bargain is only realized slowly after the BIT is signed. The fact that many nations have realized that bargain with respect to foreign investment has important implications for the future. First, it may give added impetus to future negotiations of still more BITs in the years ahead. With evidence that BITs really do protect and promote investment, both developed and developing countries may display an increased willingness to negotiate new BITs with new partners. Developing country governments that may have been reluctant to sign BITs because of concerns that BITs would prove costly and bring them little additional investment may now see evidence of increased capital flows as reason to justify treaty participation, particularly if other countries with whom they compete for foreign capital have signed BITs and obtained substantial foreign investment. Although BIT critics204 in developing countries point to the increased number of arbitration awards against developing countries as justification for their opposition, evidence of substantially increased investment flows severely weakens their position. Second, evidence that the bargain implicit in the BITs has advanced the interests of both developed and developing countries may serve to give renewed impetus to multilateral efforts to negotiate a global treaty on international investment. Such efforts were thwarted at the end of the 1990s with the failure of the OECD to negotiate a multilateral investment agreement205 but may now be renewed in the current Doha Round of Negotiations of the WTO.206

204. For an example of criticism of current arbitration procedures, see von Moltke & Mann, supra note 18, at 30–31. 205. See Kelley, supra note 63, at 484. 206. World Trade Organization, Ministerial Declaration of 14 November 2001, WT/ MIN(01)/DEC/1, 41 I.L.M. 746, 749 (2002).

156 jeswald w. salacuse and nicholas p. sullivan

Finally, and admittedly most problematically, in light of the proliferation of BITs and the perception that they really do work, one may speculate on the role of BITs as a source of international law applicable beyond the two parties to the BIT itself. The potential for this wider application is crucially important because it could portend the development—indeed perhaps the current existence—of a multilateral international investment regime built on sources of law other than treaties. BITs could provide such a source if they are seen in one of two ways:207 (1) as influences on and evidence of customary international law, or (2) as embodiments of general principles of law common to the world’s legal systems.208 A. BITs as Customary International Law Virtually since the beginning of the BIT movement, scholars have debated the extent to which BITs constitute or form customary international law with respect to foreign investment. One argument is that BITs “establish and accept and thus enlarge the force of traditional conceptions” of the law of state responsibility for foreign investment.209 Others have countered that, despite their prevalence, BITs are lex specialis, and have effect only between the parties to the BIT.210 According to this view, BIT provisions are not sufficiently uniform to establish custom accepted by the international community. In order to qualify as customary international law, BITs must have “result[ed] from a general and consistent practice of states followed by them from a sense of legal obligation.”211 The first requirement, “state practice,” is likely not difficult to prove, given that the common structure and language of certain BITs provisions, particularly those of more recent vintage, as well as their widespread adoption, arguably reflect a practice “both extensive and virtually uniform.”212 With more than 2,200 BITs with similar language and obligations in operation around the world, state practice seems readily demonstrable. Although only a relatively short period of time has passed since BITs were widely adopted, “the passage of only

207. In both cases, practices would need to be looked at on an individual basis; that is, one would have to go through the following analysis for each area in question (e.g. expropriation), and could not simply assume that the content of all BITs is necessarily law if any of it is. 208. See generally Statute of the International Court of Justice, supra note 6 (identifying the major sources of international law). 209. F. A. Mann, British Treaties for the Promotion and Protection of Investments, 52 Brit. Y.B. Int’l L. 241, 249 (1981). 210. See Bernard Kishoiyian, The Utility of Bilateral Investment Treaties in the Formulation of Customary International Law, 14 Nw. J. Int’l L. & Bus. 327, 329 (1994). See also Sornarajah, supra note 1, at 276. 211. Restatement (Third) of the Foreign Relations Law of the United States, supra note 9, § 102. 212. North Sea Continental Shelf (F.R.G. v. Den.), 1969 I.C.J. 3, 43 (Feb. 20).

do bit s really work?: an evaluation of bilateral investment treaties 157

a short period of time is not necessarily, or of itself, a bar to the formation of a new rule of customary international law.”213 The second requirement of opinion juris sive necessitatis, that the state practice “should have occurred in such a way as to show a general recognition that a rule of legal obligation is involved,”214 could prove to be more problematic to satisfy. Are the principles embodied in the BITs lex specialis, constituting specially agreed upon rules between treaty partners? Or have the principles embodied in more than 2,200 BITs become sufficiently generalized that countries respect them because they recognize that a legal obligation is involved? Certainly as the number of countries involved in BITs increases and the total number of BITs grows, certain BIT provisions, if sufficiently common, approach the status of custom and seem less and less like mere lex specialis. Moreover, if the BITs have proved their effectiveness, as this chapter has argued, then arguably, countries desiring to have an effective legal system to attract foreign investment would follow certain basic BIT provisions as the goal of having an effective foreign investment legal regime encourages them to do so. Beyond this purely theoretical argument is the much more powerful persuasion of actual practice. Two recent arbitration awards have taken the view that BITs do indeed constitute or at least contribute to international custom. In Pope & Talbot, Inc. v. Canada,215 the tribunal interpreted NAFTA Article 1105(1)216 to be similar to many BITs in providing that: “Each Party shall accord to investments of investors of another Party treatment in accordance with international law, including fair and equitable treatment and full protection and security.”217 In making the interpretation, the tribunal took account of the evolution of investor rights caused by the BIT movement and found that the current content of international custom reflects the provisions of the many BITs concluded by the nations of the world, stating that: [A]ll parties agree, that the language of Article 1105 grew out of the provisions of bilateral commercial treaties negotiated by the United States and other industrialized countries. As Canada points out, these treaties are a ‘principal source’ of the general obligations of states with respect to their treatment of foreign investment.218

213. Id. 214. Id. 215. Pope & Talbot, Inc. v. Canada, Award on the Merits, Phase Two (NAFTA Ch. 11 Arb. Trib. Apr. 10, 2001), available at http://www.dfait-maeci.gc.ca/tna-nac/documents/ Award_Merits-e.pdf (last visited Nov. 22, 2004). 216. NAFTA, supra note 45, at 639. 217. The tribunal cites the Model Bilateral Investment Treaty of 1987, reprinted in Vandevelde, supra note 1, as particularly similar, which is important because it influenced a number of countries’ BITs. See Pope & Talbot, Award on the Merits, Phase Two, at ¶ 111. 218. Id. at ¶ 110 (emphasis added) (citation omitted).

158 jeswald w. salacuse and nicholas p. sullivan

In Mondev Int’l Ltd. v. United States,219 all three NAFTA member countries made submissions challenging Pope & Talbot, Inc., particularly because the Pope & Talbot, Inc. tribunal did not consider the necessary element of opinio juris in establishing the asserted custom.220 A three-member tribunal, which included a former president of the International Court of Justice, found the issue to be “entirely legitimate” in making its own interpretation of NAFTA’s Article 1105.221 It also had to take into account a July 31, 2001, interpretation by the NAFTA Free Trade Commission222 that found that Article 1105(1) “prescribes the customary international law minimum standard of treatment of aliens as the minimum standard of treatment to be afforded investments of investors of another Party” and that “the concepts of ‘fair and equitable treatment’ and ‘full protection and security’ do not require treatment in addition to or beyond that which is required by the customary international law minimum standard of treatment of aliens.”223 Canada had suggested that the meaning of those provisions in customary international law should be interpreted by reference to the standard set down in the decision of the Mexican Claims Commission in the 1926 Neer case.224 After lengthy discussion, the tribunal rejected that view, holding that the Free Trade Commission’s interpretation of Article 1105(1) “incorporate[s] current international law, whose content is shaped by the conclusion of more than two thousand bilateral investment treaties.”225 Thus, the process of creating an international law of investment has seemingly evolved from a situation where the absence of appropriate custom prompted the creation of more than 2,200 BITs, which in turn has led to the creation of custom. B. BITs as General Principles of Law The notion that the principles embodied in BITs could represent “general principles of law”226 and thus constitute a source of international law has not received

219. Mondev Int’l Ltd. v. United States, Case No. ARB(AF)/99/2 (NAFTA Ch. 11 Arb. Trib. Oct. 11, 2002), 42 I.L.M. 85 (2003). The members of the tribunal were Sir Ninian Joseph (President), Professor James Crawford, and Judge Stephen Schwebel. 220. See Pope & Talbot, Inc. v. Canada, Award on the Merits, Phase Two, at ¶ 110. 221. Id. at ¶ 111. 222. Under NAFTA article 1131, interpretations by the Free Trade Commission of NAFTA provisions are binding on investment arbitration tribunals. NAFTA, supra note 45, at 645. 223. NAFTA Free Trade Commission, NAFTA Commission Notes of Interpretation of Certain Chapter 11 Provisions (2001), available at http://www.dfait-maeci.gc.ca/tna-nac/ NAFTA-Interpr-en.asp (last modified May 17, 2002) (last visited Nov. 22, 2004). 224. Counter-Memorial of Canada, Pope & Talbot, Inc. v. Canada, Phase Two, at ¶¶ 256 – 61 (NAFTA Ch. 11 Arb. Trib. 2000), available at http://www.dfait-maeci.gc.ca/tnanac/documents/b-2.pdf (last visited Nov. 6, 2004). 225. See Mondev Int’l Ltd. v. United States, Case No. ARB(AF)/99/2 (NAFTA Ch. 11 Arb. Trib. Oct. 11, 2002), 42 I.L.M. 85 (2003), at ¶ 125. 226. See Mann, supra note 209, at 249.

do bit s really work?: an evaluation of bilateral investment treaties 159

extensive consideration by scholars; however, as BITs proliferate, more and more countries incorporate BITs into their domestic legal systems. Thus, there is scope for arguing that BITs manifest certain concepts on the treatment of investors and investments that represent general principles of law. The argument is strengthened to the extent that individual countries have adopted foreign investment codes and laws that embody and amplify the rights accorded to investors in the BITs that host countries have signed.227 In a more recent formulation concerning general principles of law as a source of international law other than the Statute of the International Court of Justice, the Restatement (Third) of the Foreign Relations Law of the United States provides that “a rule of international law is one that has been accepted as such by the international community of states . . . by derivation from general principles of law common to the major legal systems of the world.”228 In search for those general principles of law with respect to investment, it would seem that a court or arbitral tribunal should be able to take into account the common principles found in BITs as well as domestic legislation influenced by the BITs individual countries have signed. Given the commonalities among BITs, this final approach to developing an international law of investment is a tantalizing, unexplored possibility for the creation of a more comprehensive international investment law regime. Whether or not BIT provisions will find their way into customary international law or general principles of law is not yet clear. What is clear is that BITs, despite early misgivings, are here to stay for the foreseeable future, as they have become a permanent part of the international system. Evidence that BITs are attaining their goals, that they really do work, will only add impetus to the movement to create the international investment law that the International Court of Justice could not discern in 1970. As a result, government officials, international executives, lawyers, and financiers will increasingly have to take them into account in planning, negotiating, undertaking, and managing international investment transactions.

227. See Jeswald W. Salacuse, Direct Foreign Investment and the Law in Developing Countries, 15 ICSID Rev.-Foreign Inv. L.J. 382, 382–400 (2000). 228. Restatement (Third) of the Foreign Relations Law of the United States, supra note 9, § 102.

160 jeswald w. salacuse and nicholas p. sullivan

data appendix a Effect of U.S. BITs on Aggregate FDI Flows to Developing Countries (results in Tables 1 and 2) oecd bits (and nonoecd) numeric variables indicating how many BITs a given country has signed with OECD countries other than the United States. Data are taken from UNCTAD and World Bank databases, which list specific treaties from 1959 to 1996. Post-1996, BITs are collected in aggregate form, divided into United States, Developed, Developing, and Central/Eastern European countries.

exports of goods and services (% of gdp) Value of all goods and other market services provided to the rest of the world, as a percent of GDP.229

net fdi inflows, (bop, current u.s.$) Represents the net inflows of investment to acquire a lasting management interest (10 percent or more of voting stock) in an enterprise operating in an economy other than that of the investor, and is defined as the sum of equity capital, reinvestment of earnings, other long-term capital, and short-term capital, as shown in the balance of payments. Data are in current U.S. dollars.230

gdp Data are in current U.S. dollars (for any given year). Dollar figures for GDP are converted from domestic currencies using single year official exchange rates. For a few countries where the official exchange rate does not reflect the rate effectively applied to actual foreign exchange transactions, an alternative conversion factor is used.231 229. World Bank, World Development Indicators (WDI) Data Query, available at http:// www.worldbank.org/data/dataquery.html (last visited Nov. 29, 2004) (Annual data reported in this site are derived, either directly or indirectly, from official statistical systems organized and financed by national governments). 230. Id. 231. Id.

do bit s really work?: an evaluation of bilateral investment treaties 161

gdp per capita, ppp (current international $) Gross domestic product converted to international dollars using purchasing power parity rates, where an international dollar has the same purchasing power over GDP as the U.S. dollar has in the United States. Data are in current international dollars (for each year in question).232

inflation, consumer prices (annual percent) Measured by the consumer price index, which reflects the annual percentage change in the cost to the average consumer of acquiring a fixed basket of goods and services that may be set or changed at specified intervals, such as yearly.233

real effective exchange rate index (1995 = 100) The nominal effective exchange rate (a measure of the value of a currency against a weighted average of several foreign currencies) divided by a price deflator or index of costs.234

population Statistics from 1998 are used for years 1998, 1999, 2000.235

rule of law Measure from –2.5 to 2.5 (worst to best) of a country’s ability to enforce laws and regulations without bribery and corruption, and of the independence of the judiciary.236

232. World Bank Group Online Media Briefing Center, World Development Indicators (WDI) Data Query, at http://media.worldbank.org/secure/data/qquery.php (last visited Nov. 29, 2004). 233. Id. 234. Id. 235. Id. 236. Daniel Kaufmann et al., Aggregating Governance Indicators 35 (World Bank, Working Paper No. 2195, Oct. 1999), available at http://econ.worldbank.org/docs/918. pdf (last visited Nov. 14, 2004).

162 jeswald w. salacuse and nicholas p. sullivan

data appendix b Ten-Year Study on Impact of U.S. BITs on Bilateral U.S. FDI Outflows (results in Table 5) gdp Data are in current U.S. dollars for each of 10 years (1991–2000). Dollar figures for GDP are converted from domestic currencies using single year official exchange rates. For a few countries where the official exchange rate does not reflect the rate effectively applied to actual foreign exchange transactions, an alternative conversion factor is used.237

gdp per capita, ppp (current international $) Gross domestic product converted to international dollars using purchasing power parity rates, where an international dollar has the same purchasing power over GDP as the U.S. dollar has in the United States. Data are in current international dollars for each year, 1991–2000.238

bilateral real exchange rate (lnrer) Calculated using the Ep∗/p ratio, where E is the United States equivalent to a given foreign currency (i.e., .05 U.S. dollars per peso), p∗ is the foreign producer price index, and p is the U.S. producer price index for a given year. In cases where the producer price index was not available, the best available substitute was used (wholesale price index, industrial price index, or domestic and import goods index). These indices are designed to monitor changes in prices of items at the first important commercial transaction, covering agricultural and industrial sectors (but not services). IMF International Statistics Yearbook239 (scaled so 1995 = 100). An increase in the real exchange rate is equivalent to a depreciation of the dollar, making it more expensive to do business in a foreign country. Transition countries, as well as China, with no reliable data until the early 1990s, presented a problem, as both nominal exchange rates and producer and wholesaler prices 237. World Bank, supra note 229. 238. World Bank Group Online Media Briefing Center, supra note 232. 239. International Monetary Fund, International Financial Statistics Yearbook 2000 (2000).

do bit s really work?: an evaluation of bilateral investment treaties 163

are often presented as a percentage increase over the previous year, with no base year. Thus, we set values in year 1995 to 100 as the base, mirroring the International Monetary Fund (IMF) model, and created an index showing the percentage increase or decrease from that base.

bit s The BIT variables used are usbit (U.S. BIT), bitsign (date of signature by U.S.), bitforce (date of ratification by U.S. Congress), and treaty age (age of BIT from date of signature). Data come from the U.S. Department of State.240 Mexico is listed as having signed a BIT in 1994, the year the NAFTA agreement was ratified, as Chapter 11 of NAFTA is modeled closely on a typical BIT.

u.s. fdi Indicates outflows from the United States into a given country. Data come from OECD International Direct Investment Statistics Yearbook 1980–2000,241 which includes comparable data from all OECD countries divided into a select group of host countries. Data from the U.S. Bureau of Economic Analysis are considerably different (and not used), but the assumption is that OECD data are somewhat more reliable as they cross-check FDI inflows and outflows to achieve consistency across countries.

countries Countries both on the UNCTAD World Investment Report 2002242 of developing countries and in the OECD International Direct Investment Statistics Yearbook 1980–2000.243 These countries include: Algeria, Argentina, Brazil, Bulgaria, Chile, China, Colombia, Costa Rica, Czech Republic, Egypt, Greece, Hong Kong, Hungary, India, Indonesia, Israel, Korea, Malaysia, Mexico, Morocco, Panama, Philippines, Poland, Romania, Russia, Singapore, Slovakia, South Africa, Thailand, Turkey, Ukraine, Venezuela. (Hong Kong and Singapore are considered developing countries by UNCTAD, although as international money centers with relatively high per capita incomes, they are clear outliers.) 240. See U.S. Department of State, supra note 195. 241. See OECD, supra note 194. 242. See UNCTAD, supra note 49. 243. See OECD, supra note 194.

164 jeswald w. salacuse and nicholas p. sullivan

Countries with which United States has signed (but not necessarily concluded) BITs that are not included are: Albania, Armenia, Azerbaijan, Bahrain, Bangladesh, Belarus, Bolivia, Cameroon, Congo (D.R)., Congo (Republic), Croatia, Ecuador, El Salvador, Estonia, Georgia, Grenada, Haiti, Honduras, Jamaica, Jordan, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Moldova, Mongolia, Mozambique, Nicaragua, Senegal, Sri Lanka, Trinidad & Tobago, Tunisia, and Uzbekistan.

table 1. fdi inflows to nearly 100 developing countries (1998, 1999, 2000)

Dependent variable: total FDI inflows Variable of interest: U.S. BIT Using the natural log on the dependent variable (FDI inflows) allows us to chart percentage changes, rather than absolute dollar amounts. The regression results indicate that the presence of a U.S. BIT has a large, positive, and significant impact on a country’s overall FDI inflows. Coefficients on the U.S. BIT variable range from.77 to.85—which correlates with increased global FDI to a given country in a given year by 77% to 85% (at 1%–5% significance levels). As expected, a country’s GDP (lagged one year) was overall the main determinant of FDI flows, with a 1% increase in GDP correlating with a 90%–97% increase in FDI to that country. The “Rule of Law” (a measure of a country’s ability to enforce laws and regulations without bribery and the independence of the judiciary) and “Exports/GDP” (a proxy for openness to trade) variables exhibited positive significance in a number of regressions, while inflation showed a weak negative influence. OECD BITs (not including a U.S. BIT) showed a positive correlation, but without statistical significance. Non-OECD BITs showed a slight negative correlation, again without significance.

LnGDP (lagged 1 yr.) Exports/GDP (lagged 1 yr.) Population (1998 data) Rule of law 1998 data Inflation (lagged 1 yr.) non-OECD BITs U.S. BIT

1998

1999

2000

.9075312

.9738132

.9133896

(.1204325)∗

(.110394)∗

(.1009)∗

.005873

.0131968

.0124008

(.0084271)∗∗

(.00772)∗∗∗

(.0064984)∗∗∗

7.01e-10

7.02e-10

7.31e-10

(1.09e-09)

(1.03e-09)

(8.78e-10)

.5803937

.3462706

.2947284

(.2958015)∗∗

(.2928306)

(.2470544)

−.0002231

−.0019812

−.0039564

(.0014893)

(.0116331)

(.003992)

−.0161399

−.0198837

-.0167696

(.0168424)

(.0155609)

(.013368)

.7735253

.8580597

.7910127

(.3769416)∗∗

(.3515354)∗

(.3116577)∗

do bit s really work?: an evaluation of bilateral investment treaties 165

OECD BITs Constant R-squared (adj.) Observations

1998

1999

2000 .0332509

.0432709

.0309769

(.0430122)

(.0376993)

(.0323745)

−2.429067

−4.174583

−2.679342

(2.714293)

(2.510826)

(2.297455)

.58 99

.63 97

.66 94

∗ = 1% significance ∗∗ = 5% significance ∗∗∗ = 10% significance

table 2. fdi inflows to nearly 100 developing countries (1998, 1999, 2000) when oecd bits are below or above the mean (7.3)

Dependent variable: Total FDI Inflows (natural log) Variable of Interest: U.S. BIT The presence of a U.S. BIT has a strong positive correlation with FDI inflows (associated with a 112% to 157% increase in the three different years) when a country’s overall OECD BITs are below the mean (7.3 BITs). Conversely, a U.S. BIT has a very weak and statistically insignificant correlation when OECD BITs are above the mean. In both cases, a host country’s GDP has a positive (71% to 90% increase) and significant correlation (1% significance) on FDI inflows. when oecd bits are below mean (7.3) 1998 lnGDP (lagged 1 yr.) Exports/GDP (lagged 1 yr.) Population (1998 date) Rule of Law (1998 data) Inflation (lagged 1 yr.) non-OECD BITs (total #) U.S BIT (yes or no) constant –

R-squared (adj.) Observations

1999

2000

.9043972

.7193307

.8482358

(.2500843)∗

(.2067879)∗

(.2166589)∗

.0013238

.7193307

.0140727

(.0117836)

(.2067879)∗

(.0091637)

1.01e-08

1.37e-08

5.21e-09

(1.42e-08)

(1.21e-08)

(1.11e-08)

.8577123

.857234

.532797

(.4782312)∗∗∗

(.4066552)∗∗

(.3704629)

−.0210193

.0236987

.0147427

(.0104127)∗∗

(.0338313)

(.0234731)

−.0412004

−.0206846

−.0386053

(.0498417)

(.0390781)

(.0398685)

1.573043

1.848765

1.12411

(.5958077)∗

(.5258377)∗

(.4920293)∗∗

−2.018528

1.195825

−1.258251

(5.350241)

(4.435875)

(4.654583)

.46 54

.55 53

.52 50 Continued

166 jeswald w. salacuse and nicholas p. sullivan

lnGDP (lagged 1 yr.) Exports/GDP (lagged 1 yr.) Population (1998 data) Rule of Law (1998 data) Inflation (lagged 1 yr.) non-OECD BITs (total #) U.S. BIT Constant R-squared (adj.) Observations

1998

1999

2000

.7440917

1.051477

.923907

(.1468008)∗

(.1771252)∗

(.1465855)∗

.0096643

.0084455

.0101457

(.0112118)

(.0118909)

(.0098827)

5.73e-10

−2.65e-10

2.68e-10

(8.80e-10)

(1.06e-09)

(8.78e-10)

.0181132

−.0321905

.053219

(.3634513)

(.4679818)

(.3821787)

.0002231

−.0082241

−.0060239

(.0011608)

(.0126935)

(.0041829)

−.0042503

−.0051649

−.0079657

(.0128175)

(.0154286)

(.0125828)

.0054703

.3042012

.5381399

(.4080991)

(.4916508)

(.4144613)

2.303778

−5.306884

−2.285846

(3.678956)

(4.402117)

(3.631881)

.49 45

.53 44

.58 44

∗ = 1% significance ∗∗ = 5% significance ∗∗∗ = 10% significance

table 3. united states bilateral investment treaties through 2003 244 Country

Date of signature

Date entered into force

Albania Argentina Armenia Azerbaijan Bahrain Bangladesh Belarus

January 11, 1995 November 14, 1991 September 23, 1992 August 1, 1997 September 29, 1999 March 12, 1986 January 15, 1994

January 4, 1998 October 20, 1994 March 29, 1996 August 2, 2001 May 30, 2001 July 25, 1989 N/A245

244. U.S. investment in Canada and Mexico is covered by Chapter Eleven of the North American Free Trade Agreement (NAFTA) which contains provisions similar to BIT obligations. 245. Entry into force pending exchange of instruments of ratification.

do bit s really work?: an evaluation of bilateral investment treaties 167

Country

Date of signature

Date entered into force

Bolivia Bulgaria Cameroon Congo, Democratic Republic of the246 Congo, Republic of the (Brazzaville) Croatia Czech Republic247 Ecuador Egypt El Salvador Estonia Georgia Grenada Haiti Honduras Jamaica Jordan Kazakhstan Kyrgyzstan Latvia Lithuania Moldova Mongolia Morocco Mozambique Nicaragua

April 17, 1998 September 23, 1992 February 26, 1986 August 3, 1984

June 6, 2001 June 2, 1994 April 6, 1989 July 28, 1989

February 12, 1990

August 13, 1994

July 13, 1996 October 22, 1991 August 27, 1993 March 11, 1986 March 10, 1999 April 19, 1994 March 7, 1994 May 2, 1986 December 13, 1983 July 1, 1995 February 4, 1994 July 2, 1997 May 19, 1992 January 19, 1993 January 13, 1995 January 14, 1998 April 21, 1993 October 6, 1994 July 22, 1985 December 1, 1998 July 1, 1995

June 20, 2001 December 19, 1992 May 11, 1997 June 27, 1992 N/A248 February 16, 1997 August 17, 1997 March 3, 1989 N/A249 July 11, 2001 March 7, 1997 June 12, 2003 January 12, 1994 January 12, 1994 December 26, 1996 November 22, 2001 November 25, 1994 January 1, 1997 May 29, 1991 N/A250 N/A251 Continued

246. Formerly Zaire. 247. Treaty signed on October 22, 1991, with the Czech and Slovak Federal Republic and has been in force for the Czech Republic and Slovakia as separate states since January 1, 1993. 248. Entry into force pending exchange of instruments of ratification. Entry into force pending ratification by both Parties and exchange of instruments of ratification. 249. Entry into force pending exchange of instruments of ratification. 250. Entry into force pending other Party’s ratification and exchange of instruments of ratification. 251. Entry into force pending U.S. ratification and exchange of instruments of ratification.

168 jeswald w. salacuse and nicholas p. sullivan table 3. united states bilateral investment treaties through 2003 (cont’d...) Country

Date of signature

Date entered into force

Panama Panama (Amendment) Poland Romania Russia Senegal Slovakia253 Sri Lanka Trinidad & Tobago Tunisia Turkey Ukraine Uzbekistan

October 27, 1982 June 1, 2000 March 21, 1990 May 28, 1992 June 17, 1992 December 6, 1983 October 22, 1991 September 20, 1991 September 26, 1994 May 15, 1990 December 3, 1985 March 4, 1994 December 16, 1994

May 30, 1991 May 14, 2001 August 6, 1994 January 15, 1994 N/A252 October 25, 1990 December 19, 1992 May 1, 1993 December 26, 1996 February 7, 1993 May 18, 1990 November 16, 1996 N/A254

Source: U.S. Department of State, Bureau of Economic and Business Affairs, http://www. state.gov/e/eb/rls/fs/22422.htm (last visited Nov. 22, 2004)

table 4. time series data (1991–2000) U.S. FDI, total FDI, U.S. BITs (and treaty age 255 ) for thirty-one countries (year 2000) Country

Algeria Argentina Brazil Bulgaria Chile China Colombia

U.S. FDI (in $U.S. millions) 418 676 2285 11 855 1245 693

Non-U.S. FDI (in $U.S. millions) 438 11152 32779 1002 3674 40772 2374

U.S. BIT (Y or N) N Y N Y N N N

Treaty age 0 9 0 8 0 0 0

Continued

252. Entry into force pending other Party’s ratification and exchange of instruments of ratification. 253. Treaty signed on October 22, 1991, with the Czech and Slovak Federal Republic and has been in force for the Czech Republic and Slovakia as separate states since January 1, 1993. 254. Entry info force pending exchange of instruments of ratification. 255. Number of years since signing of treaty.

do bit s really work?: an evaluation of bilateral investment treaties 169

Country

Czech Republic Egypt Greece Hong Kong Hungary India Indonesia Israel Korea Malaysia Mexico Morocco Panama Philippines Poland Romania Russia Singapore Slovakia South Africa Thailand Turkey Ukraine Venezuela

U.S. FDI (in $U.S. millions) 274 603 98 −67 −1882 -67 1182 972 1244 260 3542 8 1819 49 432 33 −257 2690 32 74 539 225 24 1256

Non-U.S. FDI (in $U.S. millions) 4986 1235 –

61938 1643 2319 −4550 –

9283 3788 14706 201 603 1241 9342 1025 2714 5407 2075 888 2813 982 595 4464

U.S. BIT (Y or N) Y Y N N N N N N N N Y Y Y N Y Y N N Y N N Y N N

Treaty age 9 11 0 0 0 0 0 0 0 0 7 10 11 0 10 8 0 0 9 0 0 10 6 0

table 5. impact of u.s. bit on u.s. fdi flows to 31 developing countries, 1991–2000 Dependent Variable: U.S. FDI inflows ($U.S. millions) Variables of interest: U.S. BIT, date of signing, date of ratification, age of treaty

A U.S. BIT exhibits a very strong correlation (at a 1% significance level) with U.S. FDI outflows to a developing country, ceteris paribus, compared to U.S. flows to developing countries with no U.S. BIT. The results suggest that a U.S. BIT is correlated with an extra 1 billion U.S. dollars (see below, the coefficient on a U.S. BIT is 1019, or $1.02 billion) in increased FDI per year. Conversely, an appreciation in the bilateral exchange rate, making it more expensive for foreigners to do business, is correlated with a drop of approximately $200 million U.S. FDI per year (coefficients range from –191 to –263, at 5%–10% significance levels). The timing variables (the date of signing, the date of ratification, and the number of years since the signing) have no coefficients that are statistically significant.

170 jeswald w. salacuse and nicholas p. sullivan table 5. impact of u.s. bit on u.s. fdi flows to 31 developing countries, 1991–2000 (cont’d...) Explanatory variable of interest

U.S. BIT (yes or no, 1 or 0)

BIT signing (impact in year of signing)

BIT ratification (impact in year of ratification)

Treaty age (impact based on #/ years since signing)

GDP ($U.S. millions) (lagged 1 yr.) GDP per capita ($U.S. millions) (lagged 1 yr.) OECD BITs (total #) Bilateral real exchange rate (lagged 1 yr.) Total FDI inflows ($U.S. millions) (excluding U.S.) U.S BIT (yes or no) BIT signed (year)

2.40e-09

2.65e-09

2.65e-09

2.64e-09

(7.69e-10)∗

(7.71e-10)∗

(7.71e-10)∗

(7.69e-10)∗

.0696751

.0646923

.0715003

.0538494

(.0389278)∗∗∗

(.039679)∗∗∗

(.0395873)∗∗∗

(.0409217)

.0616844

.0413414

.0669589

.0617388

(.5032287)

(.5100287)

(.5091661)

(.5081435)

−191.1011

−258.9089

−263.2332

−254.156

(115.1843)∗∗∗

(113.2046)∗∗

(113.1953)∗∗

(113.0556)∗∗

−.0013736

−.0009197

−.0009474

−.001071

(.0089285)

(.0090331)

(.0090316)

(.0090149)

1019.474

–

–

–

(403.0047)∗ –

−156.3614

–

–

BIT ratified (year)

–

(210.5369) –

177.2043

–

Treaty age (# years) Constant

–

–

(220.4913) –

29.81412

−1251.753

−1127.018

−1202.377

−1092.922

(489.3356)∗

(496.7851)∗

(495.9503)∗∗

(496.3178)∗∗

.10 30 279

.10 30 279

.11 30 279

(23.58143)

R-sq. (within groups) .12 Countries (groups) 30 Observations 279

∗ = 1% significance ∗∗ = 5% significance ∗∗∗ = 10% significance

6. bilateral investment treaties and foreign direct investment: a political analysis∗ tim büthe and helen v. milner introduction Bilateral investment treaties (BITs), which promise foreign investors nondiscriminatory treatment and give them specific additional rights, have become popular. After a slow start, with the first BIT signed between Germany and Pakistan in November 1959 and 72 signed by the end of the 1960s, the number of BITs signed grew steadily but slowly in the 1970s and 1980s before it took off in the 1990s, with 1,857 BITs signed between the 187 members of the UN by December 31, 1999 (UNCTAD 2000, p. 1).1 Clearly, some governments have thought them worthwhile. Yet, empirical studies of the impact of BITs on foreign direct investment (FDI) have had mixed results. Some studies have found that BITs increase FDI, but empirical analyses of bilateral investment flows, in particular, have tended to find that BITs fail to boost inward FDI into the developing countries that sign them.2 We advance a theoretical and ∗ This chapter is a revised version of a paper originally presented at the Annual Meeting of the American Political Science Association (APSA), Chicago, September 2004. For comments on previous versions, the authors are grateful to Kevin Davis, Jeff Frieden, Joanne Gowa, Joseph Grieco, Nate Jensen, Benedict Kingsbury, Quan Li, Edmund Malesky, Guillermo Rosas, and participants of presentations at APSA 2004, Duke University, UC San Diego, Washington University (St. Louis), and NYU School of Law. The authors thank Zach Elkins, Witold Henisz, Nancy Brune, Nate Jensen, WDI, and UNCTAD for making data available, and Raymond Hicks, Tom Kenyon, Ivan Savic, and Matt Fehrs for research assistance. 1. 17,391 bilateral treaties would have been hypothetically possible between the then187 member states of the UN, so more than 10% of all possible BITs had been realized by 2000 (UNCTAD 2000, p. iii); By the end of 2006, 2,573 BITs had been signed between the now 192 member states of the UN (UNCTAD 2007), possibly part of a more general endorsement of neoliberal ideas by developing countries (Yackee 2005; though cf. Elkins, Guzman and Simmons 2006). 2. Many scholars have considered bilateral FDI the most appropriate measure of FDI to examine the effect of BITs on FDI, since each BIT is signed between two countries, only. We will return to this issue below. Most BITs have been signed between one developing country and one advanced capitalist country (usually a member state of the Organization for Economic Cooperation and Development (OECD)), though developing countries have increasingly signed BITs with each other; attempts to negotiate a multilateral investment agreement have repeatedly failed (see e.g., Guzman 1998; Elkins, Guzman and Simmons 2006).

172 tim büthe and helen v. milner

epistemological argument to explain the mixed results of previous studies and to advance the understanding of BITs and their effect on FDI. Statistical as well as qualitative empirical analyses provide support for our argument. We argue, first, that understanding the effect of BITs on FDI requires a political analysis of BITs. BITs are legal instruments that establish specific rights and obligations; they are part of the remarkable “legalization” of international politics in recent decades (Goldstein et al. 2001). Most strikingly, most recent BITs contain arbitration clauses that allow private parties from either signatory to initiate binding arbitration proceedings against the government of the other, without any need for either government’s approval.3 But while BITs are legal instruments, they exist to address a political problem. We focus on this political dimension of BITs to advance the understanding of the effect of BITs on FDI. In the early years after post–World War II decolonization, outright expropriation was seen as the greatest threat to foreign direct investors. Starting in the 1970s, however, the changing nature of FDI led host country governments to largely refrain from expropriation. Rather, host governments started to use a wide variety of measures—including fees, regulatory requirements for financing or purchasing, and other interventions in the market—to increase ex post their share of the benefits from FDI. When negotiating with a potential foreign investor over the terms of the investment, governments of course have every incentive to promise that they will not take any such measures. Once the investment is made, however, governments have strong incentives to renege on such promises, especially in developing countries, where the rule of law often is only weakly established and domestic courts can often not be relied upon to enforce whatever contract the foreign investor might have with the host state. If the only remedy of foreign investors is to withhold future investments, then even the costs that investors collectively may in the long run impose upon the host country are unlikely to outweigh the benefits that political leaders with short time horizons can reap in the short run from reneging on their promises. BITs address the long-standing concern about expropriation by providing assurances against arbitrary expropriation and by committing the signatories to swift, substantial compensation if expropriation were to occur. We suggest, however, that BITs should be understood as attempts to reduce the likelihood of a much broader range of interventions by committing the FDI host country to economically liberal policies and by increasing the speed and costliness of punishments for breaking such commitments.

3. Giving private parties standing in a dispute with a foreign state is a fundamental deviation from long-standing traditions in public international law—though it may be consistent with a more general shift in international businesses’ preferences for private dispute resolution fora (see Mattli 2001).

bilateral investment treaties and foreign direct investment 173

The logic of this theoretical argument, developed in greater detail in Section B, also has epistemological implications. It suggests that BITs should not only boost FDI between the signatory states but more broadly increase inward FDI into the developing country signatory. We therefore argue, second, for monadic analyses of inward FDI rather than dyadic analyses of bilateral FDI, because dyadic analyses may be ill-suited to estimate the effect of BITs on FDI. After we discuss in Section A the often narrow, legalistic conceptualization of BITs in previous studies, as well as previous dyadic and monadic empirical findings, we present our own theoretical argument in Section B. We then turn to a statistical analysis of inward FDI flows into 122 developing countries with a population of more than 1 million from 1970 to 2000—a much more comprehensive sample than in most previous analyses (Section C). Finding strong correlational support for our argument in the quantitative analyses, we turn in Section D to a qualitative analysis of the hypothesized causal mechanisms. Since the evidence here is mostly anecdotal, our findings in the qualitative section are more tentative, but they suggest cumulatively quite strongly that the positive correlation that we find between BITs and subsequent FDI is indeed driven by the hypothesized causal mechanisms. In the conclusion, we explore some broader implications and note avenues for future research.

a. previous studies of the effect of bits on fdi 1. Conceptual Differences in the Existing Literature States started signing bilateral investment treaties in 1959—and more general bilateral treaties with investment and investor protection provisions have existed at least since the 1920s (e.g., Piper 1979, pp. 332–339; Vandevelde 1988, pp. 203–206).4 Yet, only in very recent years have scholars begun to study the effect of such treaties on FDI flows. The UN Centre on Transnational Corporation’s comprehensive 1992 review of empirical studies of the Determinants of Foreign Direct Investment, for instance, devoted a single sentence in its 76-page study to BITs, confirming the lack of empirical studies of their effect on FDI (UN 1992: p. 61). Most of the early scholarship on BITs (e.g., Alenfeld 1971; Bergman 1983; Mann 1982; Sornarajah 1986; UN 1988; Voss 1981) traced the history of such treaties in mostly descriptive fashion and, in the “old institutionalist” tradition that dominated scholarship in law and political science for much of the 20th century, focused on the specific provisions of such treaties, based on the implicit assumption that the rules operate as written and that the rules as such

4. Guzman (1998) warned, however, not to overestimate the importance of the (often vague) investment provisions in the pre–World War II treaties.

174 tim büthe and helen v. milner

should therefore be the main object of study.5 Much of this scholarship also was concerned with normative questions, especially whether BITs were cumulatively creating a new body of customary international law—an important question for legal doctrine but not necessarily for explaining FDI flows. Yet, despite the dearth of positive analysis, some of those early studies provide valuable insights for understanding the effects of such treaties on FDI. Particularly notable here is the discussion of BITs by Hans-Martin Burkhardt (1986, pp. 99f), based on his many years’ experience in the German Ministry of Economics. He suggested that it was no coincidence that the earliest BITs between an advanced industrialized country and a developing country were signed by West Germany and Switzerland (Germany–Pakistan, 1959; Switzerland– Tunisia, 1961), as these countries sought formal institutional links to compensate for the lack of colonial ties and networks to provide information and to safeguard investments. Major European colonial powers such as Britain and France started to sign BITs only later (see, e.g., Gallins 1984), though Japan and the United States were even later and began to take interest in investmentspecific bilateral treaties only in the late 1970s and early 1980s, respectively. Burkhardt also noted that, by the mid-1980s, developing countries had already started to sign BITs amongst themselves (1986, p. 100; on this point, see also Sornarajah 1986; UNCTAD 2000, p. 5), and that many BITs were going well beyond guarantees for compensation in the event of expropriation by including, for instance, provisions for free capital flows as well as against regulatory interventions such as content or performance requirements (Burkhardt 1986, p. 101f). Importantly, he noted that the range of issues covered in the treaties and in the negotiations leading to such treaties were far broader than “the risks covered by investment insurance” (1986, pp. 103f). In short, while the details of BITs differ and have changed over time—and would surely warrant a more detailed analysis based on coding the specific provisions of each individual treaty—Burkhardt’s 1986 discussion of BITs suggests that these international institutions have for several decades covered a broad set of issues that affect the “investment climate” in FDI host countries. This broad understanding of BITs was largely lost in two subsequent important and highly influential studies by UNCTAD and the World Bank. The authors of the UNCTAD study (Vandevelde, Aranda and Zimny 1998) focused on BITs as legal instruments that create rights and obligations for the signatories only. They consequently provided a richly informative discussion of BITs’ specific provisions (1998, pp. 29ff) but for the same reason focused their positive empirical work on analyses of the effect of BITs on bilateral FDI flows (1998, pp. 108ff). Similarly, Mary Hallward-Driemeier’s (2003) study for the World Bank emphasized

5. For a critique of this approach to political–legal issues in international business regulation, see, e.g., Mattli and Büthe (2005).

bilateral investment treaties and foreign direct investment 175

the host country obligations to pay “just compensation” for expropriation, the foreign investors’ right to “sue” the host government (2003, p. 4), and the bilateral nature of the treaties.6 She therefore conceptualized each BIT as a way to reassure potential foreign investors from the specific OECD signatory country, only—investors who might otherwise be deterred by weak domestic property rights in the developing-country signatory and potential FDI host (2003, pp. 2ff). This approach has become the conventional wisdom and has driven the predominance of dyadic empirical analyses (where bilateral FDI flows, usually from an OECD country to a developing country, are the dependent variable), discussed below.7 The broad understanding of BITs has only been recovered in a few recent studies. Jeswald Salacuse, who in an earlier article speculated that BITs might “improve the host country’s investment climate” beyond the details of the treaty and its dispute settlement provisions (Salacuse 1990, p. 674), has in his recent work with Nicholas Sullivan suggested several ways in which BITs may affect FDI flows. Salacuse and Sullivan have argued that post-1960s bilateral investment treaties were intended to promote investment flows into the developing country signatories prospectively, in contrast to investment protection provisions in earlier international treaties, which were mostly concerned with the retrospective protection of existing investments (2005, pp. 75ff). BITs may therefore boost FDI not only through investment protection provisions—such as assurances that foreign investors will receive national or most-favored-nation treatment and be compensated in the event of expropriation, as well as commitments to a thirdparty dispute settlement mechanism such as ICSID arbitration—but also by “liberalizing the developing country’s economy as a whole.” Salacuse and Sullivan’s discussion of how BITs might contribute to such liberalization and in turn boost foreign investment remained, however, rather unspecific (2005, pp. 90ff), and although their study was not designed to differentiate between different causal mechanisms, they emphasized in the end the investment protection provisions and the associated bilateral logic of BITs affecting FDI. Eric Neumayer and Laura Spess (2005) went further. They explicitly recognized that dispute settlement (especially binding arbitration) provisions not only enhance the credibility of guarantees against classic risks in foreign investment such as expropriation, but also that the provisions give foreign investors the right to challenge in favorable tribunals other government interventions in the market if such interventions distinctly affect the foreign investor. They consequently hypothesized that “the signing of BITs sends out a signal to potential investors 6. Most BITs allow the investor to initiate proceedings for binding arbitration, not a lawsuit in conventional courts. 7. Even scholars who conducted a range of empirical analyses, such as Tobin and RoseAckerman (2005) and Salacuse and Sullivan (2005), have tended to rely on the dyadic analyses above all to draw their conclusions.

176 tim büthe and helen v. milner

that the developing country is generally serious about the protection of foreign investment [regardless of the national origin of the investor]” (2005, p. 1571). A similar understanding of BITs also informed (less explicitly) the analysis of Robert Grosse and Len Trevino (2005), who hypothesized that foreign investors view BITs not just as property rights–protection for foreign investments but also more broadly as a “signal . . . that the host country has undertaken institutional reforms toward building a market economy” (2005, p. 129).8 Consequently, both Neumayer and Spess as well as Grosse and Trevino focused on total FDI inflows into developing countries rather than bilateral FDI flows as the dependent variable. As discussed in Section B, we take a similar view of BITs and adopt a similar empirical strategy, but we rely on a logic of credible commitment (for which we specify the causal mechanisms in greater detail than previous studies), rather than a signaling logic, since we do not believe BIT signing in and of itself can bring about a separating equilibrium, which would be required for a signaling model to work.9 2. Empirically Models of BITs and FDI Scholars have adopted different approaches to modeling the effect of BITs on FDI empirically, grounded in the conceptual differences discussed above. Most analyses—and certainly the most prominent ones—have taken bilateral FDI as the primary explanandum, often in a panel setting, where multiple dyad-years are each observed over multiple years. These empirical models examine whether the existence of a BIT between the two countries in a dyad leads to higher FDI flows between them over time or in comparison with dyads that have no BIT.10 In monadic analyses, by contrast, individual host countries are the unit of observation in cross-sectional analyses (country-years in panel settings). These statistical models examine whether the total number of BITs (or BITs with certain characteristics) affect aggregate inflows of FDI into the host country. a. Empirical findings from dyadic analyses The highly influential studies by UNCTAD and the World Bank both relied primarily on dyadic analyses for their conclusions. Vandervelde et al. (1998, pp. 108ff) examined bilateral FDI flows between 1971 and 1994.11 They found a positive coefficient for the existence of a BIT in most of their statistical models, but the estimated effect was substantively

8. Grosse and Trevino’s hypothesis is grounded in a general discussion of the new institutional economics and FDI, but the particular causal mechanism linking BITs to FDI is not specified. 9. We thank Jeff Frieden for discussion of this point. 10. Since existing analyses are mostly concerned with FDI into developing countries, they usually employ “directed” dyads, examining only FDI flows from the advanced industrialized country to the developing country in each dyad. 11. For each BIT, they examined the five dyad-years prior and subsequent to the BIT signing.

bilateral investment treaties and foreign direct investment 177

“marginal” (1998, p. 122) and statistically not quite significant at conventional levels. Hallward-Driemeier’s (2003) analyses of dyadic data on bilateral FDI flows from 20 OECD countries into 31 developing countries from 1980 until 200012 raised further doubts about the effectiveness of BITs: In a number of her models, the estimated coefficient for BITs was actually negative, but in most models it was completely insignificant, suggesting that developing countries could not use BITs to attract more FDI.13 As she pointed out, there also was no positive effect for the subset of developing countries with weak domestic political institutions or weak rule of law (Hallward-Driemeier 2003, pp. 20f). While based on a small subset of developing country hosts, the findings from these two studies quickly became something of a conventional wisdom, to the point where UNCTAD’s 2003 World Investment Report drew on them to conclude: “At best, BITs play a minor role in influencing global FDI flows and explaining differences in their size among countries” (2003b, p. 89)—though UNCTAD noted the possibility that “specific countries under specific circumstances” might experience a more positive effect. The belief that BITs have no effect on FDI was further strengthened by Jennifer Tobin and Susan Rose-Ackerman’s analysis of U.S. FDI flows into 54 developing countries from 1984 to 2000.14 In country-fixed-effects models, which examined changes in each country over time, they found no significant effect for BITs—nor for political risk, though they found a seemingly perverse conditional effect when they included an interaction term between BITs and a measure of political risk: The estimated net effect of having a BIT was negative except for countries with very low levels of political risk, of which there were few among the developing countries in their sample. These estimates, moreover, were not statistically significant, possibly due to the use of three-year averages, which left them with at most six or seven observations per country as the basis for estimating the within-country effects of BITs. They consequently drew the overall conclusion that a BIT between the United States and a given developing country generally has no significant effect on U.S. FDI flows to that country. Contrary to Tobin and Rose-Ackerman, Salacuse and Sullivan found that U.S. BITs boost U.S. FDI into the developing country signatory, based on an 12. Missing data left her between 434 and 537 dyads and between 4,261 and 8,153 observations (dyad-years), depending largely on the treatment of dyad-years with missing FDI data in the OECD database on which she drew, which may or may not have been indicative of no (zero) inflows into the developing country. 13. She arrived at this finding regardless of whether she used the absolute amount of directed dyadic FDI flow as her dependent variable or FDI as a share of host country GDP (Hallward-Driemeier 2003, p. 19). 14. Specifically, Tobin and Rose-Ackerman’s dependent variable was U.S. FDI outflows to developing country X as reported by the Bureau of Economic Analysis in millions of U.S. dollars (presumably deflated) and, alternatively, FDI flows to developing country X as a percentage of all U.S. FDI flows to developing countries in a given year.

178 tim büthe and helen v. milner

examination of U.S. FDI flows to 31 developing countries from 1991 to 2000, 11 of which had signed a BIT with the United States by the end of the time period analyzed.15 They found that having a signed BIT with the United States makes a developing country highly significantly more likely to experience an increase in U.S. direct investment (Salacuse and Sullivan 2005, pp. 108ff)—though given that seven of the ten countries with U.S. BITs in Salacuse and Sullivan’s 1991– 2000 analysis signed those BITs prior to 1992,16 and given that these findings are derived from a country-fixed effects model (with apparently no correction for trend in FDI), the highly significant coefficient for BITs was based on a preciously small amount of information. It may therefore not be surprising that Kevin Gallagher and Melissa Birch, contrary to Salacuse and Sullivan, found that U.S. BITs are not correlated with higher FDI flows from the United States to the developing country signatories, based on an analysis of FDI inflows into 24 Latin American and Caribbean countries from 1980 to 2002 (Gallagher and Birch 2006, pp. 969ff),17 though Yoram Haftel (2008) has provided qualified support for Salacuse and Sullivan’s argument by showing that ratified (rather than merely signed) U.S. BITs boost U.S. FDI into developing country signatories.18 In sum, dyadic analyses have yielded mixed results, with most of them finding little if any statistically significant increase in FDI into developing countries (LDCs) as a result of BITs. These mixed findings may be due to data quality. The quality of bilateral FDI data is generally considered to be even poorer than the quality of monadic FDI data. Bilateral data also are available only for a limited (and a decidedly non-random) set of countries. More probably, though, the mixed findings are a consequence of the narrow conceptualization that underpins these analyses. If a BIT does not just boost FDI from the capital-rich treaty-signatory state (SCT) into the developing treaty-signatory state (SLDCT), but also—for reasons discussed in Section B—boosts FDI from other capital-rich states that are not party (i.e., external) to the treaty (SCXT), then dyadic analyses will pool observations of FDI flows that have been affected by the BIT (SCXTpostBIT) with observations of FDI flows unaffected by the BIT (SCXTpreBIT and SCTpreBIT). Such pooling of “treated” and “untreated” observations biases the findings toward insignificance. Monadic analysis should address this problem.

15. Compare Salacuse and Sullivan (2005, p. 108 note 195) and “Bilateral Investment Agreements Concluded: United States” (http://www.unctad.org/sections/dite_pcbb/ docs/us.pdf, last visited 3/1/2008). 16. Of the ten U.S. BITs included in the study, the BIT with Bulgaria was signed in September 1992, with Panama in October 1992, and with Ukraine in March 1994. 17. They also conducted analyses of total FDI, where they found that the total number of BITs signed was positively correlated with total inward FDI, though this effect was robust only for South America, not for Mesoamerica. 18. In addition to these studies of FDI flows, Philippa Dee and Jyothi Gali’s (2003) study of dyadic FDI stock found no effect for BITs.

bilateral investment treaties and foreign direct investment 179

b. Empirical findings from monadic analyses The UNCTAD study also included cross-sectional analyses of aggregate inward FDI flows from a single year (1995) into 133 host countries. The findings changed substantially with model specification, and BITs were so rarely significant that the authors concluded that “BITs play a minor and secondary role in cross-country comparisons of FDI determinants” (Vandevelde, Aranda and Zimny 1998, p. 120). However, as the authors noted, their setup might not have allowed them to draw any definitive conclusions even if the estimated coefficients had been more consistent, given that any number of country-specific factors, for which they did not control, might explain the variation across countries. This problem also affected the monadic analyses by Salacuse and Sullivan, who estimated cross-sectional models of aggregate FDI inflows into 99, 97, and 94 developing countries in 1998, 1999, and 2000, respectively (separate annual cross-sections). Using three measures of BITs—a dichotomous variable for U.S. BITs and count measures of other-OECD and non-OECD BITs—they found that signing a BIT with the United States has significantly boosted a developing country’s overall FDI inflows, whereas BITs with other countries have had no significant effect (Salacuse and Sullivan 2005, pp. 105, 120ff). They speculated that this finding may be due to the more demanding provisions for investment protection in U.S. BITs compared to other countries’ BITs, though high multicollinearity between their three BITs measures (2005, p. 106) made it nearly impossible to draw firm conclusions regarding the relative effectiveness of different BITs in boosting FDI. More recently, scholars have turned to panel data of aggregate FDI inflows (monadic FDI) to alleviate the substantial risk of omitted variable bias in crosssectional analyses. In most such analyses, however, the time series have remained short and the number of countries analyzed small. Grosse and Trevino (2005, p. 136), for instance, analyzed aggregate FDI inflows into 13 Central and Eastern European (CEE) countries from 1990 to 1999 (54 observations after listwise exclusion). They found the number of BITs signed by a country to be positively and statistically significantly correlated with FDI inflows into that country in random effects GLS estimates.19 By contrast, Tobin and Rose-Ackerman’s monadic analysis of five-year averages of aggregate FDI inflows into 63 developing countries from 1985 to 2000 (up to four observations per country in a countryfixed effects model) again found no significant effect for BITs, except in the conditional setting, where BITs had a negative and even statistically significant effect on FDI inflows for developing countries, except for LDCs with very low political risk (Tobin and Rose-Ackerman 2005, pp. 19–23). Neumayer and 19. The analysis built on the work of Trevino, Daniels and Arbeláez (2002), who found a positive and statistically significant correlation between the number of BITs signed by the FDI host country and FDI inflows in complementary empirical tests to their main pooled OLS analyses of 47 observations of inward FDI into seven Latin American countries between 1988 and 1999 (see Grosse and Trevino 2005, p. 129).

180 tim büthe and helen v. milner

Spess (2005), whose empirical strategy is most similar to ours, analyzed a comprehensive sample of developing countries (up to 120) from 1970 to 2000, which enabled them to conduct much more conclusive fixed-effects estimations. They found consistently a positive and statistically significant effect for BITs, both in random and in fixed effects models.20 In sum, most of the (few) monadic analyses have suffered from small sample sizes with a substantial risk of selection bias and, where panel data are employed, short time series within countries. Neumayer and Spess addressed both of these problems, though Tobin and Rose-Ackerman criticized them for the inclusion of “very small island countries” (2005, p. 23), which would violate the unit homogeneity assumption if the “true” model of FDI allocation to these countries were structurally different from the model for larger countries (as is often assumed in the literature). We address this concern by excluding from our analysis countries with a population of less than 1 million, while still analyzing FDI inflows into 122 developing (non-OECD) countries in our most comprehensive analyses.21 Neumayer and Spess also considered only a weighted subset of BITs, where the weights are themselves a function of investment flows, whereas we consider all BITs.

b. a political theory of bits and fdi In this section we will discuss our assumptions about the political problem that BITs are intended to address. We will then specify why a BIT should be seen as a broad commitment to economically liberal policies and why such a commitment should be more costly to break and therefore more credible than corresponding domestic policy choices. Finally, we will explain why a BIT should boost inward FDI into the BIT-signing developing country in general, not just bilateral FDI

20. Neumayer and Spess used total amounts of FDI inflows in constant US$ as their main dependent variable, whereas Tobin and Rose-Ackerman calculated for each FDIrecipient country(-year) what percentage those FDI inflows constitute of all developing countries’ FDI inflows; this transformed percentage measure was their dependent variable (see Tobin and Rose-Ackerman 2005, p. 16). Neumayer and Spess also used a similarly transformed measure as an alternative dependent variable (Neumayer and Spess 2005, p. 1573, 1579); their findings are robust to that change. 21. Most countries are unambiguously above or below this conventional 1 million threshold. Among the countries included by Neumayer and Spess, Antigua and Barbuda, Seychelles, St. Kitts & Nevis, and St. Vincent and the Grenadines are below 100,000; Barbados, Belize, Sâo Tomé and Principe, and St. Lucia are well below 300,000 each. Only Swaziland and Trinidad & Tobago crossed the threshold during the time period analyzed here. Swaziland was below 1 million until 1999 and was therefore mostly excluded from our analyses whenever the threshold was mechanically applied; the population of Trinidad and Tobago, by contrast, crossed the 1 million threshold from 1973 into 1974 and is therefore mostly included in our analyses. Fully including or excluding these countries did not change any of the substantive findings.

bilateral investment treaties and foreign direct investment 181

from the other BIT signatory, which suggests that monadic analyses are the most appropriate empirical strategy, where each developing country (year) is the unit of analysis and aggregate inward FDI flows are the dependent variable. 1. BITs as the Answer to a Political Problem Foreign direct investment is distinctive in that it entails the acquisition or creation of assets that are more “specific” than other transnational financial investments, i.e. they cannot be as easily sold, moved, or put to other uses in the short run without considerable loss (Coase 1937; Williamson 1985; Yarbrough and Yarbrough 1990). This asset specificity gives governments, which control how property rights can be exercised, additional leverage vis-à-vis the investor as soon as an investment is actually made (Vernon 1971). This shift in power, which is inherent in FDI, should give rise to a fundamental concern for any potential foreign direct investor, well summarized by Andrew Guzman (1998, p. 659): “Regardless of the assurances given by the host before the investment and regardless of the intentions of the host at the time, the host can later change the rules if it feels that the existing rules are less favorable to its interests than they could be.”22 This concern is real because the asset specificity creates a time inconsistency problem (Kydland and Prescott 1977), where host governments’ preferences after the investment differ predictably from their prior preferences. Distant prospects of punishment—in the form of reduced future investment or reduced benefits from the existing investments—may not solve the problem: The short-term gain of increasing their share of the benefits from existing foreign investments may lead governments to change the terms of those investments even if the costs of doing so outweigh the gains in the long run, provided that the political leader has a short time horizon.23 The time inconsistency problem is further exacerbated by the fact that governments can intervene in the market in many different ways to change the terms of a foreign direct investment (e.g., Tarzi 1991). Governments may directly change the conditions or costs of entry or exit of foreign capital through restrictions on the capital account, including new restrictions on repatriating profits. Alternatively, they may put up tariffs or nontariff barriers so as to increase the cost of importing supplies or increase the government’s take from the export of outputs. They may raise taxes, impose new fees, or change regulations in ways that diminish the value of the investment or increase the share of the profits that goes to the FDI host government. Or they may selectively enforce the law or in various other ways affect the value or profitability of the investment. We therefore

22. In principle, the same concern applies to domestic investors, but foreign investors generally have less access to the domestic political process through which such ex post changes will be decided. 23. The problem may therefore affect potential foreign investments into developing countries in particular, because politicians’ time horizons may be particularly short in cash-strapped developing countries.

182 tim büthe and helen v. milner

assume that the primary political concern for potential foreign direct investors is not so much outright expropriation (which has become a very rare event since the 1970s, see Minor 1994; Li 2005) but more broadly host governments’ commitment to economically liberal policies, as a safeguard against the broad range of interventions in the market that may diminish the value of the investment (we examine this assumption empirically in Section D.1.). And we submit that BITs should be understood as a political instrument to address this broad set of concerns by providing information and increasing the costliness of breaking commitments (thus making them more credible)—long recognized as important factors in international cooperation (e.g., Martin 2000; Milner 1992; 1997; Morrow 1994; Simmons 2000), but little explored in transnational relations between governments and foreign non-state/private actors.24 Since the unmitigated time inconsistency problem leads to suboptimal levels of investment, developing country governments, too, should have an incentive to address the problem. The unilateral (i.e., entirely domestic) adoption of liberal economic policies will surely increase a country’s attractiveness to FDI, but may not be considered a credible commitment by foreign investors unless the country has a very strong rule-of-law tradition, since short-term incentives may lead LDC governments to change or selectively apply such policies.25 By contrast, when an international agreement, formal treaty, or international organization enshrines its members’ commitment to a certain set of policies, a change in those policies has not only domestic ramifications, but also constitutes a breach of international commitments (see, e.g., Simmons 2000, pp. 821f), which makes reneging on such commitments more costly, as discussed in greater detail below. 2. BITs as a Broad Commitment to Liberal Economic Policies BITs institutionalize foreign investors’ participation in the host economy. Their specific provisions differ across OECD signatories but also among the BITs signed by a given OECD country with different LDC countries. Yet, most BITs have many common elements (see, e.g., Dolzer and Stevens 1995; Franck 2005a, pp. 1529ff; Guzman 1998, pp. 654ff; Peterson 2001; UNCTAD 2003b, pp. 87f). Almost all BITs establish entry conditions for multinational corporations/ foreign direct investors. These provisions usually do not change domestic laws, so that if a government has, for instance, a postal monopoly, foreign investors will ordinarily not be able to establish private postal services any more than

24. Mattli and Büthe (2003) argue that delegation of regulatory authority to international nongovernmental standards-developing organizations in the Agreement Establishing the World Trade Organization enhanced the credibility of WTO member states commitments not to use technical standards as non-tariff barriers to trade. Tomz (2007) explores the persistence and half-life of information and reputations in sovereign bond markets. 25. This may be why Glen Biglaiser and Karl DeRouen (2006) found that market-friendly economic reform in Latin American countries do not always lead to increased FDI.

bilateral investment treaties and foreign direct investment 183

domestic investors. However, they usually commit each signatory to allowing nationals and economic entities from the other signatory to bring capital for direct investments into the country without being subjected to taxes or fees for entry; they also commonly commit the signatories to allow the unimpeded repatriation of profits from foreign direct investments. Reflecting lingering concerns about (rare but always very prominently publicized) instances of expropriation, BITs contain provisions that may constrain nationalization and always guarantee compensation for expropriation if it were to occur (many BITs also promise compensation for any destruction of property in the event of civil conflicts or war). And among the “procedural rights that permit the enforcement of the substantive rights,” BITs in recent years have tended to contain the kinds of binding arbitration provisions discussed above, which allow foreign investors to initiate a dispute over violations of BIT provisions directly against the FDI host government. In sum, most BITs commit the signatories not to impose capital account controls that impede foreign investment or the repatriation of profits; they commit the signatories to pay compensation in the even of expropriation, and they commit host governments to submit to arbitration in the event of a dispute over these commitments. Yet, most BITs go far beyond these common provisions. They generally guarantee, for instance, national treatment for foreign investors with respect to establishing and operating a business in the FDI host country. At the same time, they grant rights to the foreign investor that no domestic investor has, such as the right for a review of any alleged violation of the government’s obligations under the treaty by arbitration panels whose composition is controlled in part by the foreign direct investor. Most BITs also guarantee most-favored-nation treatment, that is, they guarantee that each signatory state will treat investors from the other signatory state no less favorably than investors from anywhere else. Most importantly, many BITs contain general—vague but potentially sweeping—provisions by which the signatories commit not to take any regulatory, legislative, or administrative steps that would diminish the value of the investment, including policies that would impede the import of inputs or export of outputs. Thus, the Denmark-Lithuania BIT of March 1992, for instance, specifies in Article 3(1) that “neither Contracting Party shall in any way impair by unreasonable or discriminatory measures the management, maintenance, use, enjoyment or disposal of investments in its territory . . .”26 In other words, the BIT commits the government not to interfere in the market in ways that would diminish the value of the investment to the investor. Moreover, the arbitration provisions usually apply to all of the commitments undertaken in the treaty, effectively leaving it to the arbitration tribunal to interpret the breadth or depth of the governments’ 26. The text of this and most other BITs is available from the UNCTAD website’s Investment Instruments Online; http://www.unctad.org/sections/dite/iia/docs/bits/ denmark_lithuania.pdf (3/1/2008), emphasis added.

184 tim büthe and helen v. milner

obligations under the treaty (such as what would be a “reasonable” exception under the Denmark-Lithuania BIT). In sum, BITs give foreign investors rights that go well beyond unimpeded capital flows and guaranteed compensation for expropriation. Rather, BITs constitute a commitment to economically liberal policies across a broad range of issues, as underlined by arbitration decisions (though the interpretation of BITs by arbitration panels has been far from perfectly consistent; see Franck 2005a; 2005b). 3. Credibility of Internationally Institutionalized Commitments Why should we expect that foreign investors will consider the commitments undertaken via BITs to be more credible than promises that host governments might make directly to them? Why should these internationally institutionalized commitments be less susceptible to the time inconsistency problem? We argue that BITs provide information about the nature of the commitment and about any actual occurrence of a violation—and they provide mechanisms for the enforcement of those commitments. All of this increases the costs of reneging and/or the speed with which governments that renege on their commitments will likely incur those costs, which in turn should raise the credibility of those commitments in the eyes of foreign investors. a. Ex ante information & BITs Charles Lipson (1991, p. 501) suggests that states enter into formal agreements such as treaties in part to make commitments more “visible.” Such ex ante information about international agreements, especially when it makes the nature and specific provisions of the commitment known, facilitates detecting violations of the commitment and in turn should allow for swifter punishment. There are good reasons to think that this logic applies to BITs. The conclusion of negotiations for a BIT, the signing of the BIT, and (where separately needed) the ratification of a BIT are usually publicized in government press releases and often undertaken in public ceremonies reported in the media. Such initial publicity should provide at least some information about the nature of the commitments undertaken. In addition, governments (particularly governments of regular FDI “home” countries) may provide information about the country’s BITs on an ongoing basis. In the United States, for instance, designated offices in the Department of State (Office of Investment Affairs) and in the Office of the U.S. Trade Representative (Office of Services, Investment, and Intellectual Property) exist to answer American investors’ questions about their rights under U.S. BITs. Moreover, information provision is not restricted to the two signatory countries. BITs are covered by the UN requirement to notify all treaties to the UN, and UNCTAD has for a number of years by now made the full text of most BITs available on its website. We may therefore expect that a multinational corporation (MNC) or individual investors considering a foreign direct investment in a given developing country may already know or can easily learn whether this LDC has a BIT with his/her own country, toward how many other countries the

bilateral investment treaties and foreign direct investment 185

country has undertaken similar commitments, and what the terms of those agreements are. b. Ex post information and BITs After the initial publicity subsides, BITs continue to generate information that facilitates identifying and punishing those who renege on their commitments. Such information is most likely to be generated by the aggrieved investor him/herself through the dispute settlement process (discussed in greater detail under “Enforcement” below). Investment treaties tend to include provisions for binding arbitration of any disputes under the treaty and increasingly specify the World Bank’s International Centre for Settlement of Investment Disputes (ICSID) as the arbitrator before which private investors may bring a dispute directly against the FDI host government. Arbitration before an ICSID panel makes the host government’s alleged violation of its commitment a matter of public record: Pending cases as well as decisions are published on the World Bank website and in the ICSID newsletter.27 In addition, governments and private actors are more likely to provide information about noncompliance with commitments under a BIT than about noncompliance with commitments that have not been institutionalized internationally. A government’s compliance with its treaty obligations is often monitored by the other governments that are parties to the international agreements—more closely and continually than policy commitments that a government may undertake domestically or via individual investors. The U.S. Department of State’s Office of Investment Affairs, for example, publishes annual Investment Climate Statements for some 120 countries. Prepared by government officials based in part on reports from U.S. diplomatic missions abroad as well as private investors, each statement notes any investment agreement that the country has with the United States and reports outstanding disputes or issues that have arisen in the implementation of those agreements, including property rights and trademark protection issues as well as dispute settlement problems.28 On the nongovernmental side, international agreements create incentives for domestic groups that benefit from the commitments to make government violations of those commitments public, and they legitimate such domestic, private information provision about a government’s failure to live up to its commitments (Cortell and Davis 1996; Dai 2005). This information effect should increase with the number of BITs an FDI host country has signed, since each BIT increases the number of parties monitoring a government’s policies.

27. A number of BITs, especially older ones, also arrange for ad hoc arbitration outside the ICSID framework, but we have less systematic knowledge about them, since such arbitration proceedings may be confidential. As William Rogers (2000) points out, governments’ desire to avoid the reputational damage of even being publicly accused of treaty violations may create perverse incentives for “unhappy investors . . . to [falsely] complain that a financial or business failure was due to improper regulation, misguided macroeconomic policy or discriminatory treatment by the host government . . . ” 28. See http://www.state.gov/e/eeb/ifd/c9787.htm (3/1/2008)

186 tim büthe and helen v. milner

By generating information about each instance of reneging on (or ex post rejection of) an international agreement, BITs also raise the reputational costs of such behavior. Violating an institutionalized commitment—or not making amends to correct a violation that has occurred—damages a country’s reputation for keeping commitments, making future cooperation on the same and other issues more difficult and maybe impossible to achieve (Abbott and Snidal 2000; Simmons 2000).29 And violating any specific BIT also generates costs in interactions with countries that are not a party to that treaty, because it constitutes a violation of the broader social norms affirmed through the agreement (Snidal and Thompson 2003, p. 200). The greater the number of countries with which the host has signed BITs, the greater is the number of countries that may infer from the breach of any one BIT that the commitments undertaken vis-à-vis them are now also in jeopardy. The costs of violating BIT commitments should thus increase with the number of BITs signed by a given FDI host country. As a consequence, a potential foreign investor should have ample opportunity to assess any host government’s record of compliance with its BIT obligations, not just under a BIT with the foreign investor’s own country (if such a BIT exists) but under all BITs signed by the potential host country. If investors extrapolate from past to future behavior, adding a risk premium for uncertainty, then signing additional BITs should boost the attractiveness of a developing country to foreign investors (assuming constant levels of compliance), because each BIT provides the investor with additional information. Signing additional BITs should therefore increase a country’s attractiveness for foreign direct investors, leading to increased inward FDI in the aggregate. c. Ex post enforcement & BITs Like many other formal international agreements, BITs lead to the creation of mechanisms that make it easier for private actors to solicit assistance from their home government to bring diplomatic pressure to bear on “a government that is considering or engaging in rule violation” (Simmons 2000, p. 821). The U.S. government offices that are designated to provide information about BITs (discussed above) also exist to pursue intergovernmentally any complaints by those who believe that their rights under a treaty have been violated by a foreign government. European governments similarly have designated officials in ministries for foreign affairs or economics/ commerce (and in their embassies) as the first contacts and representatives of their nationals vis-à-vis foreign governments if the former feel that a current or forthcoming policy threatens their interests in violation of a BIT. As a European government official told us: “Of course, we always try to be supportive of our citizens if they have been harmed by the actions of a foreign government, but it 29. These reputational costs should be even higher when the international institution provides for an arbitral or judicial mechanism, as BITs generally do, which provides an independent confirmation that a country has indeed violated its treaty obligations (Abbott and Snidal 2000, p. 427).

bilateral investment treaties and foreign direct investment 187

makes a difference whether there is a treaty, because we can be more forceful if there are quasi-legal rights and obligations.”30 Others, like the government of Australia, have similar institutional access points. In short, the existence of a BIT makes it easier for foreign investors to recruit the assistance of their home governments to bring costly pressure to bear on FDI host country governments that renege on their commitments to economically liberal policies. BITs, however, go beyond this common feature of international agreements. Investment treaties now almost always include an ex ante commitment to binding arbitration and—in a radical departure from the dominant tradition in international law (Salacuse and Sullivan 2005, p. 88)—create the right for private parties in one state to take the government of the other state “to court” (i.e., to an arbitrator).31 To be sure, these provisions provide no absolute guarantee that governments will not change policies to the detriment of foreign investors. Yet, while they provide no guarantee, they indeed raise the costs of reneging on the commitments made in the BITs, as evidenced by the outcomes of such disputes. The World Bank’s International Centre for Settlement of Investment Disputes (ICSID), which has increasingly become the standard designated arbitrator in BITs, registered its first case arising under a BIT arbitration provision in 1987 (ICSID 2000, p. 7). Having proven considerable independence from governments by rendering 10 of its first 18 decisions against governments (foreign investors prevailed, at often substantial economic and/or political costs to the host country government, see IBRD 2005, p. 181), ICSID has continued to attract investment disputes, with 87 cases brought under BITs by February 2003 and 144 by the end of June 2005 (ICSID 2003ff).32 In sum, enforcement procedures established by (or as a consequence of) BITs enable foreign governments and private actors to impose higher economic and political costs on governments that renege on their policy commitments—and to do so more quickly—than in the absence of BITs. By increasing the likelihood

30. Not-for-attribution interview with BIT specialist of a European government, Nov. 2005. 31. Arbitration provisions in BITs varied considerably in the early years (Sornarajah 1986, p. 96 esp) but have become much more consistent in the last two decades (Dolzer and Stevens 1995; interviews of the authors with government officials in the United States and Europe); under most recent treaties, private parties have standing before the designated arbitrator regardless of their home government’s approval (Salacuse and Sullivan 2005). 32. This pattern raises the question why governments generally accept the outcomes of the judicial or arbitration processes even when the decisions go against them. While proper research of this question is beyond the current chapter, nonacceptance of arbitration decisions would most immediately undermine the very objective that led most countries to institutionalize their commitments in the first place, namely to gain the economic benefits of increased FDI; it also would much more broadly damage their reputation for honoring commitments and thus might deprive them of a range of benefits of international cooperation (see discussion of reputational costs above).

188 tim büthe and helen v. milner

and the time-discounted magnitude of the punishment for reneging, international institutions should reduce the time-inconsistency problem posed by FDI for developing country governments. As Salacuse (1990, p. 674) noted, a “BIT’s mandatory dispute settlement provisions and the ultimate prospect of compulsory arbitration will cause host country officials to pause before taking actions toward foreign investment.” 4. Who benefits? The effect of BITs on FDI BITs’s treaty provisions create specific legal rights for nationals of the other signatory. They consequently should boost the FDI flows from the capital-rich treaty-signatory state, SCT, to the developing country treaty-signatory, SLDCT. We submit, however, that they should be expected to affect FDI well beyond bilateral FDI flows, boosting inflows from other capital-rich states that are external to the treaty (SCXT) into SLDCT. This should occur for several reasons. First, each BIT signed by a given SLDCT may boost what we have called the ex ante informational effect of other BITs signed by that SLDCT and should increase the reputational cost of reneging on the commitments undertaken via those BITs, as discussed above. Second, economic nationality is not well defined and may be stretched in the event of a dispute. Any specific BIT thus directly creates rights for a broad set of potential or actual investors. In one of the most prominent investment disputes of recent years, for example, a mostly American-owned, Bermuda-headquartered, Nasdaq-traded multinational, Central European Media Enterprises (CEME), was able to bring a case against the Czech government under a BIT between the Czech Republic and the Netherlands, because a Dutch subsidiary of CEME had been involved in some of the transactions (CEME also brought a case over the same issue under a BIT between the Czech Republic and the United States in London). The CEME dispute turned on regulatory actions and omissions by the Czech Broadcast Regulation Council in 1999, which had allowed CEME’s local business partner, on whose Czech citizenship CEME was relying for its broadcasting license, to squeeze out CEME and render its investment in a successful Czech TV station largely worthless. CEME’s founder Ronald Lauder, a U.S. national, lost his dispute under the Czech Republic-U.S. BIT,33 but the arbitrator for the Czech-Dutch BIT (in this case a panel of the Arbitration Institute of the Stockholm Chamber of Commerce), found in September 2001 that, given the actions and omissions of the Czech regulator, the Czech state had failed its obligation under the BIT to protect and fairly treat the foreign investor; it awarded CEME US$ 355 million—more than twice the market capitalization of CEME before the award and equal to the entire annual budget of the Czech Ministry of Health (Desai and 33. The tribunal under the Czech Republic-U.S. BIT found that the Czech Regulator had discriminated against Lauder as a non-Czech citizen, but that Lauder had failed to show that his subsequent loss of the investment was due to this violation of the BIT. It therefore awarded no damages.

bilateral investment treaties and foreign direct investment 189

Moel 2006; Franck 2005a, pp. 1559ff; Green 2003). This was not an isolated case: Japanese MNC Nomura Group Companies has brought a case against the Czech government over the insolvency of a Czech bank partly owned by Nomura under the Czech Republic-Netherlands BIT (Goldhaber 2003); U.S. MNC Bechtel has brought a case against Bolivia over the privatization of a water utility based on a Bolivia-Netherlands BIT (Mahnkopf 2005, p. 134); and Italian MNC Carmuzzi has brought a case against Argentina over public utilities pricing in the aftermath of its 2001 currency crisis on the basis of BITs that Argentina had signed with Belgium and Luxembourg (Goldhaber 2003)—to name just a few. Third, many BITs contain clauses that guarantee treatment of foreign investors on par with the treatment afforded to investors from the “most favored” country. As a consequence, investors from a country that has signed a BIT with a given FDI host may be able to take advantage of more far-reaching provisions in a BIT that the host has signed with another country. Developing country governments that are tempted to treat foreign investors differently, depending on whether or not the investors have certain rights under specific BITs, will therefore have to tread carefully. Uncertainty over the applicability of BITs provision (beyond the core group of investors unambiguously associated with the treaty partner) creates incentives for adopting liberal policies toward investors in general, rather than selectively.34 In sum, there are several reasons to expect BITs to affect not just bilateral FDI flows but inward FDI flows into BITs-signing developing countries much more broadly. This has important implications for empirical analyses of the effect of BITs on FDI flows into developing countries. It suggests that such analyses should not focus on bilateral flows, where inappropriate pooling may bias the results, but on aggregate FDI inflows into any given developing country. This argument for a monadic empirical strategy is further strengthened by an artifact of FDI statistics: Bilateral FDI statistics generally record only the immediate source or destination of the funds, which may obscure the flow of funds and the reasons for investment. If the April 1992 BIT between Spain and Uruguay,

34. Moreover, irrespective of how broad the group of investors who arguably gain direct rights under a given BIT, the commitment to economically liberal policies toward these foreign economic actors also creates incentives to adopt and maintain economically liberal policies in general. This is not to say that liberal foreign economic policies and liberal domestic economic policies automatically go together; domestic policies remain subject to political contestation, even after a country makes a commitment to a liberal foreign economic policy via a BIT, and developing countries may, like OECD countries, exhibit partisan differences in the kind of foreign direct investments they like to see (Pinto and Pinto 2008). Yet, there are strong economic incentives to maintain (more) marketeconomic policies when economies are (more) open, so as to be able to reap the full benefits from international (in this case financial) liberalization (see, e.g., Chang, Kaltani, and Loayza 2005; Frieden and Rogowski 1996). BITs thus may work to institutionalize a country’s commitment to liberal foreign and domestic economic policies not just selectively vis-à-vis treaty partners, but also in general.

190 tim büthe and helen v. milner

for instance, spurred a Spanish multinational corporation to make an investment in Uruguay, but the Spanish MNC decided to make that investment via its Latin American head office at its subsidiary in Argentina, the investment would be recorded (by Uruguay and Argentina) as Argentinian FDI in Uruguay. Dyadic analyses of bilateral FDI flows between Spain and Uruguay would find no effect and might, in fact, find a negative effect if they compared the flows from Spain to Uruguay (for a dyad with BIT) with the flows from Argentina to Uruguay (for a dyad without BIT). This is not just a hypothetical scenario. A recent detailed study of U.S. FDI inflows into Vietnam, for instance, gathered data on individual investment transactions and found that a large share of the growing investments by U.S. multinational corporations into Vietnam have been administratively handled by the U.S. MNCs’ existing East-Asian subsidiaries in, for instance, South Korea and Japan, and that these investments consequently were not recorded as U.S. FDI into Vietnam, but as FDI from South Korea into Vietnam and from Japan into Vietnam (FIA 2005, p. 4).35 Bilateral FDI data thus might obscure the effect of BITs or even cause the researcher to get it entirely wrong. We therefore focus on monadic analyses of FDI inflows, where the effect on FDI should be proportional to the number of BITs the FDI host (developing) country has signed.36 Our central hypothesis for the empirical analysis is therefore: The greater the number of BITs to which a developing country is a party, the more attractive will foreign investors consider it to be as an investment location, and the more inward FDI will it receive, ceteris paribus.

c. statistical analysis 1. Setup of the core statistical analysis a. Sample To test our hypothesis statistically, we conduct statistical analyses of annual flows of inward foreign direct investment into developing (non-OECD) countries with a population of more than 1 million. Our dependent variable, annual inward FDI, is the sum of the year’s flows of direct investments into a given “host” country by capital owners that are foreign to that country (net of direct investments withdrawn by foreign capital owners), calculated as a percentage of GDP.37 We restrict the analysis to developing FDI host countries

35. We thank Eddy Malesky for bringing this study to our attention. 36. The effect need not be linear, but we lack theoretical reasons for expecting any other particular functional form ex ante and therefore estimate a linear relationship. 37. This measure of FDI, which is employed in many recent analyses of FDI, eliminates the need to deflate the dependent variable and makes it easily comparable across countries and over time. Our data is from the online version of UNCTAD’s Handbook of Statistics (see UNCTAD 2003a for details).

bilateral investment treaties and foreign direct investment 191

by excluding all OECD country-years for three reasons: (1) our theoretical logic suggests that the political determinants of FDI into developing countries may differ from those for FDI into advanced industrialized countries; (2) our empirical findings from interviews with business managers (see Section D) suggest that their key concerns about the political environment differ systematically depending on whether they are considering a foreign direct investment in a developing or an advanced industrialized country; and (3) a recent study by Bruce Blonigen and Miao Wang (2005) shows that pooling “wealthy and poor” countries in statistical analyses of the determinants of FDI may lead to biased results or erroneous inferences. We also restrict our analysis to country-years during which the FDI host was an independent country (i.e., was in a position to set its own policies toward FDI) and had a population of more than 1 million, since previous studies have suggested that the logic of FDI into very small countries may be different, as a few big investments may play a very large role for a small economy.38 There have been 129 independent non-OECD countries with a population of more than 1 million in existence at some point in time between 1970 and 2000. Our analyses cover up to 122 of these countries and up to 31 years for each of them39—a much broader sample than in most previous studies, whose findings may be biased if data are missing in non-random fashion. b. Estimation methods We conduct “fixed effects” or “within [country]” analyses of inward FDI because the theoretical logic of our argument suggests first and foremost that a given developing country should experience higher inward FDI after signing one or more BITs. Such a boost in FDI should come in addition to whatever level of inward FDI the country might experience for other reasons, including reasons that may be specific to each country. Our argument thus suggests an effect within countries over time. The logic of the argument also suggests that countries with more BITs should receive more FDI than countries

38. The 1 million cutoff is ultimately arbitrary but common in the literature. It also assures the exclusion of very small advanced industrialized countries that may not be OECD members. Restricting our analysis to countries with populations greater than 1 million has two further advantages: First, very small countries often display extreme values and wide swings on variables of interest, which make them outliers or even influential points when they are included along with larger countries. Second, data coverage is much poorer for countries with a population of less than 1 million, and casual inspection of the data suggests that data are missing nonrandomly, so that including the remaining observations for small countries would introduce bias. See footnote 21 for further discussion of the threshold. 39. As virtually all economic analyses, we have no data for Afghanistan, Cuba, Iraq, Libya, Myanmar/Burma, the Democratic People’s Republic of [North] Korea, and Somalia. Listwise exclusion for missing data and the independence criterion also restrict the average length of the time series to just under 21 years, though there is considerable variation, not least because for instance the successor states of the Soviet Union do not enter the sample until the 1990s.

192 tim büthe and helen v. milner

with fewer BITs (ceteris paribus, i.e. after all other nonrandom factors have been controlled for), but there is no reason to think that the magnitude of this “crosssectional” effect across countries should be the same as the magnitude of the within-country effect over time, and the within-country effects over time are of primary interest for assessing the theoretical argument.40 Moreover, modeling the allocation of FDI across countries (beyond a few widely agreed factors, most importantly market size and level of economic development) has proven extremely difficult in prior research due to multicollinearity and unobserved differences between countries. Analyses with “country-fixed effects” safeguard against the resulting problems of multicollinearity and omitted variable bias by including a dichotomous “dummy” variable for each country. The coefficients on the dummy variables then “predict” the average FDI inflows for each country and thus soak up all of the cross-national variance in the dependent variable, including cross-sectional variance that may be explained by the cross-sectional component of the explanatory variables in the regression model. As a consequence, for each explanatory or control variable, only the variation over time within each country is brought to bear when estimating a country-fixed effect model, and coefficients that result from such a “within estimation” only predict the variation over time in the dependent variable, that is, the change over time that remains after subtracting the country-mean level of FDI from each observation.41 Since within analyses are akin to time series in a panel setting, all the usual problems of time series analysis may interfere with drawing valid inferences. Most importantly, when there is a trend in the dependent variable and any of the explanatory variables, we might find a statistically significant but entirely spurious correlation due to these co-trending series (e.g., Davidson and MacKinnon 1993, pp. 670ff). To test for the presence of a trend, we regress each variable on a trend term (with country-fixed effects to allow for a country-specific intercept of the trend) before using the variable in our analyses. If we find a statistically significant trend, we use the de-trended residuals from this test as the explanatory variables in our main analysis, a solution to co-trending series that generalizes from time series.

40. In preliminary analyses, we in fact found that FDI data from recent decades fails standard tests for using simple OLS on the pooled data (treating all country-years in our panel as independent and unit-homogenous); the data also fails standard tests for using “random effect” feasible generalized least squares (GLS) estimation (Neumayer and Spess (2005) obtain similar findings). Note that failing random effects tests can be due to deficiencies of the model specification and therefore is not necessarily indicative of real differences between the cross-sectional and “within” effects. 41. By predicting perfectly the average value of the dependent variable (inward FDI flows) for each country, the country dummy variables effectively capture the crosssectional component(s) of a potentially infinite number of “unobserved” factors. See Hsiao (2003) and Wooldridge (2002) for more detailed, technical discussion.

bilateral investment treaties and foreign direct investment 193

Finally, even with de-trending and country-fixed effects, the errors may exhibit heteroskedasticity or autocorrelation. We therefore use the standard errors for within estimators proposed by M. Arellano (1987), which are robust to both heteroskedasticity and autocorrelation and generally yield conservative inferences. Using these clustered robust standard errors rather than the regular standard errors of ordinary least square (OLS) regression makes it less likely to find a statistically significant coefficient when the true effect of a variable is zero. c. Baseline model As discussed in Section A, much of the existing work about the effect of BITs on inward FDI flows into developing countries has used small and differing sample sizes and short time series, which might explain the divergent findings. Since small and divergent panels often result from the inclusion of variables that are only available for a limited number of developing countries or only quite recent years (with data probably missing in a non-random fashion), we start from a baseline model that only includes variables that have persistently been found to be strong predictors of FDI flows, leaving the inclusion of additional controls for robustness checks. Specifically, our baseline model includes two political variables, one policy/outcome variable (trade), and three economic control variables: Market size. Research on the determinants of FDI in economics has focused on various characteristics of the host market (see Blonigen 2005 for a recent comprehensive review). These studies have very persistently found the size of the host market to be a substantively and statistically significant factor, with larger markets attracting more FDI, in many studies even on a per capita basis. Large markets are particularly attractive for horizontal foreign direct investments, where multinationals replicate their entire production process in multiple countries to serve (primarily) the domestic market in the host country, but the finding appears to apply to FDI in general. The finding, however, is primarily based on cross-sectional analyses, as even in panel settings almost all of the variation in available measures of market size is cross-sectional. We effectively control for market size by using FDI-as-a-percentage-of-GDP as our dependent variable. To be able to distinguish any possible remaining effect of market size cleanly from the level of economic development, we use (the natural log of) a country’s population as our measure of market size, with population data drawn from the World Bank’s World Development Indicators (WDI) database. However, 97.8% of the variation in the log of population is cross-sectional, and country-fixed effects will capture all cross-sectional variation in our models.42 Population also shows a clear upward trend over time, so that we de-trend this control variable, using the de-trended values in the regressions. Trend and country-fixed effect together account for 99.8% of the variation in the log of 42. All specific information about variance in the individual variables reported here is based on the sample used in Models 1 and 2, but changing the sample size did not lead to materially different findings on variance.

194 tim büthe and helen v. milner

population, leaving us with a variable that shows hardly any variation over time within each country and is bivariately negatively correlated with inwards FDI. We nonetheless include this variable since controlling for market size is customary, but one should not make too much of any findings, given the limited residual variance.43 Economic development. Research in economics has also very consistently shown that richer, more developed countries attract more FDI, though there is some question over whether this finding equally holds for developing countries, where vertical FDI often involves locating labor-intensive stages of the production process in countries with low wages, which might result in some FDI into developing countries seeking low rather than high levels of economic development (Blonigen and Wang 2005; Hanson, Mataloni and Slaughter 2005).44 As our measure of economic development, we use the natural log of per capita GDP in constant 1995 dollars (again from WDI). For economic development, too, the general finding is primarily based on cross-sectional variation in levels of economic development, and 97.2% of the variation in economic development is cross-sectional. Country-fixed effects and the trend term combined soak up 97.5% of the variance in this control variable. Estimated coefficients for economic development should therefore not be over-interpreted, either. Economic growth. The third and final economic variable that existing studies have consistently shown to be a very strong predictor of inward FDI is economic growth, measured as the annual growth rate in a country’s real GDP (GDP growth, again from WDI). High or above-average economic growth is often seen as indicative of investment opportunities, and previous studies have found a significant positive coefficient for GDP growth. Average growth rates also vary across countries, but only 19.7% of the variance in GDP growth is cross-sectional. Moreover, while there is a slight, statistically significant upward trend in GDP growth rates, country-fixed effects and de-trending combined soak up only 20.3% of the variance in this variable. FDI of course might and in fact should lead to subsequent increases in GDP growth, but since we lag GDP growth by one year, we hope to capture the effect of growth on FDI, rather than the other way around. Trade openness. Another variable that many previous studies have found to be an important predictor of FDI is trade. While some horizontal FDI is motivated by high costs of exporting to the host country and might therefore be a substitute for trade (which would imply a negative coefficient), the importance for developing

43. This problem is common to all available measures of market size in (strongly warranted) fixed effects estimations. We alternatively employed other measures of market size in models not shown here; their use did not change any of our findings. Dropping market size altogether did not change our results for BITs, either. 44. Categorizing any particular investment as horizontal or vertical is in practice extremely difficult, and no good aggregate data exist distinguishing between these conceptually very distinctive types of FDI (see Markusen and Maskus 2004).

bilateral investment treaties and foreign direct investment 195

countries of vertical FDI, where FDI and trade are complements, leads us to expect a positive coefficient for this variable, as in most previous studies of inward FDI into developing countries. We use the sum of exports and imports as a percentage of GDP (from WDI) as our measure of trade. This measure might be interpreted as simply a trade flows measure, but international trade does not just “happen” but is at least in part a function of policy decisions to permit or impede cross-border product market transactions to a greater or lesser extent. We therefore think of this variable (especially after country-fixed effects, which soak up 87.6% of the variance) not just as an economic but as an economic policy variable—an empirical manifestation of trade openness.45 Domestic political constraints. Investors—foreign and domestic—like predictability. They should consequently welcome constraints on policymakers’ ability to change policy. Sometimes this is seen as one of the benefits of (liberal) democracy, where media scrutiny of the government’s policies and greater openness of the political process should provide earlier and better information, which should at a minimum be a safeguard against sudden, unexpected changes in policy (Jensen 2003). But empirical findings regarding the effect of democracy on foreign direct investment have been mixed, probably because common measures of democracy force onto a single dimension a number of characteristics of political regimes that might be perceived quite differently by foreign investors (Feng 2001; Harms and Ursprung 2002; Jensen 2006; Kahler 1981; Li and Resnick 2003; Oneal 1994). A more specific measure of institutionalized domestic political constraints on national political leaders has been developed by Witold Henisz. His preference-weighted index of domestic institutional veto points, which we use as our measure of “domestic political constraints” (see Henisz 2000; 2002) is strongly positively correlated with measures of democracy, but it captures substantial variation on this particular dimension among democracies as well as among non-democracies, which may be why studies that have used this variable have consistently found it to be a good predictor of inward FDI (e.g., Henisz and Zelner 2001; Henisz and Macher 2004).46 We therefore include domestic political constraints rather than a measure of regime type in our baseline model. Political instability. Political instability and political violence should make a country less attractive for foreign investors, since they render the economic and political context less predictable (Schneider and Frey 1985). Empirical research using varied methods has found consistent support for this argument (e.g., Brunetti, Kisunko and Weder 1997; Jun and Singh 1996). We use the composite measure from Arthur Banks’ (1999) dataset of political events that indicate political violence and instability (coups, assassinations, general strikes, guerilla 45. De-trending only adds marginally to this: Country fixed effects and de-trending combined soak up 88.5% of the variance. 46. 51.3% of the variance is cross-sectional; country fixed effects and de-trending combined soak up 58.9% of the variance in Domestic Political Constraints.

196 tim büthe and helen v. milner

warfare, government crises, purges, riots, revolutions, antigovernment demonstrations) as our measure of political instability. Average levels of political instability differ across countries, but only 33.7% of the variance is cross-sectional, which leaves us with substantial variation on this variable to possibly explain changes in inward FDI within countries over time.47 2. Statistical findings a. Basic model with BITs This baseline model of FDI (Model 1 in Table 1) explains 5.2% of the variance in inward FDI flows, after country-fixed effects and de-trending have already accounted for 39.2% of the variance. The estimates from OLS regression with Arellano (1987)-type clustered standard errors suggest that economic growth, trade openness, domestic political constraints, and political instability all have a statistically significant effect on FDI in the expected direction. We then add our key variable of interest, cumulative BITs, i.e., the number of bilateral investment treaties to which a country is a signatory (from UNCTAD 2000).48 46.4% of the variance in BITs (among the developing countries in our broadest sample) is cross-sectional, and there is a significant trend, so that country-fixed effects and de-trending soak up 62% of the variance in this key explanatory variable. When we add the residual of this variable (after de-trending and country-fixed effects) in Model 2, OLS with clustered standard errors estimates a highly statistically significant positive coefficient for this variable. This finding suggests that BITs are indeed boosting FDI inflows into developing countries. More precisely, since country-fixed effects by design account for average levels of FDI in each country, we find that, the higher the number of BITs to which any given country is a signatory, the greater will be the amount of foreign direct investment into that country, ceteris paribus and after de-trending. The estimated effect for domestic political constraints is modestly reduced and political instability misses conventional levels of statistical significance, after including (and thus controlling for) BITs.49 At the same time, the explanatory leverage of the model increases notably.50

47. Political Instability is also the only variable in our baseline model that does not exhibit a significant trend. 48. We correct for the double-counting and some errors in UNCTAD’s comprehensive, official listing of all BITs. 49. The highest bivariate correlation among any of the de-trended variables in this model is 0.3584, suggesting no major concerns about multicollinearity; the correlation between domestic political constraints and our measure of BITs is 0.2292. 50. Here, as before, the 6.41% of variance explained (R2) is on top of the 39.2% explained by the country fixed effects and de-trending. Relative to the 5.2% additional variance explained by model 1, 6.4% constitutes more than a 1/5 increase in explanatory leverage.

bilateral investment treaties and foreign direct investment 197

table 1. effect of bit s on inward fdi flows into ldc s

Cumulative BITs

Model 1

Model 2

Model 3

Model 3’

Model 4

–

0.0464∗∗∗

0.0421∗∗∗

0.0423∗∗

0.0372∗∗

(.0132)

(.0137)

(.0176)

(.0174)

Financial openness index

–

Good policy index

–

–

0.148

–

∗∗∗

0.129

∗∗

0.0901∗

(.0502)

(.0519)

(.0470)

–

–

0.278∗∗∗ (.0653)

Domestic political constraints

∗∗∗

∗∗∗

2.24

(.686) ∗∗

Political instability –0.0156

(.00764)

Trade openness

∗∗∗

0.0210

(.00678)

Market size Economic development GDP growth Constant

1.75

(.684)

(.714)

(.633)

(.633)

–0.0134

–0.0128

–0.0139

–0.0138

∗∗∗

0.0199

(.00680)

(.00812)

0.0145

∗∗∗

(.00541)

(.00872)

0.0199

–1.17

–2.13

(1.42)

(1.38)

(1.38)

(1.92)

–0.840

(.526)

–0.592

(.527) ∗∗∗

–1.21

(.493) ∗∗∗

0.0347

0.0338

(.0107)

(.0105)

(.0109)

–8.90e

(1.16e )

122 2524 +0.0520

–7.97e

−10

−9

(1.16e )

122 2524 +0.0641

–4.24e

−10

−9

(1.54e )

121 2499 +0.0672

(.00851)

–0.0195∗∗∗ (.00566)

–2.00 (1.81)

∗∗

–1.15∗

(.583) ∗∗∗

0.0373

−10

∗∗∗

(.00584)

–1.98

–0.452

1.18∗

1.80

(.00836)

1.09

∗

–3.09∗∗

−9

n N R2

∗∗

0.0392

(.601) ∗∗∗

(.00966)

1.97e

−9

−9

(1.28e )

82 1785 +0.0990

0.0337∗∗∗ (.00950)

1.83e−9 (1.27e−9)

82 1785 +0.1092

Note: OLS within-estimates with Arellano (1987) robust (clustered) standard errors in parentheses; all estimates rounded to three significant figures. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01; two-tailed tests. Analyses cover 1970-2000, subject to data availability; all explanatory variables enter with a 1-year lag. All variables de-trended, except “political instability,” which exhibited no significant trend. Country-fixed effects implemented in advance via “areg” command, with “absorb(country)” in Stata 9.2. R2 information indicates additional variance explained by the variables shown, after country-fixed effects and trend have explained 39.2% of the variance in the raw FDI data (40.7% for Model 3; 48.7% for Model 4); R2 not fully comparable across models when sample size changes.

198 tim büthe and helen v. milner

To provide a sense of the substantive significance of these findings, Table 2 shows the magnitude of the estimated effect for each variable. The first column notes the magnitude of one standard deviation for each independent variable. The second column notes the change in our dependent variable (de-trended, post-fixed effects FDI as a percentage of GDP), which results from a one standard deviation increase in each explanatory or control variable, ceteris paribus. Thus, a one standard deviation increase in the number of BITs signed by a given country is estimated to result in an increase of FDI equal to just under 0.3% of that country’s GDP. The information in the third column helps to assess whether such a change is substantively significant (the dependent variable has a standard deviation of 2.32).51 A change of 0.291 thus amounts to 12.5% of a standard deviation in the dependent variable.

table 2. estimated substantive effects, model 2 1 standard deviation increase in . . .

Cumulative BITs Dom. political constraints Political instability Trade openness Market size Economic development GDP growth

. . ., which is equal to . . .

. . . results in this change in FDI as a % of GDP:

. . . which is equal to this percentage of a std deviation in the dependent variable

6.27 0.111

+0.291∗∗∗ +0.201∗∗∗

12.5% 8.6%

3.96 14.7 0.0618 0.191

–0.0531 +0.293∗∗∗ –0.122 –0.161

2.3% 12.6% 5.3% 6.9%

5.86

+0.204∗∗∗

8.8%

Note: All estimates based on de-trended values as discussed in the text. Due to the implementation of country-fixed effects in advance, all the variables have a mean of zero by design. For estimated change: ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01; two-tailed tests. Standard deviation and estimated effects rounded to three significant figures; percentage rounded to first decimal.

51. By construction—that is, as a consequence of having implemented the country fixed effects at the stage of de-trending in order to allow the intercept for the trend to vary across countries—de-trended, post-fixed-effects FDI as a percentage of GDP has, like the explanatory variables, a mean of zero.

bilateral investment treaties and foreign direct investment 199

These estimates suggest that BITs have as much of an effect on FDI as trade openness, where a standard deviation increase in trade as a percentage of GDP increases FDI by 12.6% of a standard deviation, and a substantively larger effect than any of the other variables. To put this into perspective: A country would have to achieve almost a one and a half standard deviations increase in economic (GDP) growth to attract increased FDI equivalent to the increased FDI that results from a one standard deviation increase in BITs. b. International institutionalized commitment or domestic policy choice? Our analysis so far has shown that BITs are positively and statistically significantly correlated with subsequent inward FDI into developing countries. We have argued that BITs constitute a commitment to economically liberal policies, and that BITs boost FDI because they make such commitments more costly to break and hence more credible. It is conceivable, however, that international investors merely respond to the domestic policy preferences/choices of the developing country government, not to the international institutionalization as such. If international agreements merely commit governments to doing what they already want to do anyway (Downs, Rocke and Barsoom 1996; von Stein 2005) and are signed at about the same time as the domestic policy changes, then our finding of a positive correlation might be spurious, rather than indicative of foreign investors attributing greater credibility to commitments undertaken via BITs. To test this alternative explanation for our statistical finding in Model 2, we introduce, separately and jointly, two direct measures of domestic policy preferences. First, we add Nancy Brune’s financial openness index, which measures the degree to which a country restricts capital account transactions, including FDI inflows themselves or the repatriation of profits. Higher values indicate greater openness and thus more liberal policies. This variable directly captures some of the specific policy commitments that LDC governments undertake in BITs and should therefore have a strongly positive effect on FDI (and reduce the estimated effect of BITs). And indeed, we estimate a statistically highly significant positive coefficient for this measure of domestic policy choice in Model 3. The estimated coefficient for the financial openness index suggests that a standard deviation increase in this variable results in an increase in inward FDI equivalent to 0.218% of host country GDP, which amounts to 9.5% of a standard deviation in our dependent variable. The estimated effect of BITs is slightly reduced, but remains substantively and statistically highly significant. This finding suggests that institutionalizing policy commitments via BITs indeed substantially increases the credibility of those commitments and hence boosts FDI beyond the boost that results from domestic policy choice alone. A broader measure of domestic policy choice is Craig Burnside and David Dollar’s “good policy index,” which is a composite measure of domestic economic policies and foreign economic policy, where higher values indicate more liberal policies (Burnside and Dollar 2000). Unfortunately, even the updated index (by Easterly, Levine and Roodman, 2003) is only available for 82 non-OECD

200 tim büthe and helen v. milner

countries with a population of more than one million, so that we lose about 30% of our sample when including this variable. Since there are strong reasons to suspect that reliable data on domestic policy choices is missing in a nonrandom fashion, we are less confident about the findings for models that include this index than for our other models, but the breadth of this index should make it informative. Due to the loss in sample size, we first re-estimate Model 3 for this sample (Model 3 in Table 1), then add the good policy index in the last column of Table 1 (Model 4). The strongly statistically significant positive coefficient for the good policy index suggests that foreign investors indeed respond very favorably to liberal economic policies, independent of international institutions such as BITs: A one standard deviation increase in this index is estimated to result in an increase in FDI equivalent to 11.0% of a standard deviation in FDI. As a consequence of including this variable, the estimated effect for financial openness is reduced and now is only weakly statistically significant anymore, probably because the good policy index captures some aspects of financial openness (the correlation between the two indices is 0.350).52 The estimated effect of BITs is also reduced, as expected, but remains statistically clearly significant and substantively still suggests an increase of 11.9% of a standard deviation in FDI for a one standard deviation increase in BITs.53 In sum, here again, BITs boost FDI well beyond investor-friendly domestic policy choices. We conclude from these additional analyses that our findings for BITs are not spurious. c. Alternative estimation techniques For the analyses shown in Table 1, we have used OLS with clustered standard errors, which under most conditions of fixed-effects or “within” estimations yields the most conservative estimates (Kézdi 2004; Wawro and Kristensen 2007). Since the use of clustered standard errors in the analysis of panel data is not yet very common and some readers may hence be hesitant to have confidence in the results, we re-estimate Model 3 (the most inclusive model from Table 1 in terms of variables and sample size) using alternative estimation techniques. The results are shown in Table 3. In preliminary analyses of OLS with regular standard errors, the BreuschGodfrey test for autocorrelation in the errors (which generalizes from time series to panel data) indicated first order (but no higher order) serial correlation in the error terms, suggesting that the use of OLS with regular standard errors is not appropriate for these data. We nonetheless report the estimates for OLS with regular standard errors in the first column of Table 3 to allow readers to see the

52. When the Good Policy Index is entered without the Financial Openness Index, then its coefficient is larger (0.371) and similarly strongly statistically significant, whereas the estimated coefficient for BITs is marginally larger, but the R2 is reduced to 0.0817. 53. De-trending and country fixed effects must be implemented anew every time that there is a change in the sample, resulting in a small change in the standard deviation of each variable.

bilateral investment treaties and foreign direct investment 201

results. To take account of the first order autoregressive (AR(1)) process generating the error term, we then re-estimate the model using feasible generalized least squares (GLS), once allowing for an AR(1) process that is common across the units (countries) and once allowing for a country-specific AR(1) process. Finally, we re-estimate the model with “panel-corrected standard errors” (PCSE), as proposed by Nathaniel Beck and Jonathan Katz (1995), with results shown in the last column of Table 3.54 table 3. alternative estimations of model 3 OLS (regular std. errors) Cumulative BITs

GLS (common AR(1))

GLS (countryspec. AR(1))

PCSE (common AR(1))

0.0421∗∗∗

0.0362∗∗∗

0.0436∗∗∗

0.0452∗∗∗

(.00815)

(.00417)

(.00406)

(.00980)

∗∗∗

∗∗∗

∗∗∗

0.133∗∗∗

Financial openness index

0.148

(.0314)

(.0166)

(.0168)

(.0339)

Domestic political constraints

1.75∗∗∗

0.392∗∗

0.553∗∗∗

1.23∗∗∗

(.419)

(.182)

(.146)

(.409)

Political instability

–0.0128

–0.00316

–0.00181

–0.00808

(.0114)

(.00283)

(.00184)

(.00634)

0.0145∗∗∗

0.00434∗∗

0.00575∗∗∗

0.00413

Trade openness

(.00321)

Market size

–1.17 –0.592

Constant n N

(.00179)

0.0338

–0.463

∗∗∗

0.00534

–1.24∗∗∗

–1.75∗

(.457)

(1.05)

–0.0975

–0.639∗

(.150)

(.135) ∗∗∗

(.00656)

(.00170)

(.436)

∗∗

(.263)

GDP growth

0.0942

–0.116

(.815)

Economic development

0.114

∗∗

(.352) ∗∗

0.00482

(.00790)

(.00250)

(.00214)

–4.24∗e−10

0.00127

0.000737

(.0445)

(.0232)

(.0226)

121 2499

117 2495

117 2495

0.0149∗ (.00765)

–0.0199 (.0964)

121 2499

Note: Standard errors in parentheses; all estimates rounded to three significant figures. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01; two-tailed tests. Analyses cover 1970–2000, subject to data availability; all explanatory variables enter with a 1-year lag. All variables de-trended, except “political instability,” which exhibited no significant trend. Country-fixed effects implemented in advance via “areg” command, with “absorb (country)” in Stata 9.2.

54. Since none of these methods deal with the problem of possible spurious correlation arising from trending series, we use the de-trended values for all of these re-estimations.

202 tim büthe and helen v. milner

The magnitude of the estimated effect changes, naturally, with the use of these alternative estimation techniques, but the estimated coefficient for our main variable of interest (cumulative BITs) remains positive and substantively significant, regardless of the estimation technique used. Moreover, the estimated coefficient for BITs is statistically even more highly significant with any of the alternative estimation techniques. d. Additional robustness checks To probe the soundness of our findings further, we conduct a series of additional robustness checks (results not shown but available on request). We consider various alternative measures for market size and level of economic development. We also drop from the model—separately or in combination—any variables that are often statistically insignificant: market size, level of economic development and political instability. None of these changes alter our main result that BITs have a positive, substantively and statistically significant effect on inward FDI into developing countries. We also consider a series of additional economic and political control variables. Most importantly, we consider three regime-type measures of domestic political institutions: Alvarez, Cheibub, Limongi and Przeworski’s dichotomous measure of democracy, the 21-point summary measure of regime type from the Polity IV dataset, and Freedom House’s three-point “Freedom” index. These measures are correlated with Henisz’s measure of veto points but measure primarily electoral democracy, rather than constraints. Whether these variables are entered separately or in combination with domestic political constraints, the coefficients estimated for the democracy measures are correctly signed but not even close to conventional levels of statistical significance, while BITs remain substantively and statistically significant. Another important issue is whether our findings might be driven by one or a few particular countries that are outliers. Here, a key concern is whether East Asian or Eastern European economies are driving our results. These countries arguably have experienced exceptionally high levels of FDI in the 1990s (and beyond) and also have signed a large number of BITs. We therefore re-estimate our main models after omitting countries from these regions individually and in groups. These sample restrictions change the estimated coefficients on the margins, but do not affect the overall finding that BITs boost inward FDI to a substantively and statistically significant extent. In Table 4, we report the results for the full sample in the first column, for the sample without the countries categorized by the OECD as Eastern European in the second column, and for the sample without the countries categorized as East Asian or South Asian.55

55. We use the OECD’s regional classification system, rather than the World Bank’s because the OECD uses a more fine-grained regional classification, which for instance allows us to differentiate between Eastern Europe and Central Asia, though results after also excluding the eight Central Asian countries included in the World Bank region “Europe & Central Asia” are very similar. The OECD’s definition of East Asia is very narrow (the Democratic People’s Republic of Korea, Mongolia, the People’s Republic of China, the

bilateral investment treaties and foreign direct investment 203

table 4. sample restrictions Model 3 full sample Cumulative BITs

0.0421∗∗∗ (.0137)

Model 3 without Eastern Europe

0.0402∗∗ (.0155)

∗∗∗

Model 3 without East/Southeast Asia

0.0469∗∗ (.0179)

∗∗∗

0.135∗∗

Financial openness index

0.148

(.0502)

(.0509)

(.0531)

Domestic political constraints

1.75∗∗

1.60∗∗

1.98∗∗∗

(.714)

(.792)

(.739)

Political instability

–0.0128

–0.0127

–0.0104

(.00812)

(.00808)

(.00758)

Trade openness Market size Economic development

∗∗∗

0.0145

0.0160

0.0138∗

(.00541)

(.00572)

(.00761)

Constant

–0.967

(1.38)

(1.41)

(1.40)

–0.592

–0.618

–0.841

–1.21

(.534) ∗∗∗

(.567) ∗∗∗

0.0338

0.0376

(.0109)

(.0121)

–4.24e

−10

(1.54e−9)

n N R2

∗∗∗

–1.17

(.493)

GDP growth

0.148

−10

0.0295∗∗∗ (.0109)

6.85e

2.80e−9∗∗

(1.30e−9)

(1.39e−9)

122 2499 +0.0672

103 2373 +0.0602

110 2255 +0.0616

estimated effect of 1 std dev 0.291 increase in BITs (= % of std dev in FDI:) 12.5%

0.238

0.252

10.2%

10.7%

Note: For details on sample restrictions, see text. OLS within-estimates with Arellano (1987) robust (clustered) standard errors in parentheses; all estimates rounded to three significant figures. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01; two-tailed tests. Analyses cover 1970–2000, subject to data availability; all explanatory variables enter with a 1-year lag. All variables de-trended, except “political instability,” which exhibited no significant trend. All estimates in Stata 9.2. R2 not fully comparable across models due to changes in sample size.

Republic of Korea, and Taiwan only), so we are showing the results for excluding the countries categorized as either as “East Asia” or “South Asia,” which captures (inter alia), Indonesia, Malaysia, the People’s Republic of China, Philippines, Singapore, Thailand and Vietnam, i.e., all of the East Asian countries that have arguably received disproportionately high FDI. Here, results based on excluding the countries categorized as being in “East Asia/ Pacific (Oceania)” by the World Bank again lead to very similar results. Note the relatively small loss of observations from the exclusion of Eastern Europe because many of the countries in

204 tim büthe and helen v. milner

Similar results pertain when excluding China by itself or jointly with other, smaller groups of East or Southeast Asian countries. The estimated substantive effects for these restricted samples are even less changed than the estimated coefficients might suggest, since re-implementing detrending and country-fixed effects after the sample restriction changes the standard deviations of several variables. Specifically, the estimated effect of a 1 standard deviation increase in BITs for the sample without Eastern Europe is an increase in FDI-as-a-percentage-of-GDP of 0.238, which is equal to 10.2% of a standard deviation. For the sample without the East and Southeast Asian countries, the estimated effect is 0.249 (10.7% of a standard deviation in FDI). In sum, we find that BITs boost FDI not only for a small number of countries or a few high-FDI regions. 3. Power and institutions As noted in Section A, Salacuse and Sullivan (2005) argue that a BIT between the United States and a developing country attracts more FDI into the developing country than a BIT between that developing country and any other OECD country (or at least that U.S. BITs boost FDI more consistently and therefore have a statistically more significant effect). As an empirical matter, U.S. BITs might indeed attract more FDI than other OECD countries’ BITs on average, though multicollinearity in Salacuse and Sulivan’s statistical analysis makes it impossible to draw such a conclusion with confidence. Theoretically, their hypothesis is based on the assumption that the specific provisions in U.S. BITs impose more far-reaching obligations on the FDI host government than BITs of other OECD countries. Yet, we do not actually know to what extent U.S. BITs are distinctive; some scholars emphasize the striking and increasing similarities in the terms of BITs, while others point out differences. A proper analysis of the logic of an argument about the effect of treaty provisions would therefore require a detailed coding of the provisions in a large random or representative sample of existing BITs. Some scholars have recently begun such a coding (see, e.g., Peinhardt and Allee 2007, Katzenstein 2008), building on Barbara Koremenos’s pioneering work in coding international agreements (e.g., Koremenos 2005). Their research holds great promise for advancing this literature (we will return to this issue in the conclusion), but has not yet yielded any firm findings. Nor do we know whether potential foreign investors actually consider the detailed provisions of the BITs, though we know that some investors are taking into account how many BITs have been signed by a country that they are considering as an investment location (see qualitative analysis, below). There is, however, another—more overtly political—reason to think that BITs may differ in their effectiveness, even if their specific provisions may not meaningfully differ. We have argued that BITs make reneging on one’s commitments more costly not just because they allow aggrieved investors to take FDI host governments this region became independent only after 1989, and almost all of them lacked FDI and/or GDP data for the years prior to the transition from Communism to capitalism.

bilateral investment treaties and foreign direct investment 205

to an arbitration panel, but also because they make it easier for investors to solicit the diplomatic assistance of their home governments, which might alleviate the need to initiate arbitration proceedings. And states differ in the power resources that they are able to bring to bear in intergovernmental settings, even when the formal rules of diplomacy or international institutions nominally consider all states to be equal (e.g., Krasner 1991). This logic leads to a more general hypothesis (complementary and supplemental to our main hypothesis), namely that the effectiveness of a BIT may be partly a function of the relative power of the other signatory vis-à-vis the developing country FDI host that is our unit of analysis. We test this hypothesis by creating a weighted measure of BITs, which uses each country’s GDP as the proxy for its power in the international political economy. Specifically, we use a variant of the “cumulative BITs” measure where each BIT has been weighted by the relative power (GDP) of the signatories. Specifically, we calculate our measure “weighted BITs” as the sum of the weighted BITs a country has signed:

. . . where X is our measure of weighted [cumulative] bits, BIT is a dichotomous indicator variable coded 1 if a BIT existed between countries i and j in year t (zero otherwise), subscript i signifies the FDI host country for which the weighted BITs measure is recorded, and subscript j signifies the (potential) signatory of a BIT with country i, where n is the universe of all independent countries in year t.56 Since the GDP weights result in a measure of a different order of magnitude, we divide it by 100 before using it the analysis. This measure, weighted BITs, is only correlated at 0.157 with our regular cumulative BITs, and the correlation is further reduced to 0.0503 after de-trending and country-fixed effects. We can therefore use the two variables in the same regression. Table 5 reports the results. In the left column, we re-estimate Model 3 for this slightly reduced sample (we lose some observations due to the GDP weights); Model 5 then includes the weighted BITs measure. For the weighted measure, we estimate a positive and statistically significant effect, which suggest an additional substantive boost in FDI equal to about 74% of the magnitude of the effect estimated for BITs alone. Importantly, however, cumulative BITs remains highly significant and in fact its coefficient is very little reduced from the model without weighted BITs. This finding suggests that (1) BITs have a substantial effect on FDI that is independent of the relative power of the signatories and (2) BITs signed with powerful FDI home states have a substantial additional effect. In other words, BITs alone are a credible 56. For purposes of calculating this weighted measure (as well as our Cumulative BITs measure), we thus consider all BITs, including BITs with OECD countries and BITs with countries whose populations is less than 1 million, although we do not analyze those countries as FDI hosts. In the regression tables, by contrast, n refers to the number of countries included in the analysis as FDI hosts (only).

206 tim büthe and helen v. milner table 5. power-weighted bit s

Cumulative BITs

Model 3’’

Model 5

0.0422∗∗∗

0.0397∗∗∗

(.0136)

Weighted BITs

(.0133)

0.0258∗∗

–

(.0112)

Financial openness Dom. political constraints Political instability Trade openness Market size Economic development

0.142∗∗∗

0.133∗∗∗

(.0491)

(.0452)

∗∗

1.74∗∗

1.77

(.715)

(.702)

–0.0121

–0.0150∗

(.00810)

(.00817)

∗∗∗

0.0153

0.0148∗∗∗

(.00537)

(.00513)

–1.36

–0.978

(1.39)

(1.41)

–0.554

–0.263

(.500)

(.469)

GDP growth

0.0327

∗∗∗

0.0303∗∗∗

Constant

2.41∗e−9∗∗

2.37∗e−9∗∗

(9.48∗e−10)

(9.46∗e−10)

0.0697

0.0756

(.0108)

R2

(.0105)

Note: OLS within-estimates with Arellano (1987) robust (clustered) standard errors in parentheses; all estimates rounded to three significant figures. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01; two-tailed tests. All variables de-trended, except “political instability,” which exhibited no significant trend. All estimates in Stata 9.2. Analyses cover inward FDI flows 1970–2000; all explanatory variables enter with a 1-year lag. Country clusters (n): 121; N = 2,478.

commitment mechanism and signing more of them improves the credibility of this commitment since it means more countries can punish and monitor behavior. In addition (i.e., controlling for the pure number of BITs), BITs with (economically) bigger countries bring in even more FDI because they increase the likely cost of punishment if the FDI host were to violate its commitments.

d. qualitative analysis In this final section, we want to complement the strong correlational findings of our statistical analyses with a brief qualitative analysis of key elements of the

bilateral investment treaties and foreign direct investment 207

causal mechanism that we have hypothesized. This analysis allows us to test the plausibility of the micro-logic of our argument. Our findings here strengthen the confidence we can have in our explanation. 1. Core substantive assumption: Investor concern We started from the assumption that a key concern of potential foreign investors, when they ponder investment opportunities in a developing country, are the political risks that arise from a broad range of political interventions in the market, which may diminish the value or profitability of an investment. We have assumed that this concern is not just (and in recent decades not even predominantly) a concern about outright expropriation but about the potential for detrimental policy change in a range of fields, including foreign economic policies (such as capital account controls) as well as domestic policies (regulation, law enforcement, etc.). We have deduced from this core substantive assumption the expectation that foreign investors generally prefer economically liberal policies, and that BITs boost FDI because they constitute a credible commitment to such liberal policies. While assumptions are never “true” and mostly heuristic, Ronald Coase (1982) forcefully argued that patently false assumptions may lead to models that are neither theoretically insightful nor useful for policy, and that the plausibility of central assumptions should therefore be examined. We find ample evidence to suggest that our assumption is plausible. In a survey of its members in the late 1990s, the U.S. Chamber of Commerce, for instance, found political risks to U.S. companies’ real or intangible property from government policies that amounted to “creeping expropriation” among the ten most important concerns of U.S. firms considering investments abroad (U.S. Chamber 2000). This concern about government intervention in the market does not appear to be just an American preoccupation: In a series of interviews with German senior managers on the factors that make for a good investment climate in a given country, interviewees tended to distinguish first between countries where physical and intellectual property was essentially secure (as they generally assumed to be the case in OECD countries) and countries where it was not. For the latter category, restrictive foreign economic policies, domestic regulation in general, and more broadly government intervention in the markets—not mentioned as a general concern regarding investments in OECD countries—were seen as generally undesirable, with a broad preference for economically liberal policies implied or even directly expressed.57 Whose political risks matter, though? It may be argued that investors care only about safeguarding their own investments against government intervention, rather than economically liberal policies in general. In fact, one might speculate

57. Interviews of some sixty senior managers of German firms were conducted by Tim Büthe in 2000–2002.

208 tim büthe and helen v. milner

that if a government is pressed for resources, investors prefer a blatant violation of someone else’s property rights (such as the expropriation of their assets) over an across-the-board increase in taxes or fees. A content analysis of the Wall Street Journal’s and Financial Times’ coverage of the 2004 expropriation of Yukos assets by the Russian government is insightful on this point: While the expropriation was at first presented primarily as a problem for those holding Yukos stock (Betts 2004, Jack 2004; FT 2004), it soon came to be seen as a more general threat to all current or planned investments in Russia (McDonald and Sender 2004; Karmin 2005). By the end of 2004, in explicit response to the expropriation of Yukos assets, some existing inward FDI was reversed as a consequence of “the perception [that] the risk of doing business [in Russia] has increased” (Arvedlund and Mouawad 2004). Further, as “aftershocks” kept “foreign investors on edge” (Buckley 2005), potential new FDI stayed out. Responses to the Yukos episode thus suggest that foreign investors care about the general approach of the government to private economic actors and their investments, rather than just acting on the promises made to them about the treatment of their individual assets. This finding is also supported by results of surveys among OECD country business executives (see, e.g., IMD 2001). Such surveys tend to find consistently a preference for economically liberal policies among those who make most foreign direct investment decisions. In sum, we find substantial and even quite direct support for our core assumption. 2. Causal mechanism Based on the above assumption, we developed in Section B a theoretical argument that led us to predict that BITs should be positively correlated with subsequent inward FDI into developing countries in fixed effect (within-country) analyses. We have found strong evidence of such a correlation in the empirical analyses in Section C. How plausible is the hypothesized causal mechanism? An examination of the causal mechanism turns on whether BITs indeed generate the hypothesized informational and enforcement effects and consequently make reneging more costly, whether they are indeed perceived in that way by governments, and whether investors take notice of them and perceive them as constraining governments.58 We offer here at least a tentative assessment, based on existing accounts, some documents, and interviews that we have conducted with business managers and government officials. a. Informational effects of BITs Do BITs generate ex ante information? As noted above, both governments usually announce the treaty signing. OECD countries often have designated offices to provide detailed information about their BITs to investors or at a minimum disseminate the terms of agreements 58. Investor perception would be sufficient for the effect to obtain, but if such perceptions were divorced from actual constraints, we would need to explain how such erroneous perceptions arise and persist.

bilateral investment treaties and foreign direct investment 209

such as the UK Investment Promotion and Protection Agreements (i.e., UK BITs), which are publicized by the Foreign and Commonwealth Office. Government officials in these offices report receiving inquiries about the details of a given BIT from individual firms “quite often” and sometimes even from outside consultants and legal experts who advise firms considering foreign direct investments.59 Developing countries differ in the extent to which they actively publicize information targeted at foreign investors. Botswana’s Export Development and Investment Authority, for instance, advertises BITs that the country has signed on its Web pages for foreign investors, as one of the reasons for why MNCs should choose Botswana for their investments (along with other investment protection measures, assurances of political stability, etc.).60 Other developing country governments do not quite advertise their BITs on the World Wide Web, but according to Western government officials who frequently interact with developing country officials and with investors from their own (FDI home) country, developing country officials regularly point out BITs to potential foreign investors and to foreign officials.61 In a 1997 UNCTAD meeting, for instance, Chinese representatives noted: “We are often consulted by foreign investors and our own overseas investors on BITs, especially when a large amount of investment and investment in some sensitive sectors such as natural resources, public utilities, are to be made” (Vandevelde, Aranda and Zimny 1998, p. 120 note 10). Host countries for (potential) FDI thus invoke BITs implicitly or even explicitly as an assurance that they will not go back on their commitments to provide a favorable investment climate for foreign investors. Just as importantly, business and industry associations and organizations such as the U.S. Chamber of Commerce often report about new BIT signings or existing BITs. Finally, media reporting of a given BIT rarely covers the specific provisions of an investment treaty, though publications focused on international commerce often provide more detailed information, and insofar as BITs are very similar to each other, detailed information may not be needed. Do BITs generate ex post information about (non)compliance? As noted in Section 3.b., governments regularly gather information about their treaty partners’ compliance with existing BITs, based importantly on investors’ reports. Some make that information widely public (such as in the U.S. Investment Climate Statements); others provide it more privately. An Australian government official, for instance, told us that they routinely evaluate the compliance of Australia’s treaty partners under trade and investment agreements and share that assessment as part of advising Australian investors on international investments, in particular since it is a major factor in the risk assessment for investment guarantees

59. Not-for-attribution interviews, Nov.–Dec. 2005. 60. See http://www.bedia.co.bw/ (3/1/2008). 61. Not-for-attribution interviews, Nov. 2005.

210 tim büthe and helen v. milner

or insurance.62 And when a private investor initiates as well as when he or she wins or loses a major dispute at ICSID, it not only is noted on the World Bank website and in the ICSID newsletter, but also gets reported in major international newspapers, such as the Financial Times, New York Times, Wall Street Journal, Frankfurter Allgemeine Zeitung, Figaro, or Neue Zurcher Zeitung, as well as more specialized business publications. The media also often report about disputes in front of other arbitration panels, which, seeking to maintain discretion, may not themselves advertise the dispute. By all indications, this information is reaching investors. How many BITs a given country has signed and with whom may not be quite common knowledge or the subject of cocktail party conversations among business managers. Yet, senior executives of multinational companies who have been involved in FDI decisions and foreign investment advisors to major U.S. and European MNCs have told us that they look for information about BITs and a country’s record of disputes under their BITs as one of the pieces of information that is easy and quick to obtain but also (in their assessment) genuinely informative.63 And a German official at the 1997 UNCTAD conference noted his government’s observation that “many investors . . . postpone their investments until their establishment is protected by a BIT,” which he interpreted as an indication that “the business community seems to be aware of additional benefits of these agreements” (Vandevelde, Aranda and Zimny 1998, p. 120 note 10). b. Cost of breaking BIT commitments BITs threaten punishments for breaking the commitments thus undertaken. If such threats were perfectly successful in deterring noncompliant behavior, such threats would never have to be carried out, which makes such “negative sanctions” very inexpensive in an environment where compliance is high (Baldwin 1985; 1989 [1971]). But just as domestic laws, international treaties rarely lead to perfect compliance. In fact, our theoretical argument in no way predicts perfect compliance, but only that BITs increase the costs of breaking promises or commitments toward foreign investors and therefore make it less likely that a host government will break these commitments. We therefore might see some instances of violations of BIT commitments, but we would expect governments that break their BIT commitments to incur real costs for doing so—which might in fact strengthen the BIT regime by making the costliness of noncompliance visible. Reneging on the commitments made in BITs has indeed proven costly: The arbitration provisions of BITs have been successfully used to seek compensation (of tens, even hundreds of millions of dollars in some cases) for allegedly 62. Not-for-attribution interview, June 2006. OECD (FDI home) country governments here tend to take a fairly liberal view of economic nationality, so that the “investors” whom they support through advice and assistance may include multinational corporations that have headquarters or subsidiaries in the FDI home country. 63. Not-for-attribution interviews.

bilateral investment treaties and foreign direct investment 211

damaging policies far broader than classic expropriation. Policies or actions that have been the subject of arbitration proceedings under BITs have included corruption, administrative/regulatory measures, and allegedly biased law enforcement (e.g., Franck 2005b), consistent with our interpretation of BITs as a broad commitment to liberal economic policies. In addition, governments of FDI home countries have on numerous occasions intervened diplomatically with the FDI host before a dispute has reached the arbitration stage (at which point governments ordinarily take a strict hands-off approach).64 c. BITs as credible commitments and the perceptions of investors There is strong evidence that FDI home governments view BITs not just as a way to reduce restrictions on FDI and the repatriation of profits or a way to gain an assurance against outright expropriation, but that they see them as means for bringing about (commitments to) economically liberal policies quite broadly. Jeffrey Lang (1998), for instance, reports that OECD country governments have long sought BITs as a device for strengthening developing countries’ general commitment to liberal economic policies. According to a U.S. official, serious negotiations for a BIT (which sometimes are requested by the developing country, at other times initiated at the request of U.S. private parties, usually via members of Congress) are preceded by an assessment whether the potential treaty partner is “politically willing and administratively ready or capable” to sign up to the major obligations that a BIT entails, including a “real commitment to the rule of law, intellectual property rights, etc.”65 Such screening, however, only establishes the potential for a pro-active agenda going forward. Summarizing his interviews with government officials in advanced industrialized countries and especially the United States, Salacuse (1990, p. 76) note that “although the BITs themselves do not specifically enunciate the goal of investment and market liberalization, that goal has clearly been in the minds of developed country negotiators and is sometimes reflected in background documents.” Similarly, European governments whose officials we interviewed conveyed their view of BIT negotiations as “an opportunity to have a series of conversations about the political, administrative, and economic conditions that are needed for foreign investments”—conversations that would not occur or be awkward or even inappropriate outside the context of BIT negotiations.66 These conversations, government officials told us, sometimes take place in the context of a liberalization of economic policies; at other times, they lead to such a liberalization, as “considerable learning takes place” on the part of the LDC government in the course of the negotiations, which consequently tend to take at least one to two years

64. Not-for-attribution interviews with North American and European government officials, Nov.–Dec. 2005. 65. Not-for-attribution interview, Nov. 2005. 66. Not-for-attribution interview, Nov. 2005.

212 tim büthe and helen v. milner

for each BIT.67 The sometimes vague but potentially sweeping provisions in BITs reflect the commitment to economically liberal policies in this broader sense. Anecdotal evidence suggests that developing country governments, too, have understood BITs as broad commitments to economically liberal policies. Salacuse (1990, p. 674) notes that developing countries have often sought them as “confidence-building measures . . . to improve the host country’s investment climate” and that in specific developing countries, “investor protection” has, in fact, often improved in the aftermath of a country signing BITs (Salacuse and Sullivan 2005, p. 674; see also Mahnkopf 2005, p. 130). Discussing why the Thai government had signed seven in one year, a recent article notes that “[BITs] signal to the business community worldwide and to [a country’s] own investors [the government’s] commitment to provide a predictable and stable legal framework for investors . . . and thereby boost FDI flows” (Business Day (Thailand), Feb. 22, 2000). Finally, investors seem to view BITs as a credible commitment to a broad range of economically liberal policies, and “codifying such commitments in a treaty” gives them greater “visibility” and “extra weight.”68 To be sure, individual investment decisions are highly idiosyncratic, driven in the first place, of course, by the existence of an investment opportunity. BITs can make a country only more attractive as an FDI host country if such opportunities exist, and the specific concerns of investors differ with the particular investment opportunities that they are considering. Nonetheless, our interviews lead us to believe that foreign investors are generally well aware of the BITs that have been signed by the government of the country that is the potential host for their investments. Any particular BIT (not just with the investor’s home country) may then be welcomed for a number of reasons, consistent with our broad interpretation of these agreements. As an investment consultant to many U.S. multinational corporations told us, BITs between the investor’s home country and the FDI host country reduce transaction costs because they contain as general policy commitments (i.e., as a matter of course) what potential foreign investors must otherwise negotiate as part of their individual contracts with the host governments, such as the right to repatriate profits or binding arbitration in the case of disputes. Moreover, these policy commitments are more credible because they are enshrined in an international treaty. As a German government official put it, based on many years of interactions with German foreign investors: “[O]f course, it’s a treaty, so changing it is more difficult than just changing a law.”69 He also

67. Note that such changes in policy during and because of the negotiations imply that even just prior or simultaneous changes in policies (as measured by the financial openness index and the good policy index in our statistical analyses) may in fact sometimes causally be connected to BITs. 68. Not-for-attribution interview, Nov. 2005. 69. Not-for-attribution interview, Nov. 2005.

bilateral investment treaties and foreign direct investment 213

noted that he had found corporate executives considering FDI to be generally well informed about the breadth of commitments that a developing country has undertaken in a given BIT. And in interviews, investors suggested that one of the reasons why they care about BITs signed between the host country and countries other than their own is that treaties are indicative of a country’s international engagement. Having signed more BITs means that the country has a greater stake in what the rest of the world thinks about it. This reasoning suggests that investors recognize the reputational logic discussed in Section B.3.b. The direct support for the hypothesized causal mechanism is mostly anecdotal and in that sense tentative: We have drawn on interviews with a number of government officials and senior managers in MNCs who have been involved in investment decisions, as well as investment advisors, but the information drawn from those interviews is not systematically linked to FDI outcomes. Nonetheless, the observed support for every major element of the hypothesized causal mechanism cumulatively gives us considerable confidence that the theoretical argument put forth in Section B indeed drives the correlational findings reported in Section C.

e. bit s , bites, and fdi: some conclusions We have examined the effect of bilateral investment treaties (BITs) on inward foreign direct investment flows (FDI) into developing countries (LDCs). We have argued that BITs generate information about LDC governments’ policy commitments, as well as information about governmental actions that violate those commitments, and that they facilitate imposing political and/or economic costs on such governments through intergovernmental or private enforcement mechanisms. As a consequence, foreign investors see BITs as credible commitments to economically liberal policies across a broad range of issues, which alleviates the key substantive concern of such investors about the political risks to FDI in developing countries. And because the informational benefits accrue to foreign investors in general, because the specific rights accrue to a broad range of foreign investors with often multiple economic nationalities, and because bilateral FDI data often do not record the real source of the invested capital or investment decision, we have argued that BITs should be expected to boost inward FDI in general, not just bilateral FDI. In fact, dyadic analyses may inappropriately pool observations that should be expected to exhibit a BIT effect with observations that should exhibit no such effect, thus biasing the findings toward insignificance. Statistical analyses have provided strong support for our argument. We find the predicted positive, statistically and substantively significant correlation between BITs and subsequent inward FDI into developing countries in a maximally comprehensive analysis of FDI flows into developing countries from 1970

214 tim büthe and helen v. milner

through 2000.70 This finding is robust to the inclusion of numerous control variables including measures of domestic policy choice, the use of several alternative estimation methods, and the exclusion of countries and even entire regions that have attracted a large amount of FDI in recent years. Qualitative analyses suggest that it is indeed the hypothesized causal mechanism that is driving these results, as we find support from interviews, internal documents, and secondary literature for our key substantive assumption and for all of the major elements of the causal mechanism. A recent study of BITs warned that BITs “may bite” (Hallward-Driemeier 2003) in that they impose real constraints on the ability of developing country governments to adapt their policies flexibly to their political and economic needs. Our theoretical argument suggests that this constraint—on the governments of the FDI host countries that sign them—is not an accidental by-product but intended by both sides. BITs are in that sense simply particular instances of the institutionalization of policy choices, which is often intended to “bind the hands” of policymakers or their successors (Moe 1990; 2005). In fact, we have argued and found that BITs work (i.e., attract FDI) precisely because they “bite.” Our finding that BITs indeed attract FDI does not constitute a normative endorsement. The flipside of our argument that BITs boost FDI inflows into developing countries because they “bite” is that such treaties indeed constrain policy choices or impose significant ex post costs if policies are chosen that violate the commitments institutionalized through these treaties. Such constraints may run counter to principles of democratic governance, though developing countries with democratic regimes appear no less likely to enter into BITs than nondemocratic ones.71 Especially due to the dispute settlement procedures, BITs constitute a kind of international delegation of governance, for which developing countries often pay a higher price than advanced industrialized countries in term of policy autonomy lost (e.g., Hawkins et al. 2006; Bradley and Kelley 2008; Büthe 2008). It is a cost that each developing country has to weigh against the benefits of increased FDI (and possibly other benefits). BITs certainly are not required for attracting FDI (Walter 2000, pp. 60ff), though the competitive dynamic (Elkins, Guzman and Simmons 2006; Jandhyala, Henisz and Mansfield 2006) may mean that retaining the status quo of no or few BITs might become increasingly costly over time.

70. In addition, we have found that BITs signed with large countries boost FDI further, without much diminishing the effet of the raw number of BITs. 71. In fact, there is a small positive correlation between measures of democracy (ACLP, Polity, or the inverted Freedom House index) and signing new BITs, though that correlation is very low (ranging from 0.12 to 0.20). If BITs increase FDI they might also contribute to changing the politics within the host countries, as Malesky (2008) showed that inward FDI changes the balance of power between the political center and periphery in FDI host countries, but such effects are beyond the scope of this chapter.

bilateral investment treaties and foreign direct investment 215

Finally, our analysis suggests several avenues for future research. First, the logic of our argument about BITs suggests that international institutions more generally might allow FDI host governments to make credible commitments to foreign investors. Trade agreements in particular might also boost FDI flows into developing countries. While most trade agreements lack provisions for a private enforcement mechanism that is a prominent feature of BITs, they specifically commit the signatories to liberal foreign economic policies that are particularly important for vertical FDI, and they contain provisions that should increase the amount of information about compliance and facilitate enforcement, so that policy commitments should be more credible when made via trade agreements than when they are just made unilaterally. An analysis of both BITs and trade agreements would contribute further to our understanding of the role of international institutions in the international political economy (see Martin and Simmons 1998). A proper discussion and analysis of the effect of trade agreements on FDI is beyond the scope of this chapter but promises to be a fruitful avenue for future research (see Büthe and Milner 2008). Since the logic of the argument suggests that BITs and trade agreements are partial substitutes for purposes of FDI, adding trade agreements to our models should reduce the estimated effect, but given the distinctive issues covered by these international institutions, they should also have independent effects. To give the reader a preliminary sense of the effect of including trade agreements in our models, we re-estimate Models 3 and 4 as Models 6 and 7, respectively, where we add three measures of trade agreements: the cumulative number of preferential trade agreements (PTAs) to which the FDI host country is a party (cumulative PTAs) and dichotomous measures for GATT membership and WTO membership.72 Table 6 shows the estimated coefficients for our measures of international institutions and domestic policy choice (the other control variables are omitted from the table, but were included in all estimations). We find indeed that these measures of trade agreements have a positive, substantively and statistically significant effect on inward FDI flows, except for GATT, which exhibits at most weak statistical significance—consistent with the relative weakness of GATT as an international institution. At the same time, BITs retain a significant, positive effect on FDI inflows even when we include trade agreements, though the estimated effect of BITs is reduced in these models (see Büthe and Milner 2008 for details).73

72. We here encode GATT and WTO separately because WTO involves stronger information-provision and enforcement mechanisms; using a single GATT/WTO indicator yields substantively the same results for BITs. 73. BITs are only weakly significant in the reduced-sample models that include the Good Policy Index.

216 tim büthe and helen v. milner table 6. trade agreements as alternative commitment mechanisms

Cumulative BITs

Model 3‡

Model 6‡

Model 4‡

Model 7‡

0.0421∗∗∗

0.321∗∗

0.0372∗∗

0.0284∗

(.0137)

(.143)

(.0174) ∗∗

Cumulative PTAs

–

0.176

GATT membership

–

0.662∗

(.0163)

–

0.187∗∗

–

0.304

(.0760)

(.0898) (.404)

(.379)

WTO membership

∗∗

–

0.520

0.373∗

–

(.235)

Financial openness index Good policy index

(.219)

0.148∗∗∗

0.0968∗

0.0901∗

0.0558

(.0502)

(.0515)

(.0470)

(.0506)

–

∗∗∗

–

0.278

(.0653)

Trade openness n N R2

∗∗∗

(.0778)

0.0145

0.0142

0.0195

0.0194∗∗∗

(.00541)

(.00552)

(.00566)

(.00570)

121 2499 +0.0820

82 1785 +0.1092

82 1785 +0.1226

121 2499 +0.0672

∗∗

0.216∗∗∗

∗∗∗

‡

Note: All regression models also included domestic political constraints, political instability, trade (openness), market size, economic development, GDP growth, and a constant. All variables de-trended, except “political instability,” which exhibited no significant trend. OLS within-estimates with Arellano (1987) robust clustered standard errors in parentheses. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01; two-tailed tests. Analyses cover inward FDI flows 1970–2000; all explanatory variables enter with a 1-year lag. All estimates in Stata 9.2. R2 not fully comparable across models when sample size changes.

Second, our empirical analyses have treated BITs as if they were all the same and therefore have estimated an average effect. Similarly, the analyses in Table 6 show the average effect of PTAs. Yet, while these international treaties have important common features, they also differ, and insofar as different elements of the treaties suggest more information provision or better/easier enforcement, stronger treaties might have a stronger effect on FDI. This suggests as one avenue for future research coding the specific provisions of BITs and PTAs, which would also contribute to the new research agenda on the design of international agreements and institutions (Koremenos, Lipson and Snidal 2001; Koremenos 2005, 2007; Axelrod 2008; Putnam and Shapiro 2007; Whytock 2005). Third, we have found no statistically significant effect of regime type (democracy) on FDI flows into developing countries, contrary to other recent studies (Feng 2001; Globerman and Shapiro 2003; Jensen 2003; 2006; Tures 2003;

bilateral investment treaties and foreign direct investment 217

though cf. Li and Resnick 2003; Biglaiser and DeRouen 2006; Blanton and Blanton 2006). Our finding is particularly surprising given that new, careful research on specific policies with which governments seek to encourage, discourage, or regulate foreign direct investment—from restrictions on capital flows and investment incentives to tax policy and performance requirements—finds that political competition leads more democratic countries to adopt policies that are more favorable to foreign investors (Dorobantu-Popa 2008). It may be, however, that the greater responsiveness of political leaders in democracies to domestic interests74 makes favorable policies in democracies particularly susceptible to the credible commitment problems that we took as the starting point of our analysis. If democracies indeed face greater commitment problems in the eyes of foreign investors, then democracy as such might not lead to greater inward FDI, but democracy might boost FDI conditional on international institutions like BITs and PTAs. Alternatively, international institutions may constitute a more effective constraint upon governments when domestic groups that benefit from them can push for government compliance with the country’s international obligations, suggesting that the effect of BITs or other international institutions may be conditional on democracy. In short, future research should consider the interaction between domestic and international institutions, which has received little attention in the field of International Political Economy (with a few exceptions, such as Büthe and Mattli 2009; Mattli and Büthe 2003; Snidal and Thompson 2003). The findings from our analysis, which has treated differences in domestic political institutions only as a factor to be controlled for, should therefore not be taken as evidence of the irrelevance of domestic political institutions for the international political economy but rather as suggestive of a new avenue for research on the interaction between domestic and international institutions. Such research is beyond this chapter, but might advance the growing literature showing that differences in domestic institutions explain a significant part of the variation in international outcomes, both in the governmental and non-governmental realm (see e.g., Milner and Rosendorff 1997; Milner 2006; Milner with Kubota 2005).

references Abbott, Kenneth W. and Duncan Snidal (2000). “Hard and soft law in international governance,” International Organization, 54, 3, pp. 421–456. Alenfeld, Justus (1971). Die Investitionsförderungsverträge der Bundesrepublik Deutschland (Frankfurt am Main: Athenäum Verlag). Arellano, M. (1987). “Computing robust standard errors for within-groups estimators,” Oxford Bulletin of Economics and Statistics, 49, 4 (November), pp. 431–434.

74. Foreign investors are not necessarily excluded from the political process in democracies but likely will be discounted relative to domestic interests in a democracy.

218 tim büthe and helen v. milner Arvelund, Erin and Jad Mouawad (2004). “Yukos sale deepens investor doubts,” New York Times, December 21, p. C8. Axelrod, Mark (2008). Saving Institutional Benefits: Path Dependence in International Law, Unpublished Ph.D. Dissertation (Durham: Department of Political Science, Duke University). Baldwin, David A. (1985). Economic Statecraft (Princeton: Princeton University Press). —— 1989 (1971). “The power of positive sanctions,” First published in World Politics, 24, 1, pp. 19–38. In Paradoxes of Power (New York: Basil Blackwell), pp. 58–81. Banks, Arthur S. (1999). Cross-National Time-Series Data Archive (Binghampton: Banner Software). Beck, Nathaniel and Jonathan N. Katz (1995). “What to do (and not to do) with timeseries-cross-section data,” American Political Science Review, 89, 3, pp. 634–647. Bergman, Mark S. (1983). “Bilateral investment protection treaties: an examination of the evolution and significance of the U.S. prototype treaty,” N.Y.U. Journal of International Law & Politics, 16, 1, pp. 1–43. Betts, Paul (2004). “Yukos fails to mar Russia’s attractions,” Financial Times, August 19, p. 16. Biglaiser, Glen and Karl DeRouen (2006). “Economic reforms and inflows of foreign direct investment in Latin America,” Latin American Research Review, 41, 1, pp. 51–75. Blanton, Shannon Lindsey and Robert G. Blanton (2006). “Human rights and foreign direct investment: a two-stage analysis,” Business and Society, 45, 4, pp. 464–485. Blonigen, Bruce A. (2005). “A review of the empirical literature on FDI determinants,” Atlantic Economic Journal, 33, 4, pp. 383–403. Blonigen, Bruce A. and Miao Wang (2005). “Inappropriate pooling of wealthy and poor countries in empirical FDI studies,” in Theodore H. Moran, Edward M. Graham and Magnus Blomström, eds., Does Foreign Direct Investment Promote Development? (Washington, D.C.: Institute for International Economics), pp. 221–244. Bradley, Curtis A. and Judith Kelley, eds. (2008). Law and Contemporary Problems, 71, 1 (Special Issue on The Law and Politics of International Delegation). Brunetti, Aymo, Gregory Kisunko and Beatrice Weder (1997). “Institutional obstacles to doing business: region-by-region results from a worldwide survey of the private sector,” World Bank Policy Research Paper, no. 1759 (Washington, D.C.: World Bank) Buckley, Neil (2005). “Aftershocks continue to keep foreign investors on edge,” Financial Times, May 17, p. 4. Burkhardt, Hans-Martin (1986). “Investment protection treaties: recent trends and prospects,” Aussenwirtschaft: Schweizerische Zeitschrift für internationale Wirtschaftsbeziehungen, 41, 1 (Special Issue on Promotion of Direct Investment in Developing), pp. 99–104. Burnside, Craig and David Dollar (2000). “Aid, policies, and growth,” American Economic Review, 90, 4, pp. 847–868. Büthe, Tim (2008). “The globalization of health and safety standards: Delegation of regulatory authority in the SPS-Agreement of 1994 Agreement Establishing the World Trade Organization,” Law and Contemporary Problems, 71, 1, pp. 219–255. Büthe, Tim and Walter Mattli (2009). Global Private Governance: The Politics of RuleMaking for Product and Financial Markets. (Unpublished manuscript, Duke and Oxford Universities). Büthe, Tim and Helen V. Milner (2008). “The politics of foreign direct investment into developing countries: Increasing FDI through trade agreements?” American Journal of Political Science, 52, 4, pp. 741–762.

bilateral investment treaties and foreign direct investment 219

Chang, Roberto, Linda Kaltani and Norman Loayza (2005). “Openness can be good for growth: the role of policy complementarities,” NBER Working Paper, no. 11787. Coase, Ronald H. (1937). “The nature of the firm,” Economica, 4, 16, pp. 386–405. —— (1982). How Should Economists Choose? (Washington, D.C.: American Enterprise Institute for Public Policy Research). Cortell, Andrew P. and James W. Davis (1996). “How do international institutions matter? The domestic impact of international rules and norms,” International Studies Quarterly, 40, 4, pp. 451–478. Dai, Xinyuan (2005). “Why comply? The domestic constituency mechanism,” International Organization, 59, 2, pp. 363–398. Davidson, Russell and James G. MacKinnon (1993). Estimation and Inference in Econometrics (New York: Oxford University Press). Dee, Philippa and Jyothi Gali (2003). “The trade and investment effects of preferential trading arrangements,” NBER Working Paper, no. 10160. Desai, Mihir A. and Alberto Moel (2006). “Czech mate: expropriation and investor protection in a converging world,” ECGI Working Paper, no. 62/2004. Dolzer, Rudolf and Margrete Stevens (1995). Bilateral Investment Treaties (The Hague: Martinus Nijhoff Publishers). Dorobantu-Popa, Sinziana (2008). Political Competition and the Regulation of Foreign Direct Investment. Doctoral dissertation in progress (Durham: Department of Political Science, Duke University). Downs, George W., David M. Rocke and Peter N. Barsoom (1996). “Is the good news about compliance good news about cooperation?” International Organization, 50, 3, pp. 379–406. Easterly, William, Ross Levine and David Roodman (2003). “New data, new doubts: revisiting ‘Aid, Policies, and Growth,’” World Bank Working Paper, no. 26 (Washington, D.C.: World Bank). Elkins, Zachary, Andrew T. Guzman and Beth A. Simmons (2006). “Competing for capital: the diffusion of bilateral investment treaties, 1960-2000,” International Organization, 60, 4, pp. 811–846. Feng, Yi (2001). “Political freedom, political instability, and policy uncertainty,” International Studies Quarterly, 45, 2, pp. 271–294. FIA (Foreign Investment Agency), Ministry of Planning and Investment of Vietnam (2005). The Impact of the U.S.-Vietnam Bilateral Trade Agreement on Overall and U.S. Foreign Direct Investment in Vietnam (Hanoi: National Political Publisher). Franck, Susan D. (2005a). “The legitimacy crisis in investment treaty arbitration: privatizing public international law through inconsistent decisions,” Fordham Law Review, 73, 4, pp. 1521–1625. —— (2005b). “The nature and enforcement of investor rights under investment treaties,” U.C. Davis Journal of International Law and Policy, 12, 1 (Fall 2005 Symposium Issue: Romancing the Investor, BIT by BIT), pp. 47–99. Frieden, Jeffry A. and Ronald Rogowski (1996). “The impact of the international economy on national policies: an analytical overview,” in Robert O. Keohane and Helen V. Milner, eds., Internationalization and Domestic Politics (New York: Cambridge University Press), pp. 25–47. FT, Financial Times (2004). “Yukos hammered, ” Financial Times, November 22, p. 16. Gallagher, Kevin P. and Melissa B. L. Birch (2006). “Do investment agreements attract investment? Evidence from Latin America,” Journal of World Investment and Trade, 7, 6, pp. 961–973.

220 tim büthe and helen v. milner Gallins, Glenn (1984). “Bilateral investment protection treaties,” Journal of Energy and Natural Resources Law, 2, 2, pp. 77–97. Globerman, Steven and Daniel Shapiro (2003). “Governance infrastructure and U.S. foreign direct investment,” Journal of International Business Studies, 34, 1, pp. 19–39. Goldhaber, Michael D. (2003). “Private practices: big arbitrations,” The American Lawyer, 25 (Supplement: Focus Europe). Goldstein, Judith, Miles Kahler, Robert O. Keohane, and Anne-Marie Slaughter, eds. (2001). Legalization and World Politics, (Cambridge, MA: MIT Press). Green, Peter S. “Czech republic pays $355 million to media concern,” New York Times, May 16 (late edition), p. W1. Grosse, Robert and Len J. Trevino (2005). “New institutional economics and FDI location in Central and Eastern Europe,” Management International Review, 45, 2, pp. 123–145. Guzman, Andrew T. (1998). “Why LDCs sign treaties that hurt them: explaining the popularity of bilateral investment treaties,” Virginia Journal of International Law, 38, 4, pp. 639–688. Haftel, Yoram Z. (2008). “The effect of U.S. BITs on FDI inflows to developing countries: signaling or credible commitment?” (Unpublished manuscript, University of Illinois, Chicago). Hallward-Driemeier, Mary (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and they could bite,” (Washington, D.C.: World Bank). Hanson, Gordon H., Raymond J. Mataloni, and Matthew J. Slaughter (2005). “Vertical Production Networks in Multinational Firms.” Review of Economics and Statistics 87, 4, pp. 664–678. Harms, Philipp and Heinrich W. Ursprung (2002). “Do civil and political repression really boost foreign direct investment?” Economic Inquiry, 40, 4, pp. 651–663. Hawkins, Darren G., David A. Lake, Daniel Nielson, and Michael J. Tierney, eds. (2006). Delegation and Agency in International Organizations (New York: Cambridge University Press). Henisz, Witold J. (2000). “The institutional environment for multinational investment,” Journal of Law, Economic and Organization, 16, 2, pp. 334–364. —— (2002). “The institutional environment for infrastructure investment,” Industrial and Corporate Change, 11, 2, pp. 355–389. Henisz, Witold J., and Jeffrey T. Macher (2004). “Firm- and country-level tradeoffs and contingencies in the evaluation of foreign investment: the semiconductor industry, 1994–2002,” Organization Science, 15, 5, pp. 537–554. Henisz, Witold J. and Bennet A. Zelner (2001). “The institutional environment for telecommunications investment,” Journal of Economics & Management Strategy, 10, 1, pp. 123–147. Hsiao, Cheng (2003). Analysis of Panel Data, 2nd edition (New York: Cambridge University Press). ICSID, Annual Report, various years. IMD, International Institute for Management Development (2001). World Competitiveness Yearbook, 2001 (Lausanne: IMD). Jack, Andrew (2004). “Two faces of Russia split the world’s investors,” Financial Times. August 25, p. 18. Jandhyala, Srividya, Witold J. Henisz, and Edward D. Mansfield (2006). “A bilateral analysis of bilateral investment treaties,” (Philadelphia: University of Pennsylvania).

bilateral investment treaties and foreign direct investment 221

Jensen, Nathan M. (2003). “Democratic governance and multinational corporations: Political regimes and inflows of foreign direct investment,” International Organization, 57, 3, pp. 587–616. —— (2006). Nation States and the Multinational Corporation: A Political Economy of Foreign Direct Investment (Princeton: Princeton University Press). Jun, Kwang W. and Harinder Singh (1996). “The determinants of foreign direct investment in developing countries,” Transnational Corporations, 5, 2, pp. 67–105. Kahler, Miles (1981). “Political regime and economic actors: the response of firms to the end of colonial rule,” World Politics, 33, 3, pp. 383–412. Karmin, Craig (2005). “Ahead of the tape,” Wall Street Journal, January 17, p. C1. Katzenstein, Suzanne (2008). “Home states and bilateral investment treaties in the 1970s and 1980s: The strategic appeal of private-actor standing,” Paper presented at the Midwest Political Science Association Meeting, Chicago, April. Kézdi, Gábor (2004). “Robust standard error estimation in fixed-effects panel models,” Hungarian Statistical Review, 9 (Special English edition), pp. 95–116. Koremenos, Barbara (2005). “Contracting around international uncertainty,” American Political Science Review, 99, 4, pp. 549–565. —— (2007). “If only half of international agreements have dispute resolution provisions, which half needs explaining?” Journal of Legal Studies, 36, 1. Koremenos, Barbara, Charles Lipson and Duncan Snidal, eds. (2001). The Rational Design of International Institutions (Special issue of International Organization, 55, 4). Krasner, Stephen D. (1991). “Global communications and national power: Life on the Pareto frontier,” World Politics, 43, 3, pp. 336–366. Kydland, Finn E. and Edward C. Prescott (1977). “Rules rather than discretion: the inconsistency of optimal plans,” Journal of Political Economy, 85, 3, pp. 473–491. Lang, Jeffrey. (1998). “Symposium on the international regulation of foreign direct investment: Keynote address,” Cornell International Law Journal, 31, 3, pp. 455–466. Li, Quan (2005). “Democratic institutions and expropriation of foreign direct investment,” Paper presented at the Conference on the Political Economy of Multinational Enterprises, Washington University, St. Louis, June. Li, Quan and Adam Resnick (2003). “Reversal of fortunes: Democratic institutions and foreign direct investment inflows to developing countries,” International Organization, 57, 1, pp. 175–211. Lipson, Charles (1991). “Why are some international agreements informal?” International Organization, 45, 4, pp. 495–538. Mahnkopf, Birgit. (2005). “Investition als Intervention: Wie interregionale und bilaterale Investitionsabkommen die Souveränität von Entwicklungsländern beschneiden,” Internationale Politik und Gesellschaft/International Politics and Society, 8, 1, pp. 121–141. Malesky, Edmund (2008). “Straight ahead on red: how foreign direct investment empowers subnational leaders,” Journal of Politics, 70, 1, pp. 1–23. Mann, Frederick A. (1982). “British treaties for the promotion and protection of investment,” British Yearbook of International Law, 52, pp. 241–254. Markusen, James R. and Keith E. Maskus (2004). “General-equilibrium approaches to the multinational firms: A review of theory and evidence,” in James Harrigan and Kwan Choi, eds., Handbook of International Trade (Massachusetts: Wiley-Blackwell), pp. 320–350. Martin, Lisa (2000). Democratic Commitments: Legislatures and International Cooperation (Princeton: Princeton University Press).

222 tim büthe and helen v. milner Martin, Lisa and Beth Simmons (1998). “Theories and empirical studies of international institutions,” International Organization, 52, 4, pp. 729–757. Mattli, Walter (2001). “Private justice in a global economy: From litigation to arbitration,” International Organization, 55, 4, pp. 919–947. Mattli, Walter and Tim Büthe (2003). “Setting international standards: Technological rationality or primacy of power?” World Politics, 56, 1, pp. 1–42. —— and —— (2005). “Global private governance: Lessons from a national model of setting standards in accounting,” Law and Contemporary Problems, 68, 3/4, pp. 225–262. McDonald, Ian and Henny Sender (2004). “Kremlin tests foreign holders with Yukos affair, ‘statist’ push; most stay put, but get set to exit,” Wall Street Journal, December 22, p. C1. Milner, Helen V. (1992). “International theories of cooperation among nations: Strengths and weaknesses,” World Politics, 44, 3, pp. 466–496. —— (1997). Interests, Institutions, and Information: Domestic Politics and International Relations (Princeton: Princeton University Press). —— (2006). “The digital divide: The role of political institutions in the spread of the internet,” Comparative Political Studies, 39, 2, pp. 176–199. Milner, Helen V. and Keiko Kubota (2005). “Why the move to free trade? Democracy and trade policy in the developing countries,” International Organization, 59, 1, pp. 107–143. Milner, Helen V. and B. Peter Rosendorff (1997). “Democratic politics and international trade negotiations: Elections and divided government as constraints on trade liberalization,” Journal of Conflict Resolution, 41, 1, pp. 117–146. Minor, Michael S. (1994). “The demise of expropriation as an instrument of LDC policy, 1980-1992,” Journal of International Business Studies, 25, 1, pp. 177–188. Moe, Terry M. (1990). “Political institutions: the neglected side of the story,” Journal of Law, Economics and Organization, 6, 3 (Special Issue on The Organization of Political Institutions), pp. 213–253. —— (2005). “Power and political institutions,” Perspectives on Politics, 3, 2, pp. 215–233. Morrow, James D. (1994). “Modeling the forms of international cooperation: Distribution versus information,” International Organization, 48, 3, pp. 387–423. Neumayer, Eric and Laura Spess (2005). “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33, 10, pp. 1567–1585. Oneal, John R (1994). “The affinity of foreign investors for authoritarian regimes,” Political Research Quarterly, 47, 3, pp. 565–588. Peinhardt, Clint and Todd Allee (2007). “The effect of international investment arbitration on FDI,” Presentation at the 2nd Annual Meeting of the International Political Economy Society, November. Peterson, Luke (2001). “Changing investment litigation: ‘Bit by BIT’” ICTSD Bridges, 5, 4, pp. 11–12. Pinto, Pablo M. and Santiago M. Pinto (2008). “The politics of investment: Partisanship and the sectoral allocation of foreign direct investment,” Economics and Politics, 20, 2, pp. 216–254. Piper, Don L (1979). “New directions in the protection of American-owned property abroad,” International Trade Law Journal, 4, 2, pp. 315–348. Putnam, Tonya L. and Jacob N. Shapiro (2007). “Do treaties matter to citizens’ willingness to punish foreign rights abusers?” (Unpublished manuscript, Columbia and Princeton Universities).

bilateral investment treaties and foreign direct investment 223

Rogers, William D. (2000). “Emergence of ICSID as the most significant forum for submission of bilateral investment treaties disputes,” Remarks delivered at the Inter-American Development Bank Conference on Commercial Alternative Dispute Resolution in the XXI Century: The Road Ahead for Latin America and the Caribbean, October 26–27 (on file with the authors). Salacuse, Jeswald W. (1990). “BIT by BIT: The growth of bilateral investment treaties and their impact on foreign investment in developing countries,” The International Lawyer, 24, 3, pp. 655–675. Salacuse, Jeswald W. and Nicholas P. Sullivan (2005). “Do BITs really work? An evaluation of bilateral investment treaties and their grand bargain,” Harvard International Law Journal, 46, 1, pp. 67–130. Schneider, Friedrich and Bruno S. Frey (1985). “Economic and political determinants of foreign direct investment,” World Development, 13, 2, pp. 161–175. Simmons, Beth A. (2000). “International law and state behavior: Commitment and compliance in international monetary affairs,” American Political Science Review, 94, 4, pp. 819–835. Snidal, Duncan and Alexander Thompson (2003). “International commitments and domestic politics: institutions and actors at two levels,” in Daniel W. Drezner, ed., Locating the Proper Authorities: The Interaction of Domestic and International Institutions (Ann Arbor: University of Michigan Press), pp. 197–230. Sornarajah, M. (1986). “State responsibility and bilateral investment treaties,” Journal of International Trade Law, 20, 1, pp. 79–98. Tarzi, Shah M. (1991). “Third world governments and multinational corporations: Dynamics of host’s bargaining power,” International Relations, 10, 3, pp. 237–249. Tobin, Jennifer and Susan Rose-Ackerman (2005). “Foreign direct investment and the business environment in developing countries: The impact of bilateral investment treaties,” Yale Law & Economics Research Paper, No. 293, May 2. Tomz, Michael. (2007). Reputation and International Cooperation: Sovereign Debt Across Three Centuries. (Princeton: Princeton University Press). Trevino, Len J., John D. Daniels and Harvey Arbeláez (2002). “Market reform and FDI in Latin America: An empirical investigation,” Transnational Corporations, 11, 1, pp. 29–48. Tures, John A. (2003). “The impact of instability and institutions on U.S. foreign direct investment in developing areas,” Conflict, Security and Development, 3, 2, pp. 163–183. UN, Centre on Transnational Corporations (1988). Bilateral Investment Treaties (New York-Geneva: United Nations). —— (1992). Determinants of Foreign Direct Investment: A Survey of the Evidence (New York: United Nations). UNCTAD, United Nations Conference on Trade and Development (2000). Bilateral Investment Treaties, 1959-1999 (New York-Geneva: United Nations). —— (2003a). Handbook of Statistics 2003 (New York-Geneva: United Nations). —— (2003b). World Investment Report 2003: FDI Policies for Development (New YorkGeneva: United Nations). —— (2007). “Recent developments in international investment agreements (2006– June 2007).” IIA Monitor, No.3 (UNCTAD/WEB/ITE/IIA/2007/6). U.S. Chamber of Commerce (2000). “What goes into a U.S. company’s decision to invest overseas? . . . or . . . the 12 commandments for international investors,” (on file with the authors).

224 tim büthe and helen v. milner Vandevelde, Kenneth J. (1988). “The bilateral investment treaty program of the United States,” Cornell International Law Journal, 21, 2, pp. 201–276. Vandevelde, Kenneth J., Victoria Aranda and Zbigniew Zimny (1998). Bilateral Investment Treaties in the Mid-1990s (New York—Geneva: UNCTAD). Vernon, Raymond (1971). Sovereignty at Bay: The Multinational Spread of U.S. Enterprises. (New York: Basic Books). Von Stein, Jana. (2005). “Do treaties constrain or screen? Selection bias and treaty compliance,” American Political Science Review, 99, 4, pp. 611–622. Voss, Jürgen. (1981). “The protection of European private investment in developing countries: An approach toward a concept for a European policy on foreign investment,” Common Market Law Review, 18, pp. 363–395. Walter, Andrew (2000). “Globalization and policy convergence: The case of direct investment rules,” in Richard A. Higgott, Geoffrey R. D. Underhill and Andreas Bieler, eds., Non-State Actors and Authority in the Global System (London: Routledge), pp. 51–73. Wawro, Gregory and Ida Kristensen (2007). “On the use of fixed effects estimators for time-series cross-section data,” (New York: Columbia University). Whytock, Christopher A. (2005). “A rational design theory of transgovernmentalism: The case of E.U.-U.S. merger review cooperation,” Boston University International Law Journal, 23, 1, pp. 1–53. Williamson, Oliver E. (1985). The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting (New York: The Free Press). Wooldridge, Jeffrey M. (2002). Econometric Analysis of Cross Section and Panel Data. (Cambridge, MA: MIT Press). Yackee, Jason Webb (2005). “Are BITs such a bright idea? Exploring the ideational basis of investment treaty enthusiasm,” U.C. Davis Journal of International Law and Policy, 12, 1, (Fall 2005 Symposium Issue: Romancing the Investor, BIT by BIT), pp. 195–224. Yarbrough, Beth V. and Robert M. Yarbrough (1990). “International institutions and the new economics of organization,” International Organization, 44, 2, pp. 235–259.

7. do bilateral investment treaties increase foreign direct investment to developing countries?∗ eric neumayer and laura spess introduction Developing countries sign bilateral investment treaties (BITs) in order to attract more foreign direct investment (FDI). In recent decades BITs have become “the most important international legal mechanism for the encouragement and governance” of FDI (Elkins, Guzman, and Simmons 2004, p. 0). The preambles of the thousands of existing BITs state that the purpose of BITs is to promote the flow of FDI and, undoubtedly, BITs are so popular because policy makers in developing countries believe that signing them will increase FDI. But do these treaties fulfill their stated purpose and attract more FDI to developing countries that submit to the obligations of a BIT? Despite the large and increasing number of BITs concluded, there exists very little evidence answering this question. Most existing scholarship, typically written with a legal perspective, simply restricts itself to an analysis of the BIT practice of one country or certain similar provisions in a range of BITs (Vandevelde 1996, p. 545). This omission is strange given that the question is of great importance to developing countries. They invest time and other scarce resources to negotiate, conclude, sign and ratify BITs. Such treaties represent a non-trivial interference with the host countries’ sovereignty as they provide protections to foreign investors that are enforceable via binding investor-to-state dispute settlement. While the motivations driving developing countries to incur these costs may be varied (see Guzman 1998; Elkins, Guzman, and Simmons 2004; Neumayer 2006), the costs might be justified if the ultimate outcome is an increase in the inward flow of FDI.1 But is this what actually occurs? In the absence of hard, quantitative evidence, some observers have been rather pessimistic toward the effect of BITs on FDI location. Sornarajah (1986, p. 82), for example, suggests that “in reality attracting foreign investment depends more on the political and economic climate for its existence rather than

∗ This chapter was reprinted with permission from World Development. The chapter was originally published as “Do bilateral investment treaties increase foreign direct investment to developing countries?” 33 World Development 1567 (2005). 1. For the purpose of this chapter, we presume that a higher FDI inflow is beneficial to the host nation. This presumption can of course be contested (De Soysa and Oneal 1999).

226 eric neumayer and laura spess

on the creation of a legal structure for its protection.” An expert group meeting sponsored by the United Nations Conference on Trade and Development (UNCTAD) in 1997 reportedly held a similar position (Raghavan 1997). Supportive of this view is the fact that some major hosts of FDI like Brazil or Mexico for a long time were reluctant to sign BITs. As UNCTAD (1998, p. 141) has put it in a review of BITs from almost a decade ago: “There are many examples of countries with large FDI inflows and few, if any, BITs.” And yet, most developing countries have signed a great many BITs by now. Is there evidence that those that have signed more BITs have also managed to attract more FDI? Two studies analyze this issue over the period from 1980 to 2000 (HallwardDriemeier 2003; Tobin and Rose-Ackerman 2005) and one over the period from 1991 to 2000 (Salacuse and Sullivan 2005). The first study by HallwardDriemeier (2003) does not find any statistically significant effect. The second study by Tobin and Rose-Ackerman (2005) finds a negative effect at high levels of risk and a positive effect only at low levels of risk, with the majority of developing countries falling into the high-risk category. The third study by Salacuse and Sullivan (2005) finds a positive effect only for United States BITs, but not for BITs from other countries of the Organisation for Economic Co-operation and Development (OECD). The existing evidence goes against expectation and would suggest that the enormous amount of effort developing countries have spent on BITs has basically been wasted. One of the problems of existing studies is that they infer results from a rather restricted sample of countries (31 and 63, respectively) or are based on cross-sectional regressions. In contrast, we employ a much larger panel over the period from 1970 to 2001, covering up to 119 countries. Importantly, we find a positive effect of BITs on FDI inflows that is consistent and robust across various model specifications. The effect is sometimes conditional on institutional quality, but is always positive and statistically significantly different from zero at all levels of institutional quality. To our knowledge, we provide the first hard evidence that there is a payoff to developing countries’ willingness to incur the costs of negotiating BITs and to succumb to the restrictions on sovereignty contained therein. Having demonstrated that BITs successfully increase the flow of FDI coming to a country, we also address the important question of whether BITs function as substitutes for or complements to good institutional quality. Naturally, one would expect them to be substitutes—that is, they provide security and certain standards of treatment to foreign investors where domestic institutions fail to deliver the same security and standards. However, some, like Hallward-Driemeier (2003), argue that BITs might only be seen as credible in an environment of good institutional quality. This would imply that BITs are most effective in countries where they are least needed. Our results provide some limited evidence that BITs might function as substitutes for good institutional quality, which would suggest that they are most effective where such quality is low and that they are most successful where they are needed most. However, this result is not robust to different specifications of institutional quality.

do bilateral investment treaties increase foreign direct investment 227

This chapter is structured as follows. Section A below briefly describes the well-known fact of increasing importance of foreign investment to developing countries, illustrates the growth of BITs, and analyzes the role of their main provisions for the promotion of FDI. Section B then reviews the three existing empirical studies of the role of BITs in promoting FDI and discuss their shortcomings, which we aim to overcome in our own analysis. After presenting our research design (section C), we report results (section D) and test their sensitivity to important changes in model specification (section E). The conclusion summarizes results and notes their implications. A. BITs and FDI The flow of FDI has dramatically increased in the past several decades to become a major force in the worldwide allocation of funds and technology. Prior to 1970, world trade generally grew at a pace greater than that of FDI, but in the decades since then the flow of FDI has grown at more than twice the pace of the growth of worldwide exports. By the early 1990s, the sales of worldwide exports would be eclipsed by the sales of foreign affiliates of multinational firms (Dunning 1998). Not only has the flow of FDI increased worldwide, but the importance of FDI as a source of funds to developing countries in particular has also significantly increased. Private international flows of financial resources have become increasingly important to developing countries. In the 1980s tight budgets, the debt crisis and an overall decreased interest in providing traditional development aid led to a decline in official development assistance from the developed world. When capital flows to developing nations began to rise again in the latter part of that decade, the flows would increasingly be composed of FDI (Zebregs 1998). Only very recently have aid flows slightly increased again in the wake of the so-called Monterrey Consensus. However, in 2003 FDI was the largest component of the net resource flows to developing countries, and this is bound to remain the case for some time to come (UNCTAD 2003a). Although the developed countries remain both the dominating source and the major recipient of FDI, their dominance has decreased over time, with developing countries in 2003 receiving almost 31% of FDI as opposed to only about 20% in the 1980s. Indeed, FDI inflows per unit of GDP are much higher in many developing countries than in developed ones (UNCTAD 2004). It was during this same period that BITs were introduced and eventually proliferated. In light of the importance of FDI, particularly to developing nations, the extent to which these two phenomena are causally related warrants careful scrutiny. The first BITs appeared at the end of the 1950s. Some trace their history back to the treaties of friendship, commerce, and navigation concluded by the United States over centuries (Salacuse 1990). The friendship, commerce, and navigation treaties had the expansion of international trade and the improvement of U.S. foreign relations as their prime purpose, even though some investment provisions were later added (Guzman 1998). BITs, on the other hand, are more clearly focused on foreign investment protection. Germany, having lost almost all of its

228 eric neumayer and laura spess figure 1. bit s signed per year and cumulative bit s worldwide, 1960–2003 2500

Number of BITs

2000

1500

1000

500

19

60 19 63 19 66 19 69 19 72 19 75 19 78 19 81 19 84 19 87 19 90 19 93 19 96 19 99 20 02

0

Year BITs Signed Per Year

BITs Signed Cumulative

Source: UNCTAD (2003b)

foreign investment during the Second World War, signed the very first BIT with Pakistan in 1959. After that, it took almost two decades before BITs gained momentum. By the end of the 1960s there were 75 treaties, with this number rising to 167 by the end of the 1970s and to 389 by the end of the 1980s. The number of BITs worldwide began to grow rapidly in the 1990s, and by 2002 there would be 2,181 BITs worldwide (UNCTAD 2003a). Figure 1 shows the number of BITs signed per year and the cumulative number of BITs worldwide. In order to explain the popularity of bilateral investment treaties it is necessary to understand how they fit into the larger regime of state-foreign investor relations. Prior to the advent of BITs, the only protection for foreign investors was the customary international legal rule of minimum standard of treatment and the so-called Hull Rule. The minimum standard of treatment rule provides only very minimal protection, as the name already suggests, while the Hull Rule dealt exclusively with cases of expropriation and therefore provided no general protection against discriminatory treatment. The Hull Rule grew out of a dispute between Mexico and the United States in the 1930s over properties expropriated by the Government of Mexico. In one of a series of diplomatic notes to the Mexican Minister of Foreign Affairs, the U.S. Secretary of State Cordell Hull stated that “no government is entitled to expropriate private property, for whatever purpose,

do bilateral investment treaties increase foreign direct investment 229

without provision for prompt, adequate, and effective payment therefor” (Guzman 1998, p. 645). Subsequently, the rule of “prompt, adequate, and effective” compensation would be the standard known as the Hull Rule. However, it is disputed whether the Hull Rule represents customary international law. Developing countries challenged its validity as part of their demands for a New International Economic Order with some success: Resolution 1803 of the United Nations General Assembly merely requires “appropriate compensation” for expropriation (Ginsburg 2004). Guzman (1998, p.641) suggests that by the mid-1970s “the Hull Rule had ceased to be a rule of customary international law”, if ever it had been one. The fact that there were several spectacular expropriations in the 1960s and 1970s taking place without what investors regarded as adequate compensation supports this view.2 This raises doubt as to whether the Hull Rule ever represented customary international law, for which conforming state practice is a requirement. Surprisingly, even as many developing nations resisted the Hull Rule, many of the same countries began to sign on to BITs that incorporated similar and indeed more far-reaching provisions. Guzman (1998) and Elkins, Guzman, and Simmons (2004) suggest that this seemingly contradictory behavior is explained by the increased competition among developing countries for FDI from developed countries. Collectively, developing countries would be better off refusing to sign any BIT and retaining as much control over their assets as possible, which explains their resistance to multilateral investment treaties at fora, such as UNCTAD, where they can collectively express and organize their interests. In a classic example of the prisoner’s dilemma, however, the individual country benefits from being able to provide credible commitments to investors. In the context of limited multilateral or customary protection for investors, the individual country gains a competitive advantage as an investment location by submitting to a BIT. Further, when a less developed country’s neighbor or economic competitor signs such an agreement, in order to remain competitive they must sign one as well. Somewhat at odds with this explanation is the fact that the latest trend is for developing countries to sign BITs among themselves. This has been a rather recent development, however, and the vast majority of existing BITs are concluded between a developed and a developing country. The basic provisions of a bilateral investment treaty typically guarantee certain standards of treatment for the foreign investor (see Dolzer and Stevens 1995; UNCTAD 1998). By entering into a BIT, signatories agree to grant certain relative standards of treatment such as national treatment (foreign investors may not be treated any worse than national investors, but may be treated better and, in fact, often are) and most-favored nation treatment (privileges granted to one

2. Expropriations without prompt, adequate and effective compensation also took place in many Communist countries after the Second World War.

230 eric neumayer and laura spess

foreign investor must be granted to all foreign investors). They also agree to guarantee certain absolute standards of treatment, such as fair and equitable treatment for foreign investors in accordance with international standards after the investment has taken place. BITs typically ban discriminatory treatment against foreign investors, and include guarantees of compensation for expropriated property or funds and free transfer and repatriation of capital and profits. Further, the BIT parties agree to submit to binding dispute settlement should a dispute concerning these provisions arise (UNCTAD 1998). Ostensibly, these provisions should secure some of the basic requirements for credible protection of property and contract rights that foreign investors look for in host countries. They should also protect foreign investors against political and other risks highly prevalent in many developing countries. Far from being treated neutrally, foreign investors are often granted higher security and better treatment than domestic investors (Vandevelde 1998). The basic provisions of BITs are all direct answers to the fundamental “hold-up” or “dynamic inconsistency” problem that faces developing nations attempting to attract FDI. The dynamic inconsistency problem arises from the fact that although host countries have an incentive to promise fair and equitable treatment beforehand in order to attract foreign investment, once that investment is established and investors have sunk significant costs the host country’s incentive is to exploit or even expropriate the assets of foreign investors. Even those host countries that are willing to forgo taking advantage in these circumstances will find it very difficult to credibly commit to their position. Many developing countries have adopted domestic legal changes over the last decade or so with a view toward encouraging a greater FDI inflow (UNCTAD 2004). However, these domestic legal rules cannot substitute for the commitment device offered by entering into a legally binding bilateral treaty. BITs, and their binding investor-to-state dispute settlement provisions in particular, are meant to overcome the dilemma facing host countries that are willing to denounce exploitation of foreign investors after the investment has already been undertaken. Interestingly, at the same time as BITs flourished in the 1980s and 1990s, outright expropriations of foreign investors, which were common during the 1960s and 1970s, practically ceased to take place (Minor 1994). The extent of interference with domestic regulatory sovereignty to which developing countries succumb in signing BITs is enormous. In fact, virtually any public policy regulation can potentially be challenged through the dispute settlement mechanism as long as it affects foreign investors. Often, foreign investors are not required to first exhaust domestic legal remedies and can thus bypass or avoid national legal systems, reaching straight for international arbitration—where they can freely choose one of the three panelists, their consensus is needed for one other panelist, and they can expect that the rules laid out in the BITs are fully applied (Peterson 2004). This contrasts with domestic courts, where investors have no say in the composition of judges and where domestic

do bilateral investment treaties increase foreign direct investment 231

rules might trump BIT provisions. BITs have been criticized for not conforming to a truly liberal economic model, since they fail to ban distorting government policies such as protective tariffs or tax incentives for foreign investors (Vandevelde 2000). However, even critics such as Vandevelde (2000, p. 499) admit that “BITs seriously restrict the ability of host states to regulate foreign investment.”3 In concluding BITs, developing countries are therefore “trading sovereignty for credibility” (Elkins, Guzman, and Simmons 2004, p. 4). BITs are therefore an important instrument of protection to foreign investors, for which there is currently not much legal alternative. Only a few regional free trade agreements contain investment protection provisions like those of the North American Free Trade Agreement (NAFTA). The World Trade Organization’s (WTO) Agreement on Trade-Related Investment Measures (TRIMs Agreement) imposes only rudimentary disciplines on the regulation of foreign investment that are much less comprehensive than, and fall far short of, provisions contained in BITs. Of course, not all BITs are identical in their provisions. Certain developed country investors like the United States insist on some limited rights of their investors to establish investment in host countries in the first place, whereas investor’s rights in most BITs are restricted to fair and equitable treatment after the investment has already taken place and provide no right of entrance. United States BITs often prohibit certain performance requirements, such as local content, export and employment requirements, beyond the requirements contained in the WTO’s TRIMs Agreement, whereas BIT programs of other developed countries do not contain such provisions (Vandevelde 1998). Conversely, some non-developed countries such as China and Eastern European countries have successfully managed to restrict the compulsory dispute settlement provisions to disputes concerning expropriation or the compensation thereof (Peters 1996, p. 107). However, by and large BITs tend to be rather similar in their provisions. BITs are also unlikely to be identical in their effect on incoming FDI flows. In principle, BIT provisions only protect investors from the signatory states to whom binding commitments are made.4 One would therefore expect that signing a BIT with a major capital exporter such as Germany or the United States 3. Developed countries tried to extend these restrictions in the context of a multilateral agreement on investment (MAI) in the mid-1990s, by banning pre-investment restrictions and performance requirements and by extending the measures of expropriation requiring compensation to indirect and de facto (as opposed to direct) expropriation. Most developing countries were rather happy when these negotiations imploded under the mounting critique of civil society and internal differences among developed countries (Neumayer 2001). Many developed countries were also not too keen on granting the same liberalized access to investors from other developed countries that they normally expect their own investors to receive from developing countries. 4. Depending on the BIT, this can refer to the country of the company’s incorporation, seat, registered office, or principal place of business, or to the nationality of the individuals

232 eric neumayer and laura spess

has a larger impact on FDI inflows than signing a BIT with minor capital exporters such as New Zealand or Portugal. However, the signing of BITs sends out a signal to potential investors that the developing country is generally serious about the protection of foreign investment. The encouragement of FDI flows therefore need not be restricted to investors from developed countries that are BIT partners of the developing country. Instead, BITs can have positive spill-over effects. How important is the signaling effect, which benefits investors from all countries, compared to the commitment effect, which only relates to investors from BIT partner countries, is difficult to say. Our research design, described in detail below, does not restrict the effect of BIT signature on FDI to investment from partner countries, and accounts for differences in the size of potential FDI to which the developing country has made binding commitments by weighting BITs with the relative importance of the developed country partner as a capital exporter. B. Review of Other Studies It is most astonishing that despite the rising number of BITs, there are only three other serious studies examining the effect of such treaties on the location of FDI.5 The first study has been undertaken by Hallward-Driemeier (2003), looking at the bilateral flow of FDI from 20 OECD countries to 31 developing countries over the period from 1980 to 2000. Her research design is dyadic, consisting of up to 537 country pairs. Using fixed-effects estimations, she finds that the existence of a BIT between two countries does not increase the flow of FDI from the developed to the developing signatory country. This is true whether the dependent variable is measured as absolute flows, flows divided by host country’s GDP or the share of the source countries’ FDI outflow. Interacting the BIT variable with various measures of institutional quality, she finds a positive coefficient of the interaction term that is often statistically significant. This would suggest that, contrary to theoretical expectations, BITs are complements to good institutional quality and therefore do not perform their original function, namely to provide guarantees to foreign investors in the absence of good domestic institutional quality. In the second study, Tobin and Rose-Ackerman (2005) analyze the impact of BITs on general non-dyadic FDI inflows, also in a panel from 1980 to 2000, but with data averaged over five-year periods, covering 63 countries. While both studies draw upon data provided by the International Country Risk Guide (ICRG), Hallward-Driemeier (2003) uses individual components of institutional who have control over, or a substantial interest in, the investing company (Salacuse and Sullivan 2005, 82). 5. A fourth study is provided by UNCTAD (1998). However, it is based on a purely cross-sectional stepwise regression analysis with an unspecified number of observations from 1995. Not surprisingly, such ‘garbage can’ modelling leads to inconclusive results.

do bilateral investment treaties increase foreign direct investment 233

quality, whereas Tobin and Rose-Ackerman (2005) use the aggregate political risk measure, which includes many more components than institutional quality, including some that are not directly related to political risk (such as, among others, religious and ethnic tensions, armed conflict and socio-economic conditions such as unemployment and poverty). In a fixed-effects model, Tobin and Rose-Ackerman find that a higher number of BITs either in total or signed with a high-income country lowers the FDI a country receives as a share of global FDI flows at high levels of risk and raises the FDI only at low levels of risk. In an additional dyadic analysis of 54 countries, they fail to find any statistically significant effect of BITs signed with the United States on FDI flows from the United States to developing countries, either conditionally on the level of political risk or unconditionally. The third study provides three cross-sectional analyses of FDI inflows to up to 99 developing countries in the years 1998, 1999 and 2000, respectively, as well as a fixed-effects estimation of the bilateral flow of FDI from the United States to 31 developing countries over the period from 1991 to 2000. Salacuse and Sullivan (2005) find the signature of a BIT with the United States to be associated with higher FDI inflows in both types of estimations, whereas the number of BITs with other OECD countries is always statistically insignificant. The three studies suffer from a number of shortcomings that we try to improve on in our own study. Hallward-Driemeier’s (2003) model presumes that a BIT will only have an effect on the flow of FDI from one developed country, namely the signatory, to the developing country. However, this presumption neglects the signaling effect of BITs (Elkins, Guzman, and Simmons 2004, p. 21). As pointed out in the preceding section, in concluding a BIT, the developing country explicitly commits only to protect the FDI from the signatory developed country, but also implicitly signals its willingness to protect all foreign investment. There are therefore likely to be positive spill-over effects from signing a BIT. HallwardDriemeier’s modeling cannot capture the potential of BITs to attract more FDI from other developed non-signatory countries as well, and consequently may underestimate the effect that signing a BIT has on the inward flow of FDI. In addition to not capturing this potentially important spill-over effect, the dyadic design also has another major disadvantage. Data on bilateral FDI flows are very sparse, and consequently the size of Hallward-Driemeier’s sample is significantly limited by this choice. A sample of 31 developing countries is anything but representative. Similar arguments apply to Salacuse and Sullivan’s (2005) fixed-effects analysis.6 Our own study draws from a much larger and more representative sample.

6. Their study also suffers from the absence of year-specific time dummies controlling for aggregate annual changes in U.S. FDI outflows, which could mean that the results are spurious.

234 eric neumayer and laura spess

Where Tobin and Rose-Ackerman (2005) do not use a dyadic research design, the paucity of bilateral FDI flow data does not impose a binding constraint on sample size. Nevertheless, for no clear reason their sample consists of only 63 countries. In comparison, our own sample is both deeper and wider. It covers the period from 1970, the first year for which UNCTAD provides FDI data, to 2001, the last year for which we have available data. It also covers up to 119 developing countries, which amounts to a much more representative sample. The countries included in our sample are listed in Appendix 1. Salacuse and Sullivan’s (2005) cross-sectional analysis also has the advantage of a large sample size. However, by definition this type of analysis cannot control for country-specific unobserved heterogeneity, which is likely to be important, nor does it exploit the full information available from looking at FDI flows over a longer time period. C. Research design 1. Dependent variable As our main measure of FDI attractiveness, we use the absolute amount of FDI going to a developing country, converted to constant U.S. dollar of 1996 with the help of the U.S. GDP deflator. Our definition of a developing country follows World Bank classification. We use absolute FDI flows because if one were to use FDI inflow as a percentage of host country’s GDP instead, the measure would capture changes in the relative importance of foreign investment to the host country, but not changes in inflows directly. Quite possibly, the worldwide increase in the rate of the conclusion of BITs is partly responsible for the increase in overall FDI going to developing countries. However, there is always the danger that one may find a statistically significant relationship between two upward trending variables that is spurious. We deal with this potential problem in two ways. First, we employ year dummies to account for any year-to-year variation in total FDI flows unaccounted for by our explanatory variables, which should mitigate potential spuriousness of any significant results. Second, as an alternate dependent variable to the absolute amount of FDI we use the FDI inflow a country receives relative to the sum of FDI going to developing countries. Since this variable is not trending over time no year-specific time dummies are included in these sets of estimations. Taking FDI inflow as a share of developing country FDI as the dependent variable captures the relative attractiveness of developing countries as hosts for FDI and explicitly allows for competition amongst them for a fixed-sized cake of FDI to be divided.7 Ideally, one would like to disaggregate FDI flows according to economic sectors. Unfortunately, no comprehensive information is available for a large panel of countries.

7. If we were to take the share a developing country receives relative to the sum of global FDI instead, then results are practically identical.

do bilateral investment treaties increase foreign direct investment 235

We take the natural log of the dependent variable to reduce the skewness of its distribution, thus increasing the model fit substantially. To do so, we need to recode a small number of negative FDI flows. Negative FDI flows essentially imply “instances of reverse investment or disinvestment” (UNCTAD 2001, p. 292). In our analysis we set negative FDI flows equal to positive FDI flows of one U.S. dollar. If instead one were to discard all negative flows then results are hardly affected. 2. Explanatory variables Our main explanatory variable is the cumulative number of BITs a developing country has signed with OECD countries, weighted by the share of outward FDI flow the OECD country accounts for relative to total world outward FDI flow.8 The weighting is used to account for differences in the size of potential FDI for which a developing country makes protection commitments via signing a BIT. We exclude BITs signed between developing countries since FDI flows between developing countries are rare. Tobin and Rose-Ackerman (2005) do not weight the cumulative BIT variable by the share of outward FDI flow of the developed country partner, but our results are similar if we include the unweighted BIT variable instead.9 They also take the natural log of the cumulative BIT variable. We keep the variable in its level form, not least because the log of zero (BITs) is undefined. Our control variables are very similar to the ones used by Hallward-Driemeier (2003) and Tobin and Rose-Ackerman (2005). They are also among the ones more consistently found to be determinants of FDI flows (Chakrabarti 2001). We include the natural log of per capita income, the log of total population size and the economic growth rate as indicators of market size and market potential (data from World Bank 2003).10 Developing countries that have concluded a free trade agreement with a developed country might receive more FDI as it is easier to export goods back into the developed or other countries. Such agreements sometimes also contain provisions on policies that might be beneficial to foreign investors. We account for this with a dummy variable indicating whether a country is a member of the World Trade Organization as well as a variable counting the number of bilateral trade agreements a developing country has concluded with the United States, the European Community/European Union or Japan, based

8. In future research we would like to analyze BITs in more detail. Not all BITs are the same and one would like to know whether it is certain elements in BITs that matter for FDI location more than others. 9. This is unsurprising given the high correlation between the weighted and unweighted cumulative count of BITs, which follows from the fact that by and large minor capital-exporting developed countries have signed few BITs, whereas the opposite is the case for major capital exporters. 10. Note that replacing logged population with the log of total GDP leads to identical results, which follows from the fact that the two variables plus the log of GDP per capita are not independent of each other.

236 eric neumayer and laura spess

on information from the WTO (2004) and European Union (EU; 2004).11 The inflation rate is a proxy variable for macroeconomic stability. In sensitivity analysis we also included trade openness and the secondary enrollment ratio, but these two variables are not included in the main analysis due to loss of observations following their inclusion. Data are taken from World Bank (2003a). We employ a measure of natural resource intensity to control for the fact that, all other things equal, large natural resources are a major attractor to foreign investors. Our measure is equal to the sum of rents from mineral resource and fossil fuel energy depletion divided by gross national income, as reported in World Bank (2003b). Rents are estimated as (P–AC)∗R, that is, as price minus average cost multiplied by the amount of resource extracted, an amount known as total Hotelling rent in the natural resource economics literature. There is a long tradition of studies analyzing the effect of political stability and institutional quality on FDI inflows (see, for example, Schneider and Frey 1985; Alesina and Perotti 1996; Wheeler and Mody 2000; Perry 2000; Globerman and Shapiro 2002). We include five different measures of institutional quality in separate estimations together with interaction effects with the BIT variable. First, we use the political constraints (POLCON) index developed by Henisz (2000). Henisz has designed his index as an indicator of the ability of political institutions to make credible commitments to an existing policy regime, which he argues is the most relevant political variable of interest to investors. Building on a simple spatial model of political interaction, the index makes use of the structure of government in a given country and the political views represented by the different levels of government (i.e., the executive and the lower and upper legislative chambers). It measures the extent to which political actors are constrained in their choice of future policies by the existence of other political actors with veto power who will have to consent. Using information on party composition of the executive and the legislative branches allows taking into account how alignment across branches of government and the extent of preference heterogeneity within each legislative branch impacts the feasibility of policy change. Scores range from 0, which indicates that the executive has total political discretion and could change existing policies at any point of time, to 1, which indicates that a change of existing policies is totally infeasible. Of course, in practice agreement is always feasible, so the maximum score is less than 1. The remaining measures of institutional quality are all compiled from the International Country Risk Guide (ICRG), published by Political Risk Services. They are the investment profile index, the index of government stability, an index of law and order, and the ICRG composite political risk index. These data are available from 1984 onwards and have the widest country coverage of all the 11. We do not include the Lomé Conventions or the follow-on Cotonou Agreement between the European Union and 77 countries from Africa, the Caribbean and the Pacific (ACP) since it is highly unlikely that these had a major impact on FDI.

do bilateral investment treaties increase foreign direct investment 237

sources for country risk ratings. The composite political risk index incorporates the other indices. The index varies from 0 (high risk) to 100 (low risk). The full composite political index or a comparable measure is more commonly used in studies investigating the determinants of FDI and BITs than the individual sub-components (see, e.g., Alesina and Perotti 1996; Wheeler and Mody 1992; UNCTAD 1998; Tobin and Rose-Ackerman 2005). However, as mentioned already, it also includes many items that are not strictly relevant to institutional quality, such as indicators of socioeconomic conditions, measures of conflict and ethnic tensions, and measures of military and religious involvement in the political process—see Appendix 2 for a detailed description of this variable. The composite index is therefore strictly speaking not purely a measure of institutional quality. For this reason, we also employ the three sub-components most relevant to investors and most closely connected to BIT provisions. The investment profile index varies from 0 to 12, 12 representing very low risk and 0 indicating very high risk. The index is made of ratings on 3 separate elements, each receiving equal weight: contract viability (risk of expropriation), profit repatriation, and payment delays. Similar to the investment profile index, the government stability index varies from 0 (high risk) to 12 (low risk) and is composed of three elements that receive equal weight: government unity, legislative strength, and popular support. The law and order index runs from 0 (worst) to 6 (best) and consists of a law component measuring the strength and impartiality of the legal system and an order component measuring the extent of popular observance of the law. 3. Estimation technique We estimate both random-effects and fixed-effects models, in which case we can employ robust standard errors. We suspect that there are factors making a country attractive to foreign investors that are not captured by our explanatory variables and that are (approximately) time-invariant, such as colonial history, culture, language, climate, geographical distance to the centers of the Western developed world, legal restrictions on inward FDI, and so on. Hausman tests, which test the random-effects assumption that these time-invariant factors are uncorrelated with the explanatory variables, by and large reject the assumption. We therefore focus on the fixed-effects estimation results. To mitigate potential reverse causality problems, we lag all explanatory variables by one period. Ideally, one would like to tackle this problem more comprehensively with the help of instrumental variable regression. However, practically all explanatory variables are potentially subject to reverse causality and it would be simply impossible to find adequate and valid instruments. Table 1 provides summary descriptive variable information together with a bivariate correlation matrix. Variance inflation analysis did not suggest reason for concern with multicollinearity problems. As in any regressions analysis, there is of course always the possibility of omitted variable bias. For example, we cannot account for over-time changes in domestic legislation encouraging or discouraging FDI other than what is captured by BITs, as there is no comprehensive information available. However, we see no reason why this or any other potentially omitted

238 eric neumayer and laura spess table 1. descriptive statistical variable information and bivariate correlation matrix Variable

Obs

Mean

Std. Dev.

Min

ln FDI flow ln FDI flow share BITs ln GDP p.c. ln Population Econ. Growth Inflation Resource rents Bilateral trade agreements WTO membership POLCON Composite political risk Investment profile Government stability Law and order

2767 2767 2767 2767 2767 2767 2767 2767 2767 2767 2767 1331 1351 1350 1349

3.92 –7.06 23.97 7.93 15.72 0.01 65.78 5.55 0.07 0.67 0.18 58.71 6.03 6.88 3.09

2.52 2.45 26.98 0.83 1.88 0.07 684.46 9.76 0.26 0.47 0.20 14.27 1.91 2.32 1.27

–4.69 –16.98 0 5.64 10.62 –0.42 –31.52 0 0 0 0 11.50 1 1 0

Max 10.78 –1.17 99.34 9.72 20.99 0.78 26762.02 66.60 2 1 0.67 94.17 11.13 12 6

variable should be systematically correlated with our explanatory variables to an extent that our results would be significantly biased. D. Results Table 2 presents random-effects estimation results for the logged amount of FDI in U.S. dollars of 1996 flowing to a country as the dependent variable. Column I starts with POLCON as the measure of institutional quality. Most variables test in accordance with theoretical expectations. Countries with a higher cumulative number of BITs, with fast-growing economies, and with larger populations receive more FDI. So do countries that are more intensive in natural resource extraction, that are members of the WTO, and that have a higher number of trade agreements with developed countries. A higher inflation rate deters FDI. The POLCON variable is statistically insignificant. How should one interpret this result? With interaction terms included, one cannot interpret the coefficients on the individual components in the conventional way. Instead, the coefficient on POLCON in a model with a significant interaction term BITs∗POLCON is the effect of POLCON on FDI when the BIT variable is zero (see Braumoeller [2004] for a nice exposition). It follows from our estimations that institutional quality as measured by POLCON has no effect on FDI in the absence of BITs. The interaction term is marginally significant, however, suggesting that BITs and

I: ln FDI flow share

1.00 II: ln FDI flow 0.95 III: BITs 0.41 IV: ln GDP p.c. 0.49 V: ln Population 0.43 VI: Econ. Growth 0.16 VII: Inflation –0.07 VIII: Resource rents 0.09 IX: Bilateral trade agreements 0.21 X: WTO membership 0.07 XI: POLCON 0.36 XII: Investment profile 0.38 XIIII: Government stability 0.31 XIV: Law and order 0.35 XV: Composite political risk 0.40

II

III

IV

– 1.00 0.28 0.48 0.45 0.15 –0.06 0.14 0.15 0.02 0.28 0.23 0.19 0.17 0.28

– – 1.00 0.31 0.21 0.08 –0.05 –0.12 0.30 0.05 0.16 0.26 0.25 0.37 0.28

– – – 1.00 –0.02 0.13 –0.07 0.13 0.30 –0.03 0.41 0.33 0.29 0.28 0.38

V

VI – – – –

1.00 0.07 0.02 0.07 0.05 0.02 –0.02 0.04 –0.02 0.07 0.03

– – – – – 1.00 –0.16 –0.04 0.03 –0.02 0.08 0.14 0.14 0.15 0.15

VI

VIII

IX

X

XI

– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 1.00 – – – – 0.02 – – – 1.00 –0.03 1.00 – – 0.04 0.02 –0.21 –0.09 1.00 – –0.02 –0.13 0.08 0.19 1.00 –0.12 –0.08 0.22 0.18 0.29 –0.15 –0.09 0.20 0.20 0.24 –0.12 0.00 0.23 0.04 0.18 –0.10 –0.05 0.24 0.03 0.22

XII – – – – – – – – – – –

1.00

XIII – – – – – – – – – – – –

XIV – – – – – – – – – – – – –

XV – – – – – – – – – – – – – –

0.49 1.00 0.58 0.68 1.00 0.58 0.35 0.46 1.00

do bilateral investment treaties increase foreign direct investment 239

I

240 eric neumayer and laura spess table 2. random-effects estimation results (logged fdi flows in 1996 dollars)

BITs ln GDP p.c. ln Population Econ. growth Inflation Resource rents Bilateral trade agreements WTO membership POLCON

I

II

III

IV

V

0.015

0.031

0.012

0.020

0.020

(5.34)∗∗∗

(3.79)∗∗∗

(2.17)∗∗

(3.66)∗∗∗

(3.96)∗∗∗

0.548

0.265

0.238

0.294

0.245

(5.53)∗∗∗

(1.68)∗

(1.50)

(1.86)∗

(1.64)∗

0.506

0.594

0.626

0.610

0.625

(10.27)∗∗∗

(6.98)∗∗∗

(7.24)∗∗∗

(7.08)∗∗∗

(7.96)∗∗∗

1.195

1.683

1.553

1.602

1.597

(2.52)∗∗

(2.42)∗∗

(2.19)∗∗

(2.25)∗∗

(2.26)∗∗

–0.0001

–0.0001

–0.0001

–0.0001

–0.0001

(2.07)∗∗

(1.73)∗

(1.46)

(1.65)∗

(1.64)

0.025

0.030

0.023

0.022

0.021

(4.34)∗∗∗

(3.61)∗∗∗

(2.76)∗∗∗

(2.62)∗∗∗

(2.59)∗∗∗

0.343

0.278

0.160

0.344

0.288

(1.83)∗

(1.34)

(0.72)

(1.56)

(1.34)

0.212

–0.018

–0.024

0.066

0.137

(1.98)∗∗

(0.12)

(0.16)

(0.44)

(0.92)

0.350

–

–

–

–

–

–

–

–

0.011

–

–

–

–

–

–

0.091

–

–

–

–

0.091

–

(1.17)

BITs∗POLCON

0.012 (1.78)∗

Composite political risk

–

BITs∗Comp. pol. risk

–

(1.81)∗

–0.0002 (1.73)∗

Investment profile

–

–

(2.16)∗∗

BITs∗Inv. profile

–

–

0.001 (1.08)

Government stability

–

–

–

(2.34)∗∗

BITs∗Gov. stability

–

–

–

–0.001

–

(0.77)

Law and order

–

–

–

–

BITs∗Law and order

–

–

–

–

0.247 (3.56)∗∗∗

–0.001 (0.87)

Observations Countries

2767 120

1346 91

1369 91

1368 91

1367 91 Continued

do bilateral investment treaties increase foreign direct investment 241

I Period R-squared (overall) Hausman test

II

III

IV

V

1970–2001 1984–2001 1984–2001 1984–2001 1984–2001 0.46 0.49 0.50 0.49 0.51 72.11 68.46 73.93 90.24 28.58 0.8659

0.0000

0.0000

0.0000

0.0000

Notes: Absolute t-values in parentheses. Year-specific time dummies included, but coefficients not reported. Hausman test is asymptotically χ2 distributed with p-values in brackets. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

POLCON-measured institutional quality are complements since the positive effect of cumulative BIT signature is higher when the POLCON index is high as well. In column II we replace Henisz’s (2000) policy constraints variable with the ICRG composite political risk index (where higher values on the index mean lower risk). Note the significant drop in the number of observations. This is not so much due to a loss of countries included in the sample, which drops from 120 to 91, but due to the fact that we lose all observations before 1984, the first year for which this variable has been coded. Despite the substantial reduction in sample size, the weighted sum of BITs variable is statistically significant with the expected positive sign. Results on the other variables are also relatively consistent in terms of sign and statistical significance of variables, but the trade agreement variables are no longer statistically significant. There is one further important further exception: the interaction term between institutional quality and our BIT variable is now statistically significant with a negative coefficient sign. This would suggest that BITs function as substitutes for high domestic institutional quality since the positive effect of cumulative BIT signature is higher when the ICRG composite index is low, that is, in high-risk environments. Importantly, while the positive effect of BITs on FDI decreases as political risk is reduced, the effect always remains positive, even at very low levels of risk.12 In the remaining columns, we replace the ICRG composite index with the selected individual sub-components. In column III, we look at a country’s investment profile, in column IV at a country’s governmental stability and in column V at its 12. Since the coefficient of the BIT variable represents the effect of BITs when the value of institutional quality is zero, one can re-scale the institutional quality variable such that zero represents the lowest level of risk. At the lowest observed level of risk in the sample, the positive effect of BITs is reduced from 0.030 (at the highest observed level of risk in the sample) to 0.011.

242 eric neumayer and laura spess

degree of law and order, respectively. Results are largely consistent with the ones for the ICRG composite index. Importantly, the weighted cumulative number of signed BITs is always statistically significant with the expected positive sign. However, the interaction term is not statistically significant for these subcomponents of the ICRG composite index. The Hausman test fails to reject the random-effects assumption only in column I, which underlines the importance of controlling for country fixed effects. Table 3 therefore reports results from fixed-effects estimation.

table 3. fixed-effects estimation results (logged fdi flows in 1996 dollars) I ln GDP p.c. ln Population Econ. growth Inflation Resource rents Bilateral trade agreements WTO membership POLCON

II

III

IV

V

1.916

3.771

3.304

4.052

3.691

(4.04)∗∗∗

(4.37)∗∗∗

(3.60)∗∗∗

(4.47)∗∗∗

(4.21)∗∗∗

–1.344

–4.942

–4.513

–5.176

–5.033

(2.66)∗∗∗

(5.21)∗∗∗

(4.51)∗∗∗

(5.19)∗∗∗

(5.24)∗∗∗

1.134

2.372

2.366

2.343

2.464

(1.81)∗

(3.35)∗∗∗

(3.13)∗∗∗

(3.12)∗∗∗

(3.24)∗∗∗

–0.0001

–0.0001

–0.0001

–0.0001

–0.0001

(3.16)∗∗∗

(2.36)∗∗

(2.18)∗∗

(2.41)∗∗

(2.48)∗∗

0.030 (3.63)∗∗∗

0.036 (2.96)∗∗∗

0.031 (2.41)∗∗

0.031 (2.47)∗∗

0.028 (2.16)∗∗

0.532

0.119

0.061

0.289

0.199

(2.17)∗∗

(0.66)

(0.32)

(1.44)

(1.02)

0.218

–0.081

–0.111

–0.047

0.027

(1.98)∗∗

(0.52)

(0.73)

(0.31)

(0.18)

0.233

–

–

–

–

–

–

–

–

0.014

–

–

–

–

–

–

0.117

–

–

–

–

(0.71)

BITs∗POLCON

0.011 (1.29)

Composite political risk

–

BITs∗Comp. pol. risk

–

(2.05)∗∗

–0.0003 (2.17)∗∗

Investment profile

–

–

(2.84)∗∗∗

BITs∗Inv. profile

–

–

–0.000 (0.01)

Continued

do bilateral investment treaties increase foreign direct investment 243

I

II

III

IV

V

Government stability

–

–

–

0.128

–

BITs∗Gov. stability

–

(2.88)∗∗∗

–

–

–0.001

–

(2.12)∗∗

Law and order

–

–

–

–

0.290 (4.13)∗∗∗

BITs∗Law and order

–

–

–

–

–0.002 (1.14)

Observations

2767

1346

1369

1368

1367

Countries

120

91

91

91

91

Period

1970–2001 1984–2001 1984–2001 1984–2001 1984–2001

R-squared (within)

0.22

0.30

0.30

0.30

0.30

Notes: Absolute t-values in parentheses. Year-specific time dummies included, but coefficients not reported. Robust standard errors. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

Fixed-effects estimation results are rather similar to the ones from randomeffects estimation with two important exceptions. First, the interaction term between POLCON and the BIT variable is now insignificant, whereas the interaction term with governmental stability is significant and negative, suggesting that BITs and governmental stability function as substitutes. As before, the positive effect of BITs on FDI becomes smaller the more stable governments are, but never to an extent that the effect would become negative. Second, the log of population size now switches signs and is statistically significant with a negative coefficient. Keeping in mind that the fixed-effects estimation is based on the within-variation of the data in each country only, whereas the random-effects estimation is based on both cross-country variation and within-variation, this can be interpreted to the effect that countries with larger populations receive more FDI conditional on the other explanatory variables, but as a country’s population grows, it receives less rather than more FDI conditional on the other explanatory variables and the country-specific fixed effects. Table 4 presents results for the logged country share of developing country FDI as the dependent variable. The reported results are based on fixed-effects estimation, since Hausman tests overwhelmingly rejected the random-effects assumption in all cases. Results are very similar to the ones reported above for

244 eric neumayer and laura spess table 4. fixed-effects estimation results (logged fdi flows as share of developing country fdi)

BITs ln GDP p.c. ln Population Econ. growth Inflation Resource rents

I

II

III

IV

V

0.013

0.035

0.018

0.027

0.023

(3.47)∗∗∗

(4.04)∗∗∗

(3.33)∗∗∗

(5.02)∗∗∗

(4.65)∗∗∗

3.008

1.916

1.754

2.144

2.047

(10.10)∗∗∗

(3.13)∗∗∗

(3.03)∗∗∗

(3.42)∗∗∗

(3.52)∗∗∗

–2.613

–2.686

–2.654

–2.864

–2.844

(11.45)∗∗∗

(5.13)∗∗∗

(5.66)∗∗∗

(5.37)∗∗∗

(5.98)∗∗∗

1.334

2.268

2.348

2.225

2.146

(2.15)∗∗

(3.18)∗∗∗

(3.06)∗∗∗

(2.93)∗∗∗

(2.81)∗∗∗

–0.0001

–0.0001

–0.0001

–0.0001

–0.0001

(3.27)∗∗∗

(2.33)∗∗

(2.10)∗∗

(2.38)∗∗

(2.53)∗∗

0.027

0.032

0.027

0.028

0.027

(3.47)∗∗∗

(2.70)∗∗∗

(2.18)∗∗

(2.24)∗∗

(2.18)∗∗

Bilateral trade agreements

0.539

0.216

0.113

0.343

0.302

(2.25)∗∗

(1.29)

(0.60)

(1.82)∗

(1.65)∗

WTO membership

0.133

–0.049

–0.075

–0.017

0.031

(1.26)

(0.31)

(0.49)

(0.11)

(0.20)

0.184

–

–

–

–

–

–

–

–

0.013

–

–

–

–

–

–

0.108

–

–

–

–

0.104

–

POLCON

(0.57)

BITs∗POLCON

0.008 (0.96)

Composite political risk

–

BITs∗Comp. pol. risk

–

(2.00)∗∗

–0.0003 (2.04)∗∗

Investment profile

–

–

(2.69)∗∗∗

BITs∗Inv. profile

–

–

–0.000 (0.02)

Government stability

–

–

–

(2.59)∗∗∗

BITs∗Gov. stability

–

–

–

–0.001

–

(2.11)∗∗

Law and order

–

–

–

–

0.231 (3.36)∗∗∗

BITs∗Law and order

–

–

–

–

–0.002 (1.20)

Continued

do bilateral investment treaties increase foreign direct investment 245

Observations Countries Period R-squared (within)

I

II

III

IV

V

2767 120 1970–2001 0.09

1346 91 1984–2001 0.07

1369 91 1984–2001 0.07

1368 91 1984–2001 0.06

1367 91 1984–2001 0.07

Notes: Absolute t-values in parentheses. Robust standard errors. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

the other dependent variable. In particular, a higher cumulative number of signed BITs is associated with a higher share of FDI inflows. So is institutional quality with the exception of POLCON. As before, there is some limited evidence that BITs and institutional quality are substitutes for each other, but the interaction term is only statistically significant with a negative coefficient sign for the composite ICRG index and its governmental stability sub-component. E. Sensitivity Analysis Lagging the explanatory variables by one year mitigates potential simultaneity bias, but this lag length is somewhat arbitrary. The positive impact of the BIT variable on FDI inflows is maintained if the lag length is two, three or four years instead.13 Tobin and Rose-Ackerman (2005) use five-year period averages to avoid the impact of year-to-year variation. Maintaining a lag of one year but averaging the data over five-year periods does not dramatically change our results, as the fixed-effects estimation results reported in Table 5 attest, with the share of developing country FDI inflow as the dependent variable. With less variation in the data, we are not surprised to find that some variables lose statistical significance. Importantly, however, the BIT variable remains statistically significant with the expected positive sign throughout. The institutional variables and interaction terms also test almost as before in terms of sign and whether they are statistically significantly different from zero. The exception is the interaction between the BIT variable and governmental stability, which becomes only marginally insignificant. In further non-reported sensitivity analysis, we briefly explored the issue of whether BITs are all about a signaling effect. Maybe the only thing that matters is that a developing country has signed one BIT (perhaps with a major capital exporter), and signing any more BITs has no additional effect. When we added a dummy variable that is set to one if a developing country has signed a BIT with 13. All results that are not explicitly reported are available on request.

246 eric neumayer and laura spess table 5. five-year period averages fixed-effects estimation results (logged fdi flows as share of developing country fdi)

BITs ln GDP p.c. ln Population Econ. growth Inflation Resource rents Bilateral trade agreements WTO membership POLCON

I

II

III

IV

V

0.014

0.048

0.030

0.042

0.034

(2.15)∗∗

(3.80)∗∗∗

(2.91)∗∗∗

(4.16)∗∗∗

(3.68)∗∗∗

3.036

1.948

1.792

2.099

1.957

(6.47)∗∗∗

(2.33)∗∗

(2.17)∗∗

(2.39)∗∗

(2.47)∗∗

–2.601

–2.766

–2.702

–2.849

–2.944

(6.92)∗∗∗

(3.50)∗∗∗

(3.63)∗∗∗

(3.58)∗∗∗

(4.25)∗∗∗

4.337

3.488

3.300

3.564

4.014

(2.14)∗∗

(1.21)

(1.03)

(1.11)

(1.29)

–0.0001

–0.0001

–0.0001

–0.0001

–0.0001

(1.95)∗

(0.97)

(1.06)

(1.21)

(1.23)

0.029

0.039

0.042

0.045

0.045

(1.94)∗

(1.55)

(1.69)∗

(1.90)∗

(1.84)∗

1.112

0.482

0.336

0.634

0.530

(2.26)∗∗

(1.33)

(0.79)

(1.50)

(1.38)

0.242

–0.245

–0.231

–0.146

–0.099

(1.17)

(0.82)

(0.77)

(0.49)

(0.34)

–0.066

–

–

–

–

–

–

–

–

0.020

–

–

–

–

–

–

0.138

–

–

–

–

0.127

–

(0.10)

BITs∗POLCON

0.008 (0.49)

Composite political risk

–

BITs∗Comp. pol. risk

–

(1.72)∗

–0.000 (1.79)∗

Investment profile

–

–

(1.64)∗

BITs∗Inv. profile

–

–

–0.001 (0.37)

Government stability

–

–

–

(1.71)∗

BITs∗Gov. stability

–

–

–

–0.002

–

(1.61)

Law and order

–

–

–

–

0.341 (3.01)∗∗∗

BITs∗Law and order

–

–

–

–

–0.002 (0.91)

Continued

do bilateral investment treaties increase foreign direct investment 247

Observations Countries Period R-squared (within)

I

II

III

IV

V

637 120 1970–2001 0.19

317 91 1984–2001 0.19

314 91 1984–2001 0.20

314 91 1984–2001 0.19

314 91 1984–2001 0.21

Notes: Absolute t-values in parentheses. Robust standard errors. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

any developed country, this dummy variable was statistically significant with the expected positive sign. However, as before, the weighted cumulative sum of BITs variable remained positive and statistically significant as well with its magnitude only slightly reduced. This remains true if we replace the dummy variable with one for BITs signed with any of the six major capital exporters, namely France, Germany, Japan, the Netherlands, the United Kingdom and the United States. We interpret this as further evidence that BITs are likely to fulfill the dual function of both signaling and commitment. Next, we checked whether our main results are due to the presence of problematic countries. In particular, we excluded all Eastern European and former Soviet Union countries from the sample since, with the exception of Hungary, these countries do not seem to be included in the analysis by Tobin and RoseAckerman (2005). However, results hardly change. We also excluded countries with a population size of less than one million from the sample to eliminate the influence of very small countries, but results were again hardly affected. We even restricted the list of countries to be exactly the same as in Tobin and Rose-Ackerman (2005) and still found a positive effect of BITs on FDI throughout, with one exception (investment profile as measure of institutional quality), in which case the BIT variable is marginally insignificant with a positive coefficient sign. This holds true both in the annual and in the five-year period model specification. Next, we wanted to test more formally whether results are driven by a few influential outliers. Belsley, Kuh, and Welsch (1980) suggest that observations that have both high residuals and a high leverage deserve special attention. We excluded an observation if its so-called DFITS is greater than twice the square root of (k/n), where k is the number of independent variables and n the number of observations. DFITS is defined as the square root of (hi/(1–hi)), where hi is an observation’s leverage, multiplied by its studentized residual. Applying this criterion leads to the exclusion of up to 385 observations. Results are remarkably consistent, however.

248 eric neumayer and laura spess

In further sensitivity analysis, we also included a measure of trade openness, which has a theoretically ambiguous effect on FDI (Taylor 2000). On one hand, countries that are more open to trade can be more attractive host countries if the main purpose of foreign investment is to export the goods or services produced. On the other hand, high trade barriers could make it in a company’s best interest to locate production within the host country in order to circumvent the import barriers. Trade openness tested sometimes insignificant, and sometimes (marginally) significant with a negative coefficient sign in the estimations, hardly affecting the results of the other variables. Since its inclusion would reduce the sample size by about 20 countries, we did not include this variable in the reported estimations. Following Noorbakhsh, Paloni, and Youssef (2001), we also included the secondary enrollment ratio to account for human capital as a determinant of FDI in non-reported analysis. The number of observations dropped substantially, leaving results for the other variables mainly unaffected. The enrollment ratio itself is never statistically significant, even though it is positively signed in line with expectations.

conclusion Developing countries that sign more BITs with developed countries receive more FDI inflows. The effect is robust to various sample sizes, model specifications, and whether or not FDI flows are normalized by the total flow of FDI going to developing countries. There is some limited evidence that BITs function as substitutes for institutional quality, as in a few estimations the interaction term between the accumulated number of BITs variable and institutional quality is negative and statistically significant. The message to developing countries, therefore, is that succumbing to the obligations of BITs does have the desired payoff of higher FDI inflows. To our knowledge, ours is the first study to provide robust empirical evidence that BITs fulfill their stated objective. Those with particularly poor domestic institutional quality possibly stand to gain most from BITs, but there is no robust and consistent evidence for this. Why do we come to different conclusions than the three other relevant studies? Hallward-Driemeier’s (2003) study does not allow for a signaling effect and suffers from a small non-representative sample due to the dyadic research design. Salacuse and Sullivan’s (2005) analysis is cross-sectional and therefore cannot detect how a higher number of BITs raises the flow of FDI to signatory developing countries over time. Our difference from the results presented by Tobin and Rose-Ackerman (2005) is more puzzling. As we noted in the sensitivity analysis, our results uphold if we adopt their five-year period averages approach and restrict the list of countries to be exactly the same as in their analysis. It is therefore difficult to know where the difference comes from. One possibility is that we do not log the number of BITs, not least because the log of zero (BITs)

do bilateral investment treaties increase foreign direct investment 249

is not defined. Whatever the cause, we find Tobin and Rose-Ackerman’s (2005) result that each additional BIT lowers (rather than raises) the flow of FDI to developing countries with high political risk as extremely difficult to believe. BITs might not raise FDI flows in contexts of high risk, but there is no reason whatsoever to expect that they should lower FDI flows. Statistical significance is not equivalent to substantive importance. We therefore need to know how strong is the effect of the BIT variable on FDI flows. How much more FDI can a developing country expect if it aggressively engages in a program to sign BITs with developed countries? To answer this question, we look at a one standard deviation increase in the BIT variable (equivalent to an increase of around 27 in the weighted cumulative BIT variable running from 0 to 99). Since in some regressions the interaction effect between the BIT variable and institutional quality is statistically significant, the overall effect of signing up to BITs sometimes depends on the level of institutional quality, in which case for simplicity we fix institutional quality at its median.14 Based on the estimations in Table 3, a country experiencing an increase of one standard deviation in the BIT variable is predicted to increase its FDI inflow by between 43.7% and 93.2%. Based on the results from Table 4, such a country is predicted to increase its share of FDI inflow relative to the total inflow to developing countries by between 42.0% and 104.1%. Clearly, these are non-negligible increases following a substantial increase in BIT activity. But whether the demonstrated benefits of concluding BITs in the form of increased FDI inflows are higher than the substantial costs developing countries incur in negotiating, signing, concluding and complying with the obligations typically contained in such treaties is impossible to tell. What we do know is that BITs fulfill their purpose, and those developing countries that have signed more BITs with major capital-exporting developed countries are likely to have received more FDI in return.

14. Strictly speaking, the coefficient of the BIT variable always represents the effect of BITs at a zero value of the institutional quality measure, but if the interaction term is not statistically significantly different from zero, then, as a first approximation, it can be taken as the effect of BITs at any value of institutional quality.

250 eric neumayer and laura spess

appendix 1 List of Countries included in Sample Albania, Algeria, Angola, Antigua and Barbuda, Argentina, Armenia, Azerbaijan, Bangladesh, Belarus, Belize, Benin, Bolivia, Botswana, Brazil, Bulgaria, Burkina Faso, Burundi, Cambodia, Cameroon, Cape Verde, Central African Republic, Chad, Chile, China, Colombia, Comoros, Congo (Dem. Rep.), Congo (Rep.), Costa Rica, Côte d’Ivoire, Croatia, Czech Republic, Dominica, Dominican Republic, Ecuador, Egypt, El Salvador, Equatorial Guinea, Estonia, Ethiopia, Fiji, Gabon, Gambia, Georgia, Ghana, Grenada, Guatemala, Guinea, Guinea-Bissau, Guyana, Haiti, Honduras, Hungary, India, Indonesia, Iran, Jamaica, Jordan, Kazakhstan, Kenya, Korea (Rep.), Kyrgyz Republic, Latvia, Lebanon, Lesotho, Lithuania, Macedonia FYR, Madagascar, Malawi, Malaysia, Mali, Mauritania, Mauritius, Mexico, Moldova, Morocco, Mozambique, Namibia, Nepal, Nicaragua, Niger, Nigeria, Pakistan, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Poland, Romania, Russian Federation, Rwanda, São Tomé and Principe, Senegal, Seychelles, Sierra Leone, Slovak Republic, South Africa, Sri Lanka, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Swaziland, Syria, Tanzania, Thailand, Togo, Trinidad and Tobago, Tunisia, Turkey, Uganda, Ukraine, Uruguay, Uzbekistan, Venezuela, Vietnam, Yemen, Zambia, Zimbabwe.

appendix 2 ICRG Composite Political Risk Index Sequence

Political Risk Component

A B C D E F G H I J K L Total

Government Stability Socioeconomic Conditions Investment Profile Internal Conflict External Conflict Corruption Military in Politics Religion in Politics Law and Order Ethnic Tensions Democratic Accountability Bureaucracy Quality

Source: www.prsgroup.com

Max. points 12 12 12 12 12 6 6 6 6 6 6 4 100

do bilateral investment treaties increase foreign direct investment 251

references Alesina, A. and R. Perotti (1996). “Income distribution, political instability, and investment,” European Economic Review, 40, pp. 1203–1228. Belsley, D.A., E. Kuh and R.E. Welsch (1980). Regression Diagnostics (New York: John Wiley). Braumoeller, B.F. (2004). “Hypothesis testing and multiplicative interaction terms,” International Organization, 58, pp. 807–820. Chakrabarti, A (2001). “The determinants of foreign direct investment: sensitivity analyses of cross-country regressions,” Kyklos, 54, pp. 89–114. De Soysa, I. and J.R. Oneal (1999). “Boon or bane? Reassessing the productivity of foreign direct investment,” American Sociological Review, 64, pp. 766–782. Dolzer, R. and M. Stevens (1995). Bilateral Investment Treaties (The Hague: Martinus Nijhoff Publishers). Dunning, J.H. (1998). “The changing geography of foreign direct investment,” in N. Kumar, ed., Globalization, Foreign Direct Investment and Technology Transfers: Impacts on and prospects for developing countries. (New York: Routledge). Elkins, Z., A. Guzman and B. Simmons (2004). “Competing for capital: the diffusion of bilateral investment treaties, 1960–2000,” Working Paper (Champaign, Berkeley, and Cambridge: University of Illinois, University of California at Berkeley and Harvard University). EU [European Union] (2004). EC Regional Trade Agreements. (Brussels: European Union Directorate General Trade), http://europa.eu.int/comm/trade/index_en.htm. Globerman, S. and D. Shapiro (2002). “Global foreign direct investment flows: the role of governance infrastructure,” World Development, 30, pp. 1899–1919. Ginsburg, T. (2004). “International substitutes for domestic institutions: bilateral investment treaties and governance,” Working Paper (Champaign: University of Illinois College of Law). Guzman, A. (1998). “Why LDCs sign treaties that hurt them: explaining the popularity of bilateral investment treaties,” Virginia Journal of International Law, 38, pp. 639–688. Hallward-Dreimeier, M. (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and it might bite,” World Bank Policy Research Working Paper No. 3121 (Washington, D.C.: World Bank). Henisz, W. J. (2000). “The institutional environment for economic growth,” Economics and Politics, 12, 1, pp. 1–31. Minor, M.S. (1994). “The demise of expropriation as an instrument of LDC policy, 1980–1992,” Journal of International Business Studies, 25, pp. 177–188. Neumayer, E. (2001). Greening Trade and Investment. (London: Earthscan). —— (2006). Self-interest, Foreign Need and Good Governance: Are Bilateral Investment Treaty Programs Similar to Aid Allocation? Foreign Policy Analysis, 2, pp. 245–267 Noorbakhsh, F., A. Paloni and A. Youssef.(2001). “Human capital and FDI inflows to developing countries: new empirical evidence,” World Development, 29, pp. 1593–1610. Perry, A. (2000). “Effective legal systems and foreign direct investment: in search of the evidence,” International and Comparative Law Quarterly, 49, pp. 779–799. Peters, P. (1996). “Review of Dolzer and Stevens: bilateral investment treaties,” Netherlands International Law Review, 43 (1), pp. 103–108.

252 eric neumayer and laura spess Peterson, L.E. (2004). Bilateral Investment Treaties and Development Policy-Making. (Winnipeg: International Institute for Sustainable Development). Raghavan, C. (1997). “Bilateral investment treaties play only a minor role in attracting FDI,” Third World Economics, June 1–15, p. 162. Salacuse, J.W. (1990). “BIT by BIT: The growth of bilateral investment treaties and their impact on foreign investment in developing countries,” The International Lawyer, 24, 3, pp. 655–675. Salacuse, J.W. and Nicholas P. Sullivan (2005). “Do BITs really work? An evaluation of bilateral investment treaties and their grand bargain,” Harvard International Law Journal, 46, pp. 67–130. Schneider, F. and B. Frey (1985). “Economic and political determinants of foreign direct investment,” World Development, 13, pp. 161–175. Sornarajah. M. (1986). “State responsibility and bilateral investment treaties,” Journal of World Trade Law, 20, pp. 79–98. Taylor, C.J. (2000). “The impact of host country government policy on U.S. multinational investment decisions,” The World Economy, 23, pp. 635–647. Tobin, J. and S. Rose-Ackerman (2005). “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Economics and Public Policy Research Paper No. 293, (New Haven: Yale Law School Center for Law). United Nations Conference on Trade and Development (UNCTAD) (1998). Bilateral Investment Treaties in the Mid-1990s (New York and Geneva: United Nations). —— (2001). Bilateral Investment Treaties 1959–1999 (New York and Geneva: United Nations). —— (2003a). World Investment Report 2003: FDI Policies for Development National and International Perspectives (New York and Geneva: United Nations). —— (2003b). World Investment Directory (New York and Geneva: United Nations). —— (2004). World Investment Report 2004: The Shift Toward Services (New York and Geneva: United Nations). Vandevelde, K (1996). “Review of Dolzer and Stevens: bilateral investment treaties,” American Journal of International Law, 90, 3, pp. 545–547. —— (1998). “The political economy of a bilateral investment treaty,” The American Journal of International Law, 92, 4, pp. 621–641. —— (2000). “The economics of bilateral investment treaties,” Harvard International Law Journal, 41, 2, pp. 469–502. Wheeler, D. and A. Mody (1992). “International investment location decisions: the case of U.S. Firms,” Journal of International Economics, 33, pp. 57–76. World Bank (2003a). World Development Indicators CD-ROM (Washington, D.C.: World Bank). —— (2003b). Adjusted Net Savings Data (Washington, D.C.: World Bank). http:// lnweb18.worldbank.org/ESSD/envext.nsf/44ByDocName/ GreenAccountingAdjustedNetSavings. World Trade Organization (WTO) (2004). Regional Trade Agreements. (Geneva: World Trade Organization). www.wto.org. Zebregs, H. (1998). “Can the neoclassical model explain the distribution of foreign direct investment across developing countries?,” IMF Working Paper WP/98/139. (Washington DC: International Monetary Fund).

8. the impact of bilateral investment treaties on foreign direct investment∗ peter egger and michael pfaffermayr introduction The first bilateral investment treaty (BIT) was signed between Germany and Pakistan in 1959 and came into force in 1962. Up to 1999, another 1,856 BITs have been signed and further BITs are expected in the future (United Nations, 2000). BITs are designed to facilitate foreign direct investment (FDI) from economies with abundant capital and skilled labor, that is, mainly Organisation for Economic Co-operation and Development (OECD) countries, to the less developed economies. Many of the existing BITs between the current OECD economies involve one old and one new OECD member. For example, the Czech Republic, Hungary, Poland, and the Slovak Republic concluded BITs with old OECD members in the early 1990s and then joined the OECD afterwards. The theoretical literature on the expected impact of BITs on FDI is not conclusive. Hoekman and Saggi (2000) argue that, due to some differences in national rules, BITs may be the source of higher transaction costs and uncertainty from a firm’s perspective. Although this point would support an argument for a harmonized global BIT, that is, a multilateral investment treaty, these authors concede that differences in cultural, political, and general business climate characteristics are more important determinants of the transaction costs associated with FDI. Formally, BITs regulate FDI-related issues such as admission, treatment, expropriation, and the settlement of disputes at the bilateral level. Ex ante, they establish transparency about risk and, thus, reduce the risk of investing in a country. Ex post, BITs ensure that firms have certain rights, for example property rights, and preserve them from expropriation.1 According to the Fact Sheet on the U.S. bilateral investment treaty program released by the Office of Investment Affairs of the Bureau of Economic Business Affairs, the program’s basic aims are the following.2 First, BITs should protect U.S. FDI in those countries where

∗ This chapter was reprinted with permission from the Journal of Comparative Economics. The chapter was originally published as “The impact of bilateral investment treaties on foreign direct investment,” 32 Journal of Comparative Economics 788 (2004). Egger acknowledges financial support by the Fonds zur Förderung der wissenschaftlichen Forschung through the Erwin Schroedinger Auslandsstipendium Grant J2280-G05. 1. Maskus (2000) addresses the issue of protecting intellectual property. Drabek (2002) considers the importance of the risk of expropriation. 2. Hallward-Driemeier (2003) provides further details.

254 peter egger and michael pfaffermayr

U.S. investors’ rights are not protected through existing agreements. Second, they should encourage host countries to adopt market-oriented domestic policies that treat private investment fairly. Third, they should support the development of international law standards consistent with these objectives.3 In some sense, BITs extend an investor’s property rights and regulate how host governments must arbitrate disputes covered by the treaty. Further, BITs define what is deemed expropriation, formulate how and under which conditions property may be expropriated, and determine how quickly and comprehensively investors must be compensated. The United Nations Conference on Trade and Development (UNCTAD 1998) study summarizes the following features of BITs, which are designed to attract FDI. First, BITs facilitate and encourage bilateral FDI between the contracting parties. To achieve this goal, most BITs guarantee foreign investors fair and equitable, non-discriminatory, most-favored-nation, and national treatment in addition to access to international means of dispute resolution. Moreover, BITs usually provide legal protection of both physical and intellectual property under international law and investment guarantees with a special focus on the transfer of funds and expropriation, including the rules of compensation. In this way, they facilitate insurance and reduce insurance premia. Some BITs provide even more reliable and transparent conditions for investors than do national laws. Hence, they allow transition economies to provide guarantees for foreign investors while undertaking national legislative reforms at the same time.4 From this perspective, BITs reduce the costs of investing abroad, including risk premia, so that FDI should increase if new BITs are implemented. In addition, BITs should make new inward investment attractive and also reduce the likelihood of investment outflows. The theoretical trade literature incorporates multinational enterprises (MNEs) in trade models characterized by increasing returns and considers both horizontal MNEs and vertical MNEs. Horizontal MNEs have production facilities in both the parent and host countries (Markusen 1984; Markusen and Venables 1998, 2000) and tend to be found in the similarly endowed economies, for example within the OECD.5 Vertical MNEs unbundle completely the headquarter services from production to exploit factor cost differentials (Helpman 1984; Helpman and Krugman 1985). Therefore, vertical FDI tends to occur in dissimilar economies,

3. Hoekman and Saggi (2000) remark that, with the notable exception of those negotiated by the U.S., BITs do not usually address the question of market access liberalization. 4. Whereas BITs aim to avoid additional fixed costs by reducing these types of risk, bilateral tax treaties deal with the repatriation of profits. Davies (2004) and Chisik and Davies (2004) present a thorough theoretical treatment of tax treaties. Blonigen and Davies (2003) provide an empirical assessment of the impact of bilateral tax treaties on FDI. 5. Horstmann and Markusen (1987 and 1992) model horizontal MNEs in a somewhat different framework.

the impact of bilateral investment treaties on foreign direct investment 255

for example between the OECD and the developing countries. Carr et al. (2001), Markusen (2002), and Markusen and Maskus (2002) develop knowledgecapital models of MNEs in which both horizontal and vertical activities arise endogenously. In a panel econometric framework, Hallward-Driemeier (2003) finds little evidence of any positive impact of BITs on FDI. A study based on cross-section analysis by UNCTAD (1998) supports only a weak nexus between signing BITs and changes in FDI flows and stocks. By contrast, the United Nations (2000) views BITs as the most important instrument for protecting FDI at the international level. In this paper, we undertake an empirical assessment of the impact of BITs on FDI stocks. We estimate several variants of the knowledge-capital model of MNEs using the largest available panel of outward FDI stocks provided by OECD, which contains FDI of OECD countries into both OECD and nonOECD economies. Information on BITs, both signed and ratified, is available from the World Bank. We find a significant and positive impact of ratified BITs throughout. The estimated effect of BITs on real outward FDI stocks amounts to about 30% in the preferred specification. Additionally, we look at whether simply signing a BIT will have a positive anticipation effect. We find a positive impact from signing a treaty, although its magnitude is smaller than that associated with the ratification of an existing treaty. However, the estimated anticipation effect is insignificant, in most specifications, leading us to conclude that the advantages to simply signing a BIT are inconsequential. Section A below provides details on the econometric specification and discusses the construction of the variables. In Section B, we present the main estimation results together with extensive sensitivity analysis. The conclusion presents a summary of the empirical findings. A. Specification and Data Base In the estimated empirical models, we focus on variants of the knowledge-capital model estimated by Carr et al. (2001), Egger and Pfaffermayr (2004b), and Markusen and Maskus (2002). Carr et al. (2001) and Markusen and Maskus (2002) use foreign affiliate sales as the dependent variable because their twofactor knowledge-capital model does not include physical capital. Egger and Pfaffermayr (2004a, 2004b) present a three-factor model and derive explicitly the hypotheses for a specification having FDI stocks as the dependent variable. However, Blonigen et al. (2003) illustrate that the key parameters are qualitatively similar if sales instead of FDI are used. This literature supports using four types of variables to explain the stock of outward FDI at the bilateral level, namely, country size, factor endowments, trade and FDI frictions, and interaction terms. Table 1 summarizes the definitions of our explanatory variables and reports their expected signs from the knowledge-capital model, the horizontal models, and the vertical models.

256 peter egger and michael pfaffermayr

Both absolute bilateral country size (ΣGDP) and similarity in bilateral country size (SIMI) affect horizontal FDI positively.6 In contrast to national exporting firms and vertical MNEs, horizontal MNEs run a production plant in each market and, thus, incur higher fixed costs. A larger size of both the home and the host market increases the likelihood that horizontal MNEs cover these fixed costs. The bilateral difference in the endowment ratio of skilled to unskilled labor (ΔSK) increases vertical FDI, because vertical MNEs arise only if countries differ in terms of production costs, that is, if ΔSK >0. The difference in the skilled to unskilled labor endowment ratio supports vertical FDI to a lesser extent, if the bilateral distance is large, the home country is large, or bilateral size is large. Transport costs impede trade and, thus, sales of vertical MNEs. The positive nexus between distance and transport costs motivates the interaction term, denoted by DIST·ΔSK, which is nonzero if ΔSK>0. The home country size effect supports the inclusion of the interaction term, denoted by ΔGDP·ΔSK, which is nonzero if ΔSK>0.7 The bilateral size-related interaction term is defined as ΣGDP·ΔSK. In Table 1, the last three interaction terms refer to specifications estimated in Markusen and Maskus (2002). Finally, BITs should reduce the impediment to foreign investment and foster FDI, irrespective of whether horizontal or vertical MNEs are considered. We distinguish between the anticipation effect (BITs) and the ratification effect (BITR) of BITs. Hence, the variable BITs is coded 1 after the date of signing and 0 before and the variable BITR is defined analogously for ratification. One set of specifications includes BITR only, while the other one includes both BITs and BITR. By itself, BITR measures the overall effect of a BIT after it has been implemented; however, if BITs is also included in the regressions, BITR reflects the additional impact of ratification. In the latter case, the overall effect equals the sum of the estimated BITs and BITR coefficients.

6. Carr et al. (2001) and Markusen and Maskus (2002) use the squared difference in bilateral GDP instead of SIMI. Whereas SIMI rises if two countries are similar with respect to GDP, the squared difference in GDP declines. Therefore, we expect a positive sign on the coefficient for SIMI (Egger and Pfaffermayr 2004a). 7. A large difference between the parent and the host countries’ skilled to unskilled labor endowment ratio (ΔSK) is associated with both more horizontal and more vertical FDI, because skilled-labor-abundant countries have a comparative advantage in inventing blue prints and setting up firms or multinational networks. However, this effect applies less to large parent economies (ΔGDP·ΔSK). Blonigen et al. (2003) argue that (ΔGDP)2·|ΔSK| should be used instead. Ekholm (1998) motivates a similar, although more parsimonious, specification. In a reply to Blonigen et al. (2003), Carr et al. (2003) verify that the simple difference rather than the absolute difference in factor endowments should be used as a regressor. Moreover, Carr et al. (2003) and Markusen and Maskus (2002) recommend a specification that allows a positive skill difference to exert a different effect than a negative skill difference. In our case, the specification issue has an impact neither on the sign of the dummies for BITs nor on their significance.

Abbreviation

Definition

KK

HOR

ΣGDPijt SIMIijt ΔSKijt D+ = 1 if ΔSKijt>0, 0 otherwise D− = 1 if ΔSKijt<0, 0 otherwise DISTij·D+·ΔSKijt ΔGDPijt·D+·ΔSKijt ΣGDPijt·D+·ΔSKijt −ΣGDPijt·D−·ΔSKijt BITSijt (BIT signed) BITFijt (BIT into force)

ln(GDPit+GDPjt) ln{1−[GDPit/(GDPit+GDPjt)]2−[GDPjt/(GDPit+GDPjt)]2} ln(tert. school enr.it)−ln(tert. school enr.jt) dummy variable dummy variable ln(distanceij)·D+·[ln(tert. school enr.it)−ln(tert. school enr.jt)] [ln(GDPit)−ln(GDPjt)]·D+·[ln(tert. school enr.it)−ln(tert. school enr.jt)] ln(GDPit+GDPjt)·D+·[ln(tert. school enr.it)−ln(tert. school enr.jt)] −ln(GDPit+GDPjt)·D−·[ln(tert. school enr.it)−ln(tert. school enr.jt)] 1 after the BIT has been signed, 0 otherwise 1 after the BIT has come into force, 0 otherwise

+ + +/− – – +/− − − − + +

+ + +/− – – +/− 0 − − + +

VER 0

0 + – – − − + − + +

Notes: (i) The notation i indicates the parent country, j refers to the host country, and t is the time index. (ii) The labels KK, HOR, and VER denote the predicted signs from the knowledge capital, the horizontal, and the vertical models of the multinational enterprise, respectively. (iii) The model predictions are derived in Carr et al. (2001), Egger and Pfaffermayr (2004a and 2004b) and Markusen and Maskus (2002).

the impact of bilateral investment treaties on foreign direct investment 257

table 1. variables and theoretical predictions

258 peter egger and michael pfaffermayr

We deflate nominal outward FDI stocks in U.S. dollars and employ homecountry investment deflators from the World Bank, using 1995 as the base year to approximate real stocks of outward FDI. The explanatory variables consist of real GDP, tertiary school enrollment, population figures, and bilateral distance between capitals. Information on signed and ratified BITs is available from the World Bank for the years from 1959 to 1999. Table 1 provides details on the definition of variables and Table A1 in the Appendix identifies the data sources. Due to missing FDI and school enrollment data, we restrict ourselves to the period from 1982 to 1997. The full design matrix for this period would contain 16,017 observations, whereas only 4,291 remain after adjusting for missing values, mainly of FDI stock data. In Table 2, we report summary statistics for the key variables in the whole sample and also in the two most important sub-samples of the data, namely intraOECD relations with 2,789 observations and OECD–non-OECD relations with 1,446 observations. Both bilateral stocks of outward FDI and bilateral country size are higher on average for intra-OECD relations. Consistent with stylized facts, OECD countries are more similar in terms of both size (SIMI) and relative factor endowments (ΔSK). Overall, about three times as many BITs are signed and ratified between the OECD and non-OECD countries as between members of the current OECD economies in our sample. Most existing intra-OECD BITs were negotiated and signed in the early 1990s between old members and then– non-members that have since joined the OECD, for example the Czech Republic, Hungary, Poland, or the Slovak Republic. Only a smaller portion of these treaties has been concluded between two states belonging to the OECD at the date of signing the BIT, for example the United States and Turkey in 1990, Mexico and Spain in 1995, and Mexico and Switzerland in 1995. Hence, the typical BIT is negotiated between a highly developed country and a less-developed partner. Table 3 provides information on ratified/signed BITs between 1982 and 1997 for all countries in the regression sample. The Appendix supplies further details on the covered 19 home and 54 host countries. Of course, not all BITs come into force immediately, but rather there are signed BITs which are still ineffective. On average, about 66% of the BITs signed by the countries in the sample indeed entered into force within this 16-year period. Notably, we define the OECD as of 2003 in Table 3. With this database at hand, we are able to estimate the impact of ratifying a BIT on bilateral outward FDI. Moreover, we can investigate whether there are significant anticipation effects from signing a BIT only. Our choice of controls will be guided by the recent general equilibrium models of trade and FDI summarized above. Hence, the key determinants will be related to economic size, relative factor endowments, and impediments to trade and FDI.

the impact of bilateral investment treaties on foreign direct investment 259

table 2. descriptive statistics Variable

Mean

Std. Dev.

Min

Max

Full sample (4,235 observations) ln real outward FDI stocks (dependent) 5.58 27.84 ΣGDPijt SIMIijt −1.67 16.82 ΔSKijt 30.06 DISTij·D+·ΔSKijt 38.76 ΔGDPijt·D+·ΔSKijt 577.06 ΣGDPijt·D+·ΔSKijt ΣGDPijt·D−·ΔSKijt −106.39 0.20 BITSijt (BIT signed) BITRijt (BIT ratified) 0.15

2.69 1.07 1.06 26.34 84.26 73.35 590.60 287.31 0.40 0.36

−4.31 23.35 −6.27 −69.80 0.00 −150.75 0.00 −1,885.48 0.00 0.00

11.60 30.13 −0.69 94.40 544.40 389.86 2,603.92 0.00 1.00 1.00

Intra-OECD sample (2,789 observations) ln real outward FDI stocks (dependent) ΣGDPijt SIMIijt ΔSKijt DISTij·D+·ΔSKijt ΔGDPijt·D+·ΔSKijt ΣGDPijt·D+·ΔSKijt ΣGDPijt·D−·ΔSKijt BITSijt (BIT signed) BITRijt (BIT ratified)

6.00 27.92 −1.54 8.43 44.63 20.23 394.79 −158.09 0.11 0.10

2.73 1.03 0.96 24.77 100.43 53.73 499.78 341.51 0.32 0.30

11.60 −4.23 25.38 30.13 −6.27 −0.69 83.20 −69.80 0.00 544.40 376.27 −150.75 0.00 2,280.02 0.00 −1,885.48 0.00 1.00 0.00 1.00

OECD-to-non-OECD sample (1,446 observations) ln real outward FDI stocks (dependent) ΣGDPijt SIMIijt ΔSKijt DISTij·D+·ΔSKijt ΔGDPijt·D+·ΔSKijt ΣGDPijt·D+·ΔSKijt ΣGDPijt·D−·ΔSKijt BITSijt (BIT signed) BITRijt (BIT ratified)

4.77 27.69 −1.94 33.00 1.95

2.43 1.11 1.18 21.24 12.07

−4.31 23.35 −6.20 −17.00 0.00

10.13 29.72 −0.69 94.40 141.42

74.49 928.63 −6.68 0.35 0.25

90.86 593.57 41.74 0.48 0.43

−120.45 0.00 −448.32 0.00 0.00

389.86 2,603.92 0.00 1.00 1.00

Note: OECD countries are defined according to 2003 membership.

260 peter egger and michael pfaffermayr table 3. new bit s from 1982 to 1997 Country

Algeria Argentina Australia Austria Belgium-Luxembourg Brazil Bulgaria Canada Chile China Colombia Costa Rica Czech Republic Denmark Egypt Finland France Germany Greece Hong Kong Hungary Iceland India Indonesia Israel Italy Japan Korea Kuwait Malaysia Mexico Morocco Netherlands Norway Panama

New BITs signed

New BITs ratified

overall

thereof with OECD

overall

5 38 12 19 27 10 33 17 29 71 3 3 43 30 21 27 48 66 18 11 44 1 11 18 12 43 3 38 15 35 2 15 44 15 6

4 20 3 6 4 9 18 4 13 24 2 3 22 6 10 5 4 4 4 11 22 0 6 10 3 5 1 12 7 9 2 12 5 4 6

3 25 11 14 14 10 20 9 7 53 3 0 31 22 7 20 29 41 12 6 33 1 3 10 5 24 3 28 7 15 2 9 36 14 5

thereof with OECD 2 16 3 6 4 9 14 4 4 22 2 0 20 6 6 5 4 5 4 6 20 0 3 9 3 3 1 10 5 11 2 8 5 4 5 Continued

the impact of bilateral investment treaties on foreign direct investment 261

Country

Philippines Poland Portugal Romania Russia Saudi Arabia Singapore Slovak Republic Slovenia South Africa Spain Sweden Switzerland Thailand Turkey Ukraine United Arab Emirates United Kingdom United States Venezuela

New BITs signed

New BITs ratified

overall

thereof with OECD

overall

thereof with OECD

10 53 19 53 32 1 6 25 9 9 37 25 43 14 36 26 8 69 37 15

8 23 4 21 22 1 2 17 5 8 7 4 6 6 15 15 6 5 4 10

4 50 9 40 11 1 4 23 3 9 27 22 32 7 20 12 3 58 21 7

4 23 3 18 10 1 1 16 2 8 5 3 5 4 11 7 2 4 4 4

Notes: (i) Data are from the International Centre for Settlement of Investment Disputes. (ii) OECD economies as of 2003 are printed in bold face.

B. Empirical Results In this section, we estimate the following basic specification: Fijt = β1ΣGDPijt+β2SIMIijt+β3ΔSKijt+β4DISTij·D+·ΔSKijt+β5ΔGDPijt· D+·ΔSKijt +β6ΣGDPijt· D+·ΔSKijt+β7–ΣGDPijt·D−·ΔSKijt+β8ΒITSijt+β9BITRijt+μij+λt+εijt,

where Fijt denotes the log of real stocks of outward FDI of home country i in host country j for year t. The explanatory variables are defined in Table 1. In all regressions, we include fixed country-pair effects (μij) to correct for omitted, time-invariant, geographical or cultural variables, and fixed time effects (λt) to control for omitted time-variant effects that affect all country-pairs in the same way. The usual error term is denoted by εijt.

262 peter egger and michael pfaffermayr

Table 4 reports the estimated coefficients measuring the impact of the explanatory variables on outward FDI. We use tertiary school enrollment as a proxy for a country’s skilled to unskilled labor ratio. In models I and II, we include only BITR, that is, β8=0; in models III and IV, we consider additional anticipation effects captured by the coefficient of BITs. Models I and III that exclude the distance interaction term are closest to the specifications estimated by Markusen and Maskus (2002). Models II and IV both include ΔSKijt and the interaction term DISTij·D+·ΔSKijt as suggested by Egger and Pfaffermayr (2004b). In all cases, we exclude extreme outliers defined as falling in the outer two percentiles of the distribution of residuals and we report only heteroskedasticity-robust tstatistics. The country-pair and time effects always enter significantly and the random effects model is rejected firmly when compared to its fixed effects counterpart. The high adjusted R2 statistics corroborate the appropriateness of our specification. Moreover, the jointly significant size variables underscore the importance of horizontal FDI, whereas the parameters of the skill difference variable and the interaction terms indicate that vertical MNEs are also present. Hence, the signs of these knowledge-capital model coefficients are consistent with the theoretical predictions and with Markusen and Maskus (2002). Regarding the impact of BITs, several findings are worth emphasizing. First, in all the models in Table 4, the coefficient of BITR is significantly different from zero and it ranges from 0.21 to 0.26. Hence, the estimated impact is relatively unaffected by the choice of specification.8 Second, we find weak evidence for an anticipation effect from the coefficients of BITs in models III and IV. Third, the estimated effect of signing a treaty (BITs) is always smaller than that of ratifying it (BITR). Since FDI stocks are measured in logs, we must transform the BITs effect and its standard deviation to percentage figures. Following Kennedy (1983) and van Garderen and Shah (2002), the overall effect of implementing a treaty is calculated as 100·exp(β9–0.5Var(β9)–1). According to Table 4, this estimated impact amounts to about 30% in models I and II. However, the point estimate of the combined effect of BITS and BITR is even higher. Nonetheless, since the coefficient for BITs is not significant, the parsimonious models I and II are preferred. In Table 5, we check the robustness of our findings. First, we include additional control variables to ensure that the estimated BITs coefficients do not include effects from other omitted variables. We take account of European Union (EU) and North American Free Trade Agreement (NAFTA) membership for two reasons: Daude et al. (2002) argue that FDI regressions should consider the relevance of trading bloc effects for MNEs, and both the EU and NAFTA are multilateral investment treaties that might generate different effects from their

8. UNCTAD (1998) and Hallward-Driemeier (2003) do not identify any important, significant effect of BITs, probably because UNCTAD relies on cross-section analysis and Hallward-Driemeier focuses on developing economies and FDI flows rather than stocks.

the impact of bilateral investment treaties on foreign direct investment 263

table 4. the impact of bit s Explanatory variables

Model I

Model II

Model III

Model IV

ΣGDPijt

3.692∗∗∗

3.718∗∗∗

3.639∗∗∗

3.666∗∗∗

(11.92) 0.203 (1.17) −0.004∗∗∗ (−3.88) −0.000∗∗∗ (−3.17) −0.000∗∗ (−2.51) -

(12.00) 0.230 (1.32) 0.052 (1.17) −0.003 (−1.07) −0.004∗∗∗ (−3.42) −0.002 (−1.35) −0.003∗ (−1.77) -

(11.72) 0.212 (1.23) −0.004∗∗∗ (−3.93) −0.000∗∗∗ (−3.28) −0.000∗∗∗ (−2.57) 0.126

(11.80) 0.239 (1.38) 0.051 (1.16) −0.003 (−1.04) −0.004∗∗ (−3.47) −0.002 (−1.35) −0.003∗ (−1.75) 0.124

0.264∗∗∗ (2.62) 30.231∗∗ (2.32) 4,235 0.97 329.80 0.00

0.262∗∗∗ (2.60) 29.914∗∗ (2.31) 4,235 0.97 418.28 0.00

(1.62) 0.214∗∗ (2.05) 13.223 (1.51) 23.521∗ (1.83) 4,235 0.97 269.77 0.00

(1.60) 0.212∗∗ (2.03) 13.050 (1.49) 23.302∗ (1.82) 4,235 0.97 346.33 0.00

88.08 0.00 29.56 0.00

87.21 0.00 28.54 0.00

87.02 0.00 29.29 0.00

86.15 0.00 28.32 0.00

SIMIijt ΔSKijt DISTij·D+·ΔSKijt ΔGDPijt·D+·ΔSKijt ΣGDPijt·D+·ΔSKijt −ΣGDPijt·D−·ΔSKijt BITSijt (BIT signed) BITRijt (BIT ratified) BITS-effect in % (Kennedy, 1981) BITR-effect in % (Kennedy, 1981) Observations Adjusted R2 Hausman test: p-value F-tests: Country-pair effects p-value Time effects p-value

Notes: (i) The figures below the coefficients in parentheses are t-statistics. (ii) The symbols ∗∗∗, ∗∗, and ∗ denote significance levels of 1%, 5%, and 10%, respectively. (iii) The t-statistics in bold face for the BITs and BITR coefficients are based on the approximated standard errors suggested by van Garderen and Shah (2002).

264 peter egger and michael pfaffermayr table 5. sensitivity analysis BITS BITR (BIT signed) (BIT ratified) Including additional controls(a) Model I Model II Model III Model IV

– – – – 0.139∗ (1.87) 0.138∗ (1.85)

0.254∗∗∗ (4.46) 0.209∗∗ (2.31) 0.156∗ (1.67) 0.155∗ (1.66)

Secondary instead of tertiary school enrollment plus additional controls(a) Model I Model II Model III Model IV

– – – – 0.140∗∗ (2.02) 0.114 (1.60)

0.204∗∗∗ (2.61) 0.221∗∗∗ (2.81) 0.147∗ (1.77) 0.173∗∗ (2.05)

As in Table 4, but excluding transition countries(b) Model I Model II Model III Model IV

– – – – 0.100 (1.51) 0.097 (1.20)

0.218∗∗ (2.15) 0.220∗∗ (2.13) 0.182∗∗ (2.41) 0.180∗ (1.67)

BITS BITR (BIT signed) (BIT ratified) Real GDP/capita instead of tertiary school enrollment plus additional controls(a) Model I Model II Model III Model IV

– – – – 0.143∗∗ (2.55) 0.140∗∗ (2.50)

0.259∗∗∗ (4.52) 0.254∗∗∗ (4.46) 0.204∗∗∗ (3.21) 0.201∗∗∗ (3.18)

As in Table 4, but excluding Hungary, Mexico, and Poland Model I Model II Model III Model IV

– – – – 0.070 (0.79) 0.069 (0.77)

0.172∗∗ (2.36) 0.171∗∗ (2.34) 0.144∗ (1.66) 0.142∗ (1.64)

As in Table 4, but including an interaction term of BITR with DC-LDC(c) Model I Model II

– – – –

0.209∗ (1.81) 0.203∗ (1.75)

Endogenous BITR plus additional controls(d) Model I Model II

– – – –

0.252∗∗ (2.37) 0.228∗∗ (2.14)

Notes: (i) The figures below the coefficients in parentheses are t-statistics. (ii) The symbols ∗∗∗, ∗∗, and ∗ denote significance levels of 1%, 5%, and 10%, respectively. (iii) The superscript (a) indicates that additional controls are EU and NAFTA dummies, a dummy capturing the political change in the CEEC after the fall of the iron curtain (1 after 1989 for CEEC and 0 else), the average corporate tax rate, and three variables capturing infrastructure endowments in host countries (road network in km, telephone main lines per 1000 people, and electricity production in kWh).

the impact of bilateral investment treaties on foreign direct investment 265

(iv) The superscript (b) indicates that Bulgaria, Czech Republic, Hungary, Poland, Romania, Russia, Slovak Republic, Slovenia, and Ukraine are excluded from the sample. (v) In the specifications denoted by superscript (c), the main BITR effect is the basis and a dummy capturing OECD outward FDI to non-OECD countries is interacted with BITR and included as well. The corresponding parameter estimates are 0.14 with a t-statistic of 1.05 and 0.15 with a t-statistic of 1.10, respectively. (vi) The Difference-in-difference matching approach under superscript (d) proceeds in three steps. In the first step, a probit model is estimated for each year with all exogenous variables plus two outside instruments, the home and host country overall number of BITs, excluding the respective bilateral agreement. In the second step, the Mills ratio is computed for each year. In the third step, a fixed effects model is estimated, including the Mills ratio derived in the second step. (vii) In the matching approach under superscript (d), we include the same controls as listed in (a) except for the potentially endogenous EU and NAFTA dummies and the corporate tax rate. An alternative specification, which also excludes the three infrastructure variables mentioned in (a), yields very similar but slightly higher BITR parameter estimates. We obtain virtually the same results from an alternative model that does not use any outside instruments.

bilateral counterparts.9 As suggested by Hallward-Driemeier (2003), we account for change in the political system in Central and Eastern European (CEE) countries in the 1990s by introducing a dummy variable equal to 1 in 1990 or later if the host country belongs to this group. Furthermore, we include the host country’s corporate tax rate and three host country infrastructure variables, namely, telephone mainlines per 1,000 people, the size of the road network in kilometers, and the electricity supply in kilowatt hours. Tables A1 and A2 in the Appendix provide details on data sources and summary statistics. To sum up the results, the estimated BITR parameters do not change much in the augmented specifications and are slightly lower compared to their counterparts in Table 4. However, the BITs parameters are now estimated more precisely. Second, we run this augmented specification but use real GDP per capita instead of tertiary school enrollment to measure the skill difference. This change does not alter the point estimates of BITs and BITR substantially, but the t-statistics are higher than in Table 3 because we now have more observations. Third, we estimate the augmented specification measuring the skill difference by secondary rather than tertiary school enrollment. The corresponding BITs and BITR estimates are somewhat smaller than in Table 4, but BITR is always positive and significant. Fourth, we exclude the new OECD members, namely, Hungary, Mexico, and Poland, from the sample to investigate to what extent the

9. We do not report these effects in Table 5. The EU dummy tends to be positive but insignificant; the NAFTA effect is almost always negative and significant, indicating that NAFTA favored trade at the expense of FDI.

266 peter egger and michael pfaffermayr

estimation results are driven by these countries. In this regression, we obtain a slightly smaller BITR point estimate, but it remains positive and significant. Fifth, we exclude all of the transition countries in the sample, namely, Bulgaria, Czech Republic, Hungary, Poland, Romania, Russia, Slovak Republic, Slovenia, and Ukraine, and find the BITs and BITR parameter estimates to be almost unchanged. Sixth, we construct an interaction term between BITR and a dummy variable equal to 1 for FDI between OECD and non-OECD countries. We re-estimate models I and II including both the BITR main effect and this interaction term.10 If the parameter estimate of the interaction term were significant, ratified BITs would exert a different impact on FDI between OECD and nonOECD countries than on intra-OECD FDI. Most of the intra-OECD BITs were signed and ratified between old and new OECD members before the latter joined the OECD. Accordingly, the coefficient of the interaction term indicates whether fast growing, recently joining OECD members are affected differently by BITs than are old OECD members. The parameter estimates for this interaction term are 0.14 and 0.15, respectively. Since the corresponding t-statistics are only 1.05 and 1.10 and the coefficients of BITR in models I and II do not change much, we conclude that no differences of this kind exist. Finally, we check for possible endogeneity of the BITR effect in model I of Table 4. The fixed effects estimator provides an unbiased estimate of BITR only if the selection into treatment, that is, ratifying a BIT, is dominated by timeinvariant variables (Wooldridge 2002). If there is endogenous selection and the time-varying treatment is correlated with time-variant unobservables, the fixedeffects estimate is biased. Since we have no fixed-effects panel estimator for endogenous dummy variables available, we follow Blundell and Costa Dias (2002) and apply a matching estimator in a difference-in-difference framework. The proposed estimator compares the change in bilateral FDI stocks between the periods before and after a BIT was ratified with the change in a properly defined control group. Hence, we obtain the average treatment effect of the treated, which is defined as the counterfactual impact of abolishing an earlier-ratified BIT in the presence of self-selection. Furthermore, by taking the difference of differences approach, we eliminate all unobserved time-invariant effects. The estimation procedure involves four steps. First, we calculate averages of log FDI stocks for the two-year period before and after a BIT has been ratified. The second two-year period includes the ratification year itself. Second, we calculate the change in log FDI between these two periods for both the country pairs that ratified a BIT and the control group. The control group is constructed for each year separately and consists of all country pairs that did not ratify a BIT until the end of the two years defined as the ratification period. Third, we estimate a probit model using the controls of model I, with GDP per capita as a

10. Due to high correlations, BITs and a similar interaction term cannot both be used.

the impact of bilateral investment treaties on foreign direct investment 267

measure of relative factor endowments, to obtain more observations. In addition, we apply several external controls, that is, the home and host country overall number of BITs, excluding the respective bilateral agreements, the telephone main lines per 1,000 people of the importing country, the dummy variable for CEE countries in the 1990s, and time dummies.11 Fourth, based on the scores derived in step three, we apply a simple one-to-one matching estimator and a radius matching estimator. The corresponding standard errors are derived by bootstrapping with 100 replications. The additional controls in the probit model are highly significant but they do not differ significantly between the group of countries that have ratified BITs and the matched control group. Thus, the balancing property is satisfied and, conditional on the explanatory variables in the probit regression, all country pairs are equally likely to ratify a BIT. The estimated BITR-effect remains highly significant, and it is even slightly higher when the endogeneity of BITR is taken into account. The point estimate equals 0.457 with a t-statistic of 2.44 in the oneto-one matching and 0.373 with a t-statistic of 4.28 in the radius matching. The matching estimates indicate that the standard fixed-effects estimates in Tables 4 and 5 may be downward biased. Hence, they represent lower bound estimates of the impact of BITs on FDI stocks. To summarize, our findings of a positive and significant impact of BITs on bilateral FDI stocks are robust to the inclusion of infrastructure variables, corporate tax rates, and trading bloc effects, to the use of alternative measures of relative factor endowment differences, to the exclusion of new OECD members, and to an endogenous treatment of BITs. Furthermore, BITs do not affect FDI between OECD and non-OECD economies differently than intra-OECD FDI. By taking explicit account of the endogeneity of BITs, we estimate a slightly higher effect of the counterfactual abolishing of an earlier-ratified BIT.

conclusion In the last four decades, the number of BITs signed has been large and the pace is accelerating. We investigate whether these treaties are signed only for political reasons or whether they remove important economic obstacles. To do so, we concentrate on FDI and analyze whether bilateral real outward FDI stocks rise as new treaties are signed or implemented. Hence, we hypothesize that bilateral investment treaties reduce barriers to FDI. Embedded in alternative specifications of the knowledge-capital model, we estimate the ratification effect of BITs. In addition, we look at potential anticipation effects after signing and before

11. Hallward-Driemeier (2003) uses similar variables to instrument her dummy variable for BITs in a two-stage least-squares framework.

268 peter egger and michael pfaffermayr

ratifying a BIT. We use the largest available set of bilateral outward FDI stock data from the OECD and a comprehensive data set on bilateral investment treaties from the World Bank. We control for the importance of time-invariant and common-cycle fixed effects as well as for the potential influence of outliers and use heteroskedasticity-robust estimates. We find an overall BIT ratification effect on FDI of about 30% in the preferred specifications. Moreover, if it has an impact at all, signing a treaty exerts a significantly lower impact on real FDI stocks. Our results are robust to alternative measures of relative factor endowment differences, to the impact of trading blocs such as EU or NAFTA, and to infrastructure endowments. In all our estimated specifications, BITs exert a positive and significant effect on real stocks of outward FDI, with a lower bound of 15%.

the impact of bilateral investment treaties on foreign direct investment 269

appendix

List of Countries in the Sample A. Home countries: Austria, Australia, Canada, Denmark, Finland, France, Germany, Iceland, Italy, the Republic of Korea, Netherlands, New Zealand, Norway, Poland, Portugal, Sweden, Switzerland, United Kingdom, and United States. B. Host countries: Algeria, Argentina, Australia, Austria, Belgium-Luxembourg, Brazil, Bulgaria, Canada, Chile, People’s Republic of China, Colombia, Costa Rica, Czech Republic, Denmark, Egypt, Finland, France, Germany, Greece, Hong Kong (China), Hungary, Iceland, India, Indonesia, Ireland, Israel, Italy, Japan, Republic of Korea, Kuwait, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Norway, Panama, Philippines, Poland, Portugal, Romania, Russian Federation, Saudi Arabia, Singapore, Slovakia, Slovenia, South Africa, Spain, Sweden, Switzerland, Thailand, Turkey, Ukraine, United Arab Emirates, United Kingdom, United States, and Venezuela.

table a1. data sources Variable

Source

Bilateral stocks of outward FDI in current US$ Investment deflators in constant 1995 US$ GDP in constant 1995 US$

Foreign Direct Investment Statistics Yearbook 2002 (OECD) World Development Indicators (World Bank) World Development Indicators (World Bank) World Development Indicators (World Bank) World Development Indicators (World Bank) World Development Indicators (World Bank) Own calculations of greater circle distance

Population Secondary school enrollment share Tertiary school enrollment share Bilateral distance in miles Bilateral investment treaties signed (BITs)

World Bank Continued

270 peter egger and michael pfaffermayr table a1. data sources (cont’d...) Variable

Source

Bilateral investment treaties ratified (BITR) Host country corporate tax rates Host country telephone main lines (per 1,000 people) Host country roads, total network (km) Host country electricity production (kWh)

World Bank World Development Indicators (World Bank) World Development Indicators (World Bank) World Development Indicators (World Bank) World Development Indicators (World Bank)

table a2. descriptive statistics for variables used in the sensitivity analysis Variable

Mean

Std. Dev.

Minimum

Maximum

Endowment variables (GDP/capita) ΔSKijt

0.83

1.29

−2.79

4.86

DISTij·D+·ΔSKijt

0.92

2.20

0.00

19.87

ΔGDPijt·D+·ΔSKijt

1.75

3.46

−16.62

18.33

ΣGDPijt·D+·ΔSKijt

26.34

32.32

0.00

131.91

ΣGDPijt·D−·ΔSKijt

−3.35

8.05

−74.64

0.00

Endowment variables (secondary school enrolment) ΔSKijt

15.48

28.11

−55.57

102.60

DISTij·D+·ΔSKijt

35.06

71.62

0.00

513.88

ΔGDPijt·D+·ΔSKijt

28.74

61.46

−234.59

342.48

ΣGDPijt·D+·ΔSKijt

560.14

620.86

0.00

2,747.15

ΣGDPijt·D−·ΔSKijt

−132.91

266.70

−1506.17

0.00

the impact of bilateral investment treaties on foreign direct investment 271

Variable

Mean

Std. Dev.

Minimum

Maximum

Dummy variables EU membership NAFTA membership Central and Eastern Europe after 1989

0.05 0.00

0.21 0.07

0.05

0.22

0.00 0.00 0.00

1.00 1.00 1.00

Other variables Host country corporate tax rates Host country telephone main lines (per 1,000 people) Host country roads, total network (1,000 km) Host country electricity production (in billion kWh)

30.98

14.67

0.00

76.90

340.62

226.38

3.00

1,147.00

372.04

977.80

0.00

6,300.00

263.00

587.00

2.25

3,560.00

Note: The number of observations is 3,904.

references Blonigen, Bruce A. and Ronald B. Davies. (2003). “The effects of bilateral tax treaties on US FDI activity,” International Tax and Public Finance, forthcoming. Blonigen, Bruce A., Ronald B. Davies and Keith C. Head (2003). “Estimating the knowledge-capital model of the multinational enterprise: comment,” American Economic Review, 93, pp. 980–994. Blundell, Richard and Monica Costa Dias (2002). “Alternative approaches to evaluation in empirical microeconomics,” Portuguese Economic Journal, 1, pp. 91–115. Carr, David, James R. Markusen and Keith E. Maskus (2001). “Estimating the knowledge-capital model of the multinational enterprise,” American Economic Review, 91, pp. 693–708. —— (2003). “Estimating the knowledge-capital model of the multinational enterprise: reply,” American Economic Review, 93, pp. 995–1001. Chisik, Richard and Ronald B. Davies (2004). “Gradualism in tax treaties with irreversible foreign direct investment,” International Economic Review, forthcoming. Dagan, Tsilla (2000). “The tax treaties myth,” New York University Journal of International Law and Politics, 32, pp. 939–996.

272 peter egger and michael pfaffermayr Daude, Christian, Ernesto Stein and Eduardo L. Yeyati (2002). “Regional integration and the location of FDI,” Working Paper No. 492 (Washington, D.C.: Inter-American Development Bank). Davies, Ronald B. (2004). “Tax treaties, renegotiations, and foreign direct investment,” Economic Analysis and Policy, forthcoming. Drabek, Zdenek and Warren Payne (2002). “The impact of transparency on foreign direct investment,” Journal of Economic Integration, 17, pp. 777–810. Egger, Peter and Michael Pfaffermayr (2004a). “Trade, multinational sales and FDI in a three-factors model,” Review of International Economics, forthcoming. —— (2004b). “Distance, trade, and FDI: a Hausman-Taylor SUR approach,” Journal of Applied Econometrics, 19, pp. 227–246. Ekholm, Karolina (1998). “Proximity advantages, scale economies, and the location of production,” in Braunerhjelm Pontus and Ekholm Karolina, eds., The Geography of Multinational Firms (Dordrecht: Kluwer Academic Publishers). Hallward-Driemeier, Mary (2003).“Do bilateral investment treaties attract FDI? Only a bit . . . and it might bite,” World Bank Policy Research Working Paper Series No. 3121 (Washington, D.C.: World Bank). Hausman, Jerry A. (1978). “Specification tests in econometrics,” Econometrica, 46, pp. 1251–1271. Helpman, Elhanan (1984). “A simple theory of international trade with multinational corporations,” Journal of Political Economy, 92, pp. 451–472. Helpman, Elhanan and Paul R. Krugman (1985). Market Structure and Foreign Trade (Cambridge, MA: MIT Press). Hoekman, Bernard and Kamal Saggi (2000). “Multilateral disciplines for investment-related policies?” in Paolo Guerrieri and Hans-Eckart Scharrer, eds., Global Governance, Regionalism, and the International Economy (Baden-Baden: Nomos Verlagsgesellschaft). Kennedy, Peter E. (1981). “Estimation with correctly interpreted dummy variables in semilogarithmic equations,” American Economic Review, 71, p. 801. Markusen, James R. (1984). “Multinationals, multi-plant economies, and the gains from trade,” Journal of International Economics, 16, pp. 205–226. Markusen, James R. and Keith E. Maskus (2002). “Discriminating among alternative theories of the multinational enterprise,” Review of International Economics, 10, pp. 694–707. Markusen, James R. and Anthony J., Venables (1998). “Multinational firms and the new trade theory,” Journal of International Economics, 46, pp. 183–203. —— (2000). “The theory of endowment, intra-industry, and multinational trade,” Journal of International Economics, 52, pp. 209–234. Maskus, Keith E. (2000). “Intellectual property rights and foreign direct investment,” Policy Discussion Paper No. 0022 (Adelaide: University of Adelaide, Centre for International Economic Studies). United Nations Conference on Trade and Development (UNCTAD) (1998). Bilateral Investment Treaties in the Mid 1990s (New York: United Nations). United Nations (2000). Bilateral Investment Treaties: 1959–1999. (New York: United Nations). van Garderen, Kees J. and Chandra Shah (2002). “Exact interpretation of dummy variables in semilogarithmic equations,” Econometrics Journal, 5, pp. 149–159. Wooldridge, Jeffrey M. (2002). Econometric Analysis of Cross Section and Panel Data (Cambridge, MA: MIT Press).

9. new institutional economics and fdi location in central and eastern europe∗ robert grosse and len j. trevino introduction The goal of this chapter is to demonstrate that institutions matter in the context of foreign direct investment (FDI) in the transitional economies of Central and Eastern Europe (CEE). We employ the new institutional economics (NIE) as a theoretical foundation and apply related concepts in an examination of FDI activity that may respond to institutional development in CEE. Within this context, FDI is expected to respond to country-level macroeconomic, microeconomic, and institutional changes, especially institutional factors that reduce (increase) uncertainty and/or costs related to long-term capital investments. This approach follows Rumelt, Schendel, and Teece (1991) in utilizing tools from both economics and strategic management to facilitate understanding of FDI decision-making. It is well understood that firms entering a new market must adapt their overall strategies to environmental conditions in the host country (Hymer 1976; Kindleberger 1969). Along these same lines, as governments in transitional economies battle for larger shares of global FDI flows, there is increasing evidence that they have adapted their institutional environments1 in order to attract inward FDI (Meyer 2001). Although institutional hurdles2 exist in all countries, for emerging markets and economies in transition, the evolution of market-based institutions is a critical hurdle (Clague 1997; Meyer 2001). Economic transition involves replacing one governing framework with another and, in the case of CEE, this transition involved replacing a socialist system with an institutional framework consistent with a market economy. Since efficient markets depend ∗ This chapter was reprinted with permission from the Management International Review. The chapter was originally published as New institutional economics and FDI location in central and eastern Europe,” 45 Management International Review 123 (2005). The authors are grateful to Michael A. Hitt and Lorraine Eden for comments and suggestions on an earlier draft of this chapter. In addition, the authors thank Taisa Minto and Sangit Rawlley for their research assistance and Doug Thomas for his statistical help. 1. The institutional environment is defined as the set of fundamental political, social, and legal ground rules that establishes the basis for production, exchange, and distribution (Davis and North 1971). 2. Meyer (2001) measures institutional reform by a composite index based on the European Bank for Reconstruction and Development transition indices, which reflect an expert evaluation of institutional reforms. In our study, institutional factors are measured independently by several of our independent variables.

274 robert grosse and len j. trevino

on market-based supporting institutions (North 1990), all of the CEE countries have pursued transition strategies from state control toward open markets. Some of the adjustments associated with the installation of a market economy have been implemented more rapidly than anticipated, while others have lagged, partially determined by the manner in which state socialism disintegrated in a given country.3 Within the broad theoretical framework of institutional theory, normative, cultural-cognitive, and regulatory systems have all been identified as being useful approaches. The normative approach specifies how things should be done, and normative systems define goals or objectives but also suggest legitimate means by which to pursue them (Scott 2001). The cultural-cognitive approach recognizes that internal interpretive processes are shaped by external cultural frameworks. In fact, it has been proposed that cultural dimensions are the cognitive frameworks in which social interests are defined (Douglas 1982). The regulatory framework involves the capacity to establish rules, to determine who has conformed to such rules, and, as necessary, to manipulate sanctions to influence behavior (Scott 2001). Although each of these three dimensions of institutional theory provides a basis for analysis—the cultural-cognitive dimension from an anthropological perspective, the normative dimension from a sociological perspective, and the regulatory dimension from an economic perspective—we focus on the regulatory environment and use the new institutional economics as a theoretical foundation for our study. The new institutional economics focuses on the intersection of institutional environments and firms that results from market imperfections (Harris, Hunter, 3. The transition has moved forward more or less toward market-based institutions depending on whether the government has pursued a more “coordinated” or a more “liberal” policy framework (Hall and Soskice 2001). The former framework more closely parallels the German and Scandinavian institutional systems, and the latter is exemplified by the United States and other Anglo-Saxon countries. The “coordinated” structure tends to produce stronger unions and industry associations, as well as greater collaboration between firms and between firms and government; the “liberal” structure tends to produce greater arm’s-length competition, greater labor mobility between firms, and more focus on explicit contracting between parties. It could easily be argued—but is not pursued here—that the more coordinated policy environment is an easier regime to establish when moving from a communist one, and that moving to a more liberal regime will likely provoke a more difficult initial transition. A reviewer points out that the formerly-communist countries lack a history of developing “associative coordination mechanisms” such as nationwide trade unions and industry associations — and that this lack hinders these countries from moving into the sphere of coordinated market economies. This point ignores, however, the pre-communist historical evolution of countries such as Russia and the Central European nations, which developed much more similarly to Germany and the northern European countries than to the Anglo-Saxon countries. In any event, this is an empirical question.

new institutional economics and fdi location 275

and Lewis 1995). North (1990) posited that institutions provide the rules of the game that structure interactions in societies and that organizational action is bound by these rules. A central focus of NIE is the reduction of costs associated with transactions, in this case those related to FDI decision-making and implementation. This perspective emphasizes that firms incur costs when they undertake business transactions, such as the cost of obtaining information and writing and enforcing contracts. According to NIE, firms are a particular form of organization for administering transactions between one party and another, and firms exist because they may be able to reduce the costs of negotiating and enforcing terms and conditions of exchange relative to market transacting (Coase 1937). Under the managerial-choice approach of NIE (Williamson 1975, 1985; Walker and Weber 1984), managers employ a transaction-cost-economizing calculation when making contracting decisions. North (1990) suggests that the efficiency of political institutions may be measured by how closely an actual political market approximates a zero transaction cost result. In the case of the multinational enterprise (MNE) and FDI decision making, presumably the better the institutional environment is able to approximate zero transaction costs for the foreign investor, the more likely the country is to receive inward FDI flows, ceteris paribus. Thus, one way to understand the intersection of MNE investment strategy and the institutional environment in transitional economies is to analyze the ability of institutions to reduce transaction costs associated with FDI that result from uncertain environments (Hoskisson et al. 2000).4 This requires an investigation of differences in institutional models across countries and their effects on inward FDI flows. In the case of transitional economies, the cost of establishing FDI is high (Meyer 2001). In fact, Estrin, Hughes, and Todd (1997) found that inefficiencies in developing economies led to substantial extra costs and delays for foreign investment. In addition to the normal costs associated with conducting business in a foreign country, foreign firms entering transition economies may face increased levels of uncertainty resulting from high inflation, opaque regulatory environments, underdeveloped judicial and financial systems, and corruption. All of these elements increase the costs associated with FDI in the host country; and according to Coase (1937), when it is costly to transact, institutions matter. Although the new institutional economics has been studied widely in developed countries, there is a dearth of theoretical and empirical research using an institutional framework in emerging economies. Furthermore, because the new institutional economics has been applied primarily in developed markets, less is 4. Interestingly, Madhok (1997) points out that the decision-making perspective should be to raise firm value, rather than just to decrease costs. Thus, looking at government policies and institutional conditions should emphasize the impacts of this environment in altering transaction costs and in changing revenues and risks. All of these impacts are relevant in the FDI decisions in question here.

276 robert grosse and len j. trevino

known about the institutional environment–FDI decision-making interface in emerging economies (Hoskisson et al. 2000). Due to the degree and abruptness of political and economic change in transitional economies, such environments may provide an excellent framework for advancing the empirical study of the new institutional economics. We examine the extent to which institutional reform in CEE has resulted in reduced uncertainty and costs associated with long-term capital investment, ultimately impacting FDI inflows in 13 CEE countries. These countries comprise Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Russia, Slovakia, Slovenia, and the Ukraine. The formerly centrally planned economies of CEE provide a particularly noteworthy region for studying institutional reform and its impact on inward FDI. As evidence, between 1990 and 2000, inward global FDI stock increased at an average rate of about 12.8% per annum. During this same period, inward FDI in developing countries increased at an average annual rate of 21.3%. However, since 1990, inward FDI in CEE increased by an average annual rate of 33.0% (UNCTAD 2001).5 Although the FDI stock in CEE has increased 4,063% between 1990 and 2000, there is high variability among host country recipients6 (see Figure 1). The initial surge in FDI into this region was quite volatile, with absolute declines in 1994 and 1999, and with rapid growth often occurring before and after. The pace of both market-seeking and cost-reducing direct investment slacked off noticeably after the Russian financial crisis of 1998, and has not resumed its initial growth path to date (see Figure 2). Abrupt or incremental changes in institutional reform variables, resulting from the shift from state control to open markets, coupled with significant movements in inward FDI flows, provide another rationale for the present study. A. Literature review and hypotheses 1. New institutional economics and FDI We argue that, while the new institutional economics is not a unified theory, it does provide the basis for theoretical constructs to demonstrate that “institutions matter.” We have identified five such constructs—corruption, regulation, investment treaties, privatization, and political risk—that demonstrate that governmental institutions matter in FDI 5. The CEE growth rate is calculated from 1992 to 2000, since a number of the countries were only created in 1991. If we use the available data on FDI into the region for 1990–2000, the growth rate was 103.5% per year. 6. Fundamental changes in the structure of the global economy help to explain much of these increases. In addition, market reform, changes in the investment climate, and government policies toward investment offer additional explanation. In the past, host country attractiveness has been positively correlated with the country’s overall size or its natural resource endowment. With the onset of “globalization” and multiple linkages among markets in the 1990s, however, host country competitiveness as a manufacturing site and as an export platform have become additional important determinants of attractiveness.

new institutional economics and fdi location 277

figure 1. fdi/gdp (%, in decimal terms) 0.4

Czech Rep Hungary

0.35

Poland Russia

0.3

Brazil

0.25

Germany Malaysia

0.2 0.15 0.1 0.05 0 1992

1995

1999

Source: UNCTAD World Investment Reports

decisions. Evidence that supports these measures of institutional dimensions as affecting FDI decisions will thus support the broader proposition that the new institutional economics offers a useful framework for understanding these decisions, beyond the traditional economic, political, and financial dimensions. CORRUPTION Corruption may be defined as acts in which the power of public office is used for personal gain in a manner that contravenes the rules of the game (Jain 2001); thus, corruption is one element of a country’s institutional environment that may act to deter or attract FDI. The existence of corruption requires three factors to co-exist. First, someone must have discretionary power to design regulations and to administer them. Second, there must be economic rents associated with this power and identifiable groups must be capable of figure 2. fdi flows (in billions of current u.s. dollars) 8 6 4 2 0 −2 Hungary

−4

Poland

−6

Czech Rep

−8

Romania

−10 1990

Russia

−28 1992

1994

Source: UNCTAD World Investment Reports

1996

1998

278 robert grosse and len j. trevino

capturing these rents. Third, the legal/judicial system must offer sufficiently low probability of detection and/or penalty for violation (Jain 2001). Kaufman and Wei (2000) found that there is a positive relationship between the preponderance of bribery in a country and the time that international managers have to spend with bureaucrats. In this context, corruption is analogous to a tax because it raises the cost of doing business (Jain 2001). When corruption is a key factor in the policy environment, these considerations place companies’ FDI at risk or force investors to incur higher costs to reduce this risk. Seen in this light, corruption may affect resource allocation by altering foreign investors’ perception of the host country environment. While corruption is not limited to the formerly communist countries, it has been much more of a factor there than in other settings. Because recent studies have found that increases in host country corruption levels reduce inward FDI (Hines 1995; Wei 2000), we posit that firms will incorporate this variable into their FDI decision-making process, ultimately influencing FDI levels in the host country. Hypothesis 1: The greater the host country’s success in reducing corruption, the lower the foreign investor’s uncertainty and costs associated with long-term capital investment, resulting in increased inward FDI. REGULATIONS ON FDI Foreign investment regulations and changes in those regulations are one element of a country’s institutional environment that may influence inward FDI flows by increasing uncertainty and FDI-related costs. In principle, firms seek full ownership of FDI, no restrictions on profit repatriation or capital controls, adequate protection on technology transfer, and well-defined property rights. In CEE, firms face a complex web of legal rules and procedures, all of which have the potential to increase costs. Prior to the demise of the Soviet Union, FDI into these countries was highly regulated, and generally was limited to minority joint ventures and long-term cooperation agreements. Along these lines, Contractor (1990) found that these styles of government-imposed regulations and performance requirements tended to limit certain types of FDI. Although the degree of controls placed on foreign investors in CEE countries declined significantly during the 1990s, the amount of regulation did differ across countries. We posit that these decreasing regulations may act to reduce uncertainty and costs for foreign investors, ultimately influencing inward FDI flows.

Hypothesis 2: The greater the success in institutional reform in the host country, as proxied by decreased foreign investment regulations, or by increased enterprise reform (European Bank for Reconstruction and Development index of enterprise reform), the lower the foreign investor’s uncertainty and costs associated with long-term capital investment, resulting in increased inward FDI. BILATERAL INVESTMENT TREATIES The adoption of bilateral investment treaties (BITs) can be considered one of the elements of institutional reform that has helped to foster the perception that the formerly centrally planned economies of CEE

new institutional economics and fdi location 279

are moving toward market-based economies. Since World War II, many BITs, under the auspices of the International Monetary Fund (IMF) and the International Bank for Reconstruction and Development (World Bank), have been enacted in order to provide protection for foreign direct investors. BITs generally offer investors additional and higher standards of legal protection and guarantees for foreign investments beyond those offered under national laws. In fact, BITs often have provisions for the avoidance of double taxation of income and capital. Thus, from the perspective of the new institutional economics, a BIT may act to signal a favorable investment climate. During the 1990s, the number of BITs quintupled, rising from 385 at the end of the 1980s to 1,857 at the end of the 1990s (UNCTAD 2000). The number of BITs between developing countries, between developing countries and countries in Central and Eastern Europe, and between Central and Eastern European Countries increased from 63 at the end of the 1980s to 833 by the end of the 1990s. In an investigation of market reform and FDI in Latin America, Trevino et al. (2002) found a significant and positive relationship between the number of host country BITs and inward FDI. In a related study, Heinrich and Konan (2001) found that preferential trade areas with low individual trade costs see increased investment due to a more integrated market. We propose that BITs may act as an instrument for the international protection of foreign investment, thereby reducing uncertainty and FDI-related costs, while at the same time signalling to foreign investors that the host country has undertaken institutional reforms toward building a market economy. This may imply that BITs are playing an increasingly important role in the proliferation of international investment. Hypothesis 3: The greater the number of host country bilateral investment treaties, the lower the foreign investor’s uncertainty and costs associated with long-term capital investment, resulting in increased inward FDI. PRIVATIZATION Privatization of state-owned enterprises is considered one of the fundamental elements of institutional reform driving the development of the transitional economies of CEE (Frydman, Hessel and Rapaczynski 1998). Since state-owned enterprises in CEE have operated inefficiently due to a dearth of entrepreneurial and managerial skills and a lack of efficient capital markets (Meyer 1998), a common policy in developing countries, and one emphasized by all governments in CEE, is the sale of state-owned enterprises to private sector investors (Kogut 1996). Whether the privatized firm is wholly- or jointly-owned, the potential for cost reductions by transferring the foreign firm’s technology and management capabilities to the acquired state-owned enterprise is often present. In related studies of Latin American privatization programs, Devlin and Cominetti (1994) and Hartenek (1995) have concluded that privatization has given foreign companies more opportunities to invest in host Latin American countries (i.e., it has eliminated institutional barriers that constrain FDI).

280 robert grosse and len j. trevino

In general, privatization programs tend to indicate a government’s willingness to allow the private sector to play a larger role in the economy, thus supporting the entry and growth of foreign MNEs as well as domestic firms. These findings suggest that privatization is a private-sector-friendly policy, which offers the potential for cost reductions by MNEs contemplating investment in the host country. Hypothesis 4: The greater the reduction in barriers to FDI in the host country, as proxied by higher levels of host country privatization, the lower the foreign investor’s costs associated with long-term capital investment, resulting in increased inward FDI. POLITICAL RISK Another element of the institutional environment in the host country revolves around its political risk profile. Political risk may be defined as the risk that a sovereign host government will unexpectedly change the institutional environment under which businesses operate (Butler and Castelo Joaquin 1998). The institutional environment adopted creates unique political risks that contribute to FDI-related costs for foreign investors, and the MNE must consider these costs when deciding whether to invest in the host country. From a financial perspective, political risk may alter operating cash flows via discriminatory policies and regulations, leading to unexpected decreases in future cash flows. Because the uncertainty and costs associated with each country’s political risk profile alter the return to the foreign investor (Bailey and Chung 1995), a risk-averse firm, ceteris paribus, will under-allocate capital to host countries in the presence of high political risk.

Hypothesis 5: The greater the success of a country in reducing political risk, the lower the foreign investor’s uncertainty and costs associated with long-term capital investment, resulting in increased inward FDI. 2. New institutional economics and FDI (Control Variables) Additional factors that should help to explain FDI into CEE countries are those that routinely are found to be important in explaining FDI in most countries around the world. We treat these factors as “control variables” in that they do not differentiate CEE investment from that occurring elsewhere, but they still are expected to play an important role in this region. First, consider the level of macroeconomic instability or risk. With the fragile democracies and market systems of the CEE countries, instability of prices, interest rates, exchange rates, and even unemployment would be expected, in comparison with industrialized economies. We measure two types of macroeconomic risk, namely inflation and exchange rate changes, and recognize that they also mark conditions that characterize emerging markets more broadly. INFLATION Inflation generally is caused by excessive growth in the money supply. Under the former socialist regimes of CEE, inflation was often hidden because prices were fixed, and thus did not allow supply and demand to function normally. In CEE countries, studies have shown that economic instability makes FDI particularly challenging for foreign investors (Meyer 2001). A high and/or

new institutional economics and fdi location 281

variable rate of inflation is a sign of internal economic instability and of the host government’s inability to maintain consistent monetary policy. When companies face an institutional environment with high inflation, their capital budgeting and long-term planning become more uncertain. As a result, foreign investors’ capital commitment becomes riskier due to increased uncertainty and costs. In the worst case, they respond by avoiding investment in environments with these conditions. From the MNE’s perspective, high inflation creates uncertainty about the realized net present value stemming from a costly capital commitment. In addition to the uncertain revenues and, by extension, the realized net present value of costly, irreversible (in the short-run) investment in a high-inflation country, inflation also may inhibit export sales from the host country, thus making resource-seeking FDI less attractive. Trevino et al. (2002) found a negative and significant relationship between inflation in Latin American countries and inward FDI flows, while Aspergis and Katrakilidis (1998) found that inflation and inflation uncertainty had a negative and significant effect on FDI flows in Portugal, Spain, and Greece. Central and Eastern Europe has a shorter history of institutional reform than that occurring in either of the aforementioned studies. In addition, there is sufficient cross-country variability in inflation in CEE to justify an examination of its effect on FDI inflows in the region. Hypothesis 6: The greater the success of government in controlling inflation, the lower the foreign investor’s uncertainty and costs associated with long-term capital investment, resulting in increased inward FDI. CURRENCY VALUATION Another potential impediment to FDI is the absence of a stable, well-accepted currency. The possibility of unanticipated depreciation in the host country exchange rate makes long-term planning difficult and increases the risk associated with long-term investment. Currency devaluation and volatility may result from economic or political upheaval, and foreign investors must incur costs to prevent transaction and translation losses when host country currencies depreciate. If they believe that depreciation will continue after they enter the host country, they may conclude that the costs will be too high to justify their investment. Thus, ceteris paribus, foreign investors would be expected to undertake FDI in countries whose currencies are expected to maintain the value of their earnings. Another argument suggests that currency overvaluation is an example of market disequilibrium, and a currency may be defined as undervalued when, at the prevailing rate of exchange, production costs for tradable goods are on average lower than in other countries. In this case, there is an incentive to locate production of internationally traded commodities in countries with undervalued currencies and to purchase production capacity with overvalued foreign exchange. Thus, the devalued host country currency may attract foreign investors who wish to acquire “inexpensive” local assets. This second line of reasoning is consistent

282 robert grosse and len j. trevino

with recent literature (Grosse and Trevino 1996; Klein and Rosengren 1994; Froot and Stein 1991) in developed countries. If this argument holds in the transitional economies of CEE, it may be that foreign investors possess the international experience to overcome the increased costs of operating in these markets. Since exchange rate devaluation creates both problems and opportunities for MNEs, requiring them to incur costs to manage the risks inherent in a devaluing currency, but also presenting the opportunity of acquiring host country production facilities with overvalued foreign exchange, we develop this argument as an empirical question. Hypothesis 7: The greater the depreciation of the host country’s currency, the higher (lower) the host country’s inward FDI. MARKET SIZE Since country size or purchasing power has been widely used to explain global patterns of FDI, we employ it as a control variable in our model. Even though gross domestic product often declined by more than 20% per annum during the early transition phase in CEE, Western firms increasingly desire access to local markets in order to expand their businesses. This suggests that more FDI will flow into larger countries, which have higher disposable incomes, and which have greater consumption levels. Evidence from recent studies comparing FDI flows to different emerging economies has been relatively consistent. Several studies in CEE found that market-seeking motives are the principal reasons why investors undertake FDI there (Svetlicic and Rojec 1994; OECD 1994; Marinov and Marinova 1999). Trevino et al. (2002) used gross domestic product (GDP) as a surrogate for market size and found it to be highly significant in explaining FDI into Latin American countries. Similarly, Jermakowicz and Bellas (1997) concluded that inward FDI in CEE countries made to gain access to local markets is positively correlated with the population of the host country and their purchasing power. In a study of FDI in China, Wei (2000) found that GDP growth strongly influenced FDI there. The UN (UNCTAD 1998, 139) found that across all emerging markets, for separate analyses in three five-year periods between 1980 and 1995, nominal GDP was a strong positive factor explaining FDI into those countries. Based on the broad literature on FDI location, we expect that:

Hypothesis 8: The larger the host country market size, as indicated by gross domestic product, the greater the host country inward FDI.7

7. Another variable that may help to explain FDI into CEE is the amount of existing trade (imports) between countries in that region and home countries of MNEs. Higher levels of trade may act as another proxy for institutional reform, because it is an indicator of the degree of openness in a given country. Since the end of the cold war, CEE countries have become increasingly interested in integrating with the world economy, because they believe previous policies have fostered and protected inefficient production. Further, industrial countries and international organizations have pressured emerging economies

new institutional economics and fdi location 283

B. Data sources and methodology Variables used in the modeling are as follows: FDI = flow of foreign direct investment into the given host country, from the IMF International Financial Statistics CD ROM database (IMF 2001b). CORINGOV8 = corruption in government, an index scored from 1 to 6 constructed by Political Risk Services’ International Country Risk Guide (ICRG) database using subjective evaluation of country criteria. RULELAW9 = rule of law, an index scored from 1 to 6 constructed by Political Risk Services’ International Country Risk Guide database using subjective evaluation of country criteria. REPATREQ = repatriation restrictions imposed by the government on overseas funds transfers, from the IMF publication Exchange Arrangements and Exchange Restrictions (various years). EBRDER10 = European Bank for Reconstruction and Development (EBRD) index of enterprise reform, scored from 1 to 4, from the annual EBRD publication Transition Report (various years). to reduce their import barriers. Therefore, CEE countries have undertaken microeconomic reforms that promote freer movement of international trade. Logically, MNEs need more trade and more open economies for resource-seeking operations, especially as they integrate their global production with vertical and horizontal value-chain linkages. For a country to be an inherent part of this integration, it must allow MNEs to easily import and export. This integration is particularly important when MNEs seek a base to serve regional CEE markets. Because imports and exports were extremely highly correlated with gross domestic product in the host countries, we did not test this hypothesis. 8. According to Political Risk Services, “This is an assessment of corruption within the political system. Such corruption is a threat to foreign investment for several reasons: it distorts the economic and financial environment, it reduces the efficiency of government and business by enabling people to assume positions of power through patronage rather than ability, and, last but not least, introduces an inherent instability into the political process.” 9. Rule of Law “reflects the degree to which the citizens of a country are willing to accept the established institutions to make and implement laws and adjudicate disputes.” Higher scores indicate: “sound political institutions, a strong court system, and provisions for an orderly succession of power.” Lower scores indicate: “a tradition of depending on physical force or illegal means to settle claims.” Upon changes in government new leaders “may be less likely to accept the obligations of the previous regime.” Definitions from Political Risk Services, Inc. 10. This 4-point scaled index was measured as follows, aiming at “enterprise restructuring”: 1: soft budget constraints, few other reforms to promote corporate governance 2: moderately tight credit and subsidy policy but weak enforcement of bankruptcy legislation and little action to strengthen competition and corporate governance 3: significant and sustained actions to harden budget constraints and promote corporate governance effectively 4: substantial improvement in corporate governance 4+: standards and performance typical of advanced industrial economies.

284 robert grosse and len j. trevino

TREATY11 = number of bilateral investment treaties, as listed in the United Nations Conference on Trade and Development Bilateral Investment Treaties, 1959–1999 database (UNCTAD 2000). PRISHARE = private sector share in GDP (in %), from the EBRD Transition Report (various years). INFL = inflation (change in consumer price index), from IMF International Financial Statistics. POLRISK12 = an index of political risk based on 12 weighted variables, scored on a scale from 0 to 100, as measured by Political Risk Services’ International Country Risk Guide database. EXRATE = nominal exchange rate versus U.S. dollar, from IMF International Financial Statistics. GDP = gross domestic product (in current U.S. dollars), from IMF International Financial Statistics. We explored the hypotheses using a standard multiple regression modeling procedure. Given the data set of annual FDI inflows into 13 CEE countries from 1990 to 1999, we used a panel data format for the models, grouping FDI by country and by year as classifying dimensions. Standard OLS models, leastsquares dummy variable models, and random-effects GLS models were computed for each specification to explain FDI inflows. Random-effects models proved to utilize the available information most efficiently and produced the most significant model results. The correlation matrix of the variables is presented in Table A1. Before proceeding to regression estimates, we address some potentially problematic aspects of the time series nature of our data set. First, we were concerned about the potential for little or very slow variation in the independent variables, especially those related to regimes and regime changes. Closer review of the data,

11. Initially BITs were concluded between a developed and a developing country, usually at the initiative of the developed country. The developed country, typically a capitalexporting country, entered into a BIT with a developing country, typically a capital-importing country. This was done in order for the developed country to secure additional and higher standards of legal protection and guarantees for the investments of its firms beyond those offered under national laws. The developing country, on the other hand, would sign a BIT as one of the elements of a favorable climate to attract foreign direct investors. 12. “The aim of the political risk rating is to provide a means of assessing the political stability of the countries covered by the International Country Risk Guide on a comparable basis. This is done by assigning risk points to a preset group of factors, termed political risk components. The minimum number of points that can be assigned to each component is zero, while the maximum number of points depends on the fixed weight that component is given in the overall political risk assessment. In every case, the lower the risk point total the higher the risk, and the higher the risk point total the lower the risk.”

new institutional economics and fdi location 285

table 1. levin-lin unit root tests Variable

Level

t-test

FDI FDIUS INFL EXRATE GDP

–0.817∗∗∗ –0.949∗∗∗ –0.942∗∗∗ –0.609∗∗∗ –0.436∗∗∗

(–7.682) (–9.539) (–14.252) (–6.697) (–7.251)

∗ = significant at the .10 level ∗∗ = significant at the .05 level ∗∗∗ = significant at the .01 level

however, indicated that this concern was unwarranted since all of the independent variables exhibited sufficient year-to-year variation to justify their inclusion in the model without adjustment. Second, we were concerned about the possibility that we were regressing time-varying independent variables on a country-specific and time-invariant dependent variable. Thus, we tested the stationarity of our dependent variables by conducting unit root tests for panel data. The most popular form of unit root tests for panel data is Levin and Lin (Levin and Lin 1993). Results of this LevinLin test indicate that both of the dependent variables are stationary in level form (see Table 1). Although many scientific time series are stationary, most time series, especially macroeconomic time series, are trending (Kennedy 1998; Nelson and Plosser 1982). To avoid problems with using non-stationary time series, a Levin-Lin test was conducted on each of the macroeconomic variables, and the results indicate that all of them are stationary in level form (see Table 2). These two tests fail to reject the null hypothesis of stationarity of our data series in every case. Therefore, our model is estimated below in level form. C. Statistical results Table 3 shows the main model, in which the regression models explained over half of the variation in aggregate FDI flows into the thirteen countries during the 1990s. Our main interest in the statistical analysis is to demonstrate the importance of institutional (non-traditional) factors in explaining FDI into the formerly communist countries in the early years of their transition to market-based democracies. Two institutional variables proved significant in explaining this investment: the number of bilateral investment treaties entered into by the country, and the level of restrictions on repatriation of earnings of the CEE affiliates. The former may be interpreted as a measure of the country’s willingness to offer

286 robert grosse and len j. trevino table 2. regression results on foreign direct investment in central and eastern europe (dependent variable: global fdi flows into cee host countries, from unctad) INDEPENDENT VARIABLE = FDI INFLOW CONSTANT CORINGOV

HYPOTHESIS

– 1

RULELAW

1

REPATREQ

2

EBRDER

2

MODEL #1

MODEL #2

MODEL #3

–8.57

–7.86

–7.37

(–4.14)∗∗∗

(–4.66)∗∗∗

(–2.98) ∗∗∗

–

0.62

0.65

(2.62)∗∗∗

(2.74)∗∗∗

–

–

0.75 (2.53) ∗∗

1.02

.90

1.19

(3.25)∗∗∗

(2.94)∗∗∗

(3.26) ∗∗∗

–

0.62

–

(1.92)∗

TREATY PRISHARE

3 4

0.07

0.06

0.08

(4.07)∗∗∗

(3.92)∗∗∗

(3.63) ∗∗∗

–

–

–0.02 (–1.30)

POLRISK

5

0.04 (2.32)∗∗

–

–

INFL

6

0.01 (0.31)

–

–0.01 (–1.41)

EXRATE

7

GDP Observations Adjusted R2

8

–0.01

–0.01

–0.01

(–2.14)∗∗

(–1.92)∗

(–1.31)

0.01

0.01

0.01

(5.56)∗∗∗

(5.27)∗∗∗

(5.10) ∗∗∗

54 54 0.59 0.56

54 0.53

t-statistics in parentheses ∗ = significant at the .10 level ∗∗ = significant at the .05 level ∗∗∗ = significant at the .01 level These are all random effects, GLS regression models. variable definitions: CORINGOV = corruption in government [higher value = less corruption] RULELAW = extent of rule of law in policy application [higher value = less arbitrariness] REPATREQ = repatriation controls [higher value = more controls on repatriation] EBRDER = EBRD index of enterprise reform [higher value = greater reform] TREATY = # of bilateral investment treaties PRISHARE = private sector share in GDP INFL = inflation (change in consumer price index) POLRISK = political risk index [higher value = less risky] EXRATE = nominal exchange rate versus U.S. dollar GDP = gross domestic product (in current U.S. dollars)

new institutional economics and fdi location 287

equal treatment to foreign firms as compared to domestic companies. The more the host country has undertaken institution-building via BITs, the greater the amount of FDI that occurred. The latter outcome, repatriation restrictions being positively correlated with FDI, is surprising. This may be a phase in the transition from government-owned economic activity to more fully open economies, such that initial repatriation restrictions may be followed by a move toward greater openness to remittances in the future. Another measure of institutional constraints in these governments is their level of corruption. As expected, higher levels of perceived corruption led to lower levels of FDI. This finding was consistent for other specifications of corruption, such as the extent to which a country was governed by the rule of law rather than by arbitrary policymaking. Host country political risk was significantly and inversely related to FDI, as expected. Looking next at the new international economics from a macroeconomic perspective, we found that inflation did not add significantly to the explanation of FDI into the CEE countries, but that the other two measures did. This outcome was not totally unexpected, given the weak relationship of inflation to FDI in emerging markets in previous studies. The exchange rate had a significant negative correlation to the flow of FDI in most of the models, supporting the argument that a weaker-currency country attracts less investment. In addition, market size was highly significant in explaining FDI flows into CEE. Apparently, the main attraction of CEE countries for FDI during the 1990s was to serve local markets, as corroborated by the various company surveys cited previously. D. Discussion The new institutional economics is employed as a theoretical foundation to better understand the intersection of MNE investment strategy and the institutional environment in the transitional economies of CEE. We test the theory empirically with traditional macroeconomic factors as well as non-traditional environmental and political measures of institutional development, and we attempt to explain the institution-environment framework surrounding FDI decision-making. By applying the new institutional economics to transitional economies, we have responded to the call to explain organizational behavior in developing country settings utilizing this theory (Meyer 2001; Shenkar and von Glinow 1994; Hoskisson et al. 2000). Our findings confirm that corruption is an element of the institutional environment that increases uncertainty and costs associated with long-term capital investment in CEE. Since corruption forces investors to incur higher costs to reduce this risk, we expected and found that the level of corruption in government constrains inward FDI in CEE. If transitional economies want to attract greater FDI, they will have greater success by providing a business environment that operates under fair, concrete and transparent laws. The degree of restriction on foreign firms’ profit repatriation was positively related to inward FDI. This finding was surprising because capital account

288 robert grosse and len j. trevino

restrictions limit companies’ ability to allocate profits optimally on a global basis (for use where returns may be higher or to augment dividends to shareholders). Since foreign operations in the formerly communist economies of CEE are still relatively young, it is possible that retained earnings are needed for expansion in these new operations. Thus, managers may view current repatriation restrictions as unimportant at present, and perhaps likely to change in the future. The policies themselves may be in a transitional phase, with lower repatriation limits on the horizon, and with currently high limits as a false indicator of general policy restrictiveness. We found that the number of bilateral investment treaties that CEE countries had signed was highly significant in attracting FDI to the region. Our study indicates that foreign investors view BITs that assure equal treatment of foreign and domestic investors as a critical component of institution building. Such equal treatment, we posit, reduces the costs of doing business in CEE. Support for this variable demonstrates that it serves as another useful measure of institutional reform that affects FDI decision-making. This finding should be of considerable interest not only to such non-governmental organizations as UNCTAD and the World Bank, which sold developing countries on the value of BITs, but also to the host countries that eventually adopted them. As expected, we found a highly significant and positive relationship between the degree of privatization in the host country and inward FDI. This result is consistent with recent studies (Frydman et al. 1998; Meyer 1998) that view privatization as a fundamental element of institutional development. The decision by CEE countries to privatize state-owned enterprises provides two roles in terms of present and future costs for foreign investors in the institutional environment. First, it allows companies to improve operating efficiency and to reduce FDIrelated costs when they acquire a state-owned enterprise, because the foreign investor provides managerial and technological capability that can bring the acquired firm up to global standards. Second, privatization sends a message to foreign investors that the government is building an institutional environment that is consistent with market-oriented economies. Although we have shown that privatization correlates positively with FDI, the number of government-owned companies is finite. Thus, governments are limited in their ability to continue to attract FDI from companies investing in previously government-owned firms. We expected and found an inverse and significant relationship between political risk and FDI. This finding supports our institutional constraint argument, similar to the corruption variable. (In fact, corruption is a form of political risk that we were able to separate out from the overall category.) High political risk increases costs for foreign investors by forcing them to operate under extreme uncertainty. The CEE countries present unique risks, and companies are aware of these and are inclined to invest in those countries where institutional stability is greatest and FDI-related costs (i.e., political risk) are lowest.

new institutional economics and fdi location 289

We hypothesized that lower inflation would correlate with higher inward FDI, yet although we found directional support for the inflation variable, it was not significant in explaining FDI. It is possible that companies have gained so much operating experience in high inflation environments around the world that they no longer perceive high inflation as sufficiently problematic to justify keeping them out of potentially attractive and profitable markets. Just as likely, the relatively low levels of inflation worldwide in the past decade may not constitute a significant barrier to FDI at this time. By indicating that countries with stronger currencies attracted greater levels of FDI, results for the exchange rate variable supported the macroeconomic institution-building argument and contradicted previous findings related to market disequilibrium in developed countries (Grosse and Trevino 1996; Klein and Rosengen 1994; Froot and Stein 1991). This outcome is consistent with the part of the literature that finds that foreign investors prefer to invest in high-value (high real exchange rate) markets where earnings may be expected to depreciate less frequently and rapidly. That is, firms prefer to obtain assets in countries that demonstrate a greater likelihood of protecting those assets. Therefore, more FDI went into countries where the costs of dealing with depreciating currencies were lower. Consistent with existing literature in developed and developing country settings, we found that GDP was highly significant in explaining FDI flows into the CEE countries. Our findings indicate that companies are still locating in the largest country markets. This is true in spite of the growth in regional trade agreements and various bilateral agreements that allow companies to more readily gain access to other country markets without a necessary local production presence. The interesting adjustment will occur when the CEE countries are linked sufficiently to the European Union so that firms will have confidence in investing in CEE for serving all of Europe; this may produce a large flow of “offshore assembly” investment as has occurred in Mexico for sales in the U.S. market during the past twenty years.

conclusion The flow of FDI into the transitional economies of CEE tends to follow patterns similar to those in other emerging markets, with the added dimension that institutional factors appear to play a larger role than elsewhere.13 In particular, bilateral investment treaties, the degree of enterprise reform, and repatriation 13. In a statistical context, models that include the political and institutional variables explain more of FDI into the CEE countries than models that only include “traditional” factors that explain FDI in all countries. Re-running the models presented in this chapter

290 robert grosse and len j. trevino

rules tend to stimulate FDI, while political risk and the level of corruption in government tend to constrain FDI into these countries. This outcome provides strong support for this reasoning based on the new institutional economics. It also implies that country-level efforts to attract FDI in the twenty-first century will require governments with strong institutions that are committed to reducing uncertainty and FDI-related costs for foreign investors, by pursuing lower corruption, more stable political environments, and bilateral investment treaties that help to create an even playing field for foreign and domestic investors. Greater achievement of the rule of law, of enterprise reform, and of intergovernmental agreements on the treatment of FDI all form a background that is conducive to attracting FDI. CEE governments that want to use FDI as part of their engine of economic growth are clearly instructed about the direction of institutional change that is needed. As in other FDI studies, it has been shown that more comprehensive frameworks for explaining this phenomenon produce more complete understanding of FDI flows. Limiting analysis only to economic factors, or to factors that are readily measurable, may simplify the effort, but the value of adding institutional variables relating to the FDI environment has been clearly demonstrated here. With improvements in the measurement of these variables, our ability to explain and understand FDI has improved significantly. Even though our analysis is fairly comprehensive from a regulatory perspective, our narrow focus on the new institutional economics is limited in that neither the normative nor the cultural-cognitive dimensions of institutional theory were analyzed in the present study. An additional institutional perspective comparing Anglo-Saxon and German/Scandinavian models of market economies (more “liberal” versus more “coordinated”) could provide further understanding of the process through which the CEE countries are passing.14 Future research could improve upon the present study by measuring and modeling these added dimensions of institutional theory into a more comprehensive explanatory model of the multinational enterprise and investment location.

with only traditional factors produced about half of the explanatory power of the augmented models. 14. As pointed out by one reviewer, “It might be the case that more liberal governmental policies in CEE are attracting more FDI . . . but such an approach can also go hand in hand with a weakening of labor law and industrial relations and other social welfare state institutions, by favoring mainly a small political and business elite in these countries, as it is the case, for example, in Russia and China. Thus, it is not astonishing that in these countries corruption and political risks are increasing.”

appendix table a1. correlation matrix FDIU

GDP

INFL

IMPT

EXPT

XRAT

PRISH

REPA

EBRD

PRSK

RULE

1.0 n.r.

– 1.0

– –

– –

– –

– –

– –

– –

– –

– –

– –

– –

– –

– –

GDP

0.28

n.a.

1.0

–

–

–

–

–

–

–

–

–

–

–

INFL

FDIFLO FDIUS

CORI TREA

–0.05

n.a.

0.04

1.0

–

–

–

–

–

–

–

–

–

–

IMPORT

0.35

n.a.

0.91

–0.02

1.0

–

–

–

–

–

–

–

–

–

EXPORT

0.27

n.a.

0.92

0.05

0.95

1.0

–

–

–

–

–

–

–

–

EXRATE

0.01

n.a.

–0.10

–0.01

–0.05

–0.06

1.0

–

–

–

–

–

–

–

PRISHA

0.35

n.a.

0.22

–0.29

0.23

0.13

–0.01

1.0

–

–

–

–

–

–

REPATR

0.26

n.a.

0.31

0.14

0.40

0.35

0.11

–0.16

1.0

–

–

–

–

–

EBRDER

0.34

n.a.

–0.02

–0.44

0.03

–0.13

0.07

0.77

–.10

1.0

–

–

–

–

POLRSK

0.13

n.a.

–0.09

–0.39

0.03

–0.11

0.26

0.47

0.02

0.77

1.0

–

–

–

RULELA

0.01

n.a.

–0.42

–0.40

–0.30

–0.45

0.13

0.27

0.01

0.63

0.67

1.0

–

–

CORING

–0.01

n.a.

–0.46

–0.14

–0.39

–0.52

0.16

–0.20

0.22

0.35

0.42

0.54

1.0

–

TREATY

0.16

n.a.

0.09

–0.17

0.10

0.05

0.02

0.58

–0.39

0.25

0.10

–0.02

–0.29

1.0

new institutional economics and fdi location 291

FDIF

292 robert grosse and len j. trevino

references Aspergis, Nicholas and Costas Katrakilidis (1998). “Does inflation uncertainty matter in foreign direct investment decisions? An empirical investigation for Portugal, Spain and Greece,” Rivista Internazionale di Scienze Economiche e Commerciali, 45, 4, pp. 729–744. Bailey, Warren and Y. Peter Chung (1995). “Exchange rate fluctuations, political risk, and stock returns: some evidence from an emerging market,” Journal of Financial and Quantitative Analysis, 30 (December), pp. 541–561. Butler, Kirt C. and Domingo Castelo Joaquin (1998). “A note on political risk and the required return on foreign direct investment,” Journal of International Business Studies, 29, 3, pp. 599–608. Clague, C. (1997). Institutions and Economic Development: Growth and Governance in Less-Developed and Post-Socialist Countries (Baltimore, MD: Johns Hopkins University Press). Coase, Ronald H. (1937). “The nature of the firm,” Economica, 4. Contractor, Farok J. (1990). “Ownership patterns of U.S. joint venture partners abroad and the liberalization of foreign government regulations in the 1980s: evidence from the benchmark surveys,” Journal of International Business Studies, 21, 1, pp. 55–73. Davis, Lance E. and Douglass C. North (1971). Institutional Change and American Economic Growth (Cambridge, MA: Cambridge University Press). Devlin, R. and R. Cominetti (1994). “La crisis de la empresa publica, las privatizaciones y la equidad social,” Reformas de Politica Publica, 26 (LC/L 832), pp. 1–35. Douglas, Mary (1982). “The effects of modernization on religious change,” Daedalus, Winter, pp. 1–19. Estrin, Saul, Kirsty Hughes and Sarah Todd (1997). Foreign Direct Investment in Central and Eastern Europe (London: Cassel). European Bank for Reconstruction and Development (2008). Transition Report. (London: EBRD). http://www.ebrd.com/pubs/econo/series/tr.htm Froot, K.A. and J.C. Stein (1991). “Exchange rates and foreign direct investment: an imperfect capital markets approach,” Quarterly Journal of Economics, 106, pp. 1191–1217. Frydman, R.M. Hessel and A. Rapaczynski (1998). “Ownership, restructuring and performance of enterprises in Central Europe,” Paper presented at the First International Workshop on Transition and Enterprise Restructuring in Eastern Europe, Copenhagen. Grosse, Robert and Len J. Trevino (1996). “Foreign direct investment in the United States: an analysis by country of origin,” Journal of International Business Studies, 27, 1, pp. 139–155. Hall, Peter and David Soskice, eds. (2001). Varieties of Capitalism (New York: Oxford University Press). Harris, J.J. Hunter and C.M. Lewis (1995). The New Institutional Economics and Third World Development (London: Routledge). Hartenek, G. (1995). “Privatization as a catalyst for FDI: the case of Argentina,” in Foreign Direct Investment: OECD Countries and Dynamic Economies of Asia and Latin America (Washington, D.C.: OECD Publications and Information Center), pp. 89–106.

new institutional economics and fdi location 293

Heinrich, Jeffrey and Denise Konan (2001). “Prospects for FDI in AFTA,” ASEAN Economic Bulletin, 18, 2, pp. 141–160. Hines, James R. (1995). “Technology transfer and the R&D activities of multinational firms,” NBER Research Project Report (Chicago and London: University of Chicago Press), pp. 225–248. Hoskisson, Robert, Lorraine Eden, Chung Ming Lau, and Mike Wright (2000). “Strategy in emerging economies,” Academy of Management Journal, 43, 3, pp. 249–267. Hymer, Stephen (1976). The International Operations of National Firms: A Study of Direct Foreign Investment (Cambridge, MA: MIT Press), previously unpublished Ph.D. Dissertation, 1960. International Monetary Fund (IMF) (2001a). Exchange Arrangements and Exchange Restrictions (Washington, D.C.: IMF). —— (2001b). International Financial Statistics Yearbook 2001 (Washington, D.C.: IMF). —— (2008) International Financial Statistics database. http://www.imfstatistics.org/imf/ Jain, Arvind K. (2001). “Corruption: a review,” Journal of Economic Surveys, 15, 1, pp. 71–121. Jermakowicz, W. and C. Bellas (1997). “Foreign direct investment in Central and Eastern Europe,” International Journal of Commerce and Eastern Europe, 7, 2, pp. 33–55. Kennedy, P. (1998). A Guide to Econometrics (Cambridge, MA: MIT Press). Kindleberger, C. (1969). American Business Abroad (New Haven: Yale University Press). Klein, M. and E. Rosengren (1994). “The real exchange rate and foreign direct investment in the United States: relative wealth versus relative wage effects,” Journal of International Economics, 36, 3/4, pp. 373–389. Kogut, B. (1996). “Direct investment, experimentation, and corporate governance in transition economies,” in R. Frydman, C.W. Gray and A. Rapazynski, eds., Corporate Governance in Central Europe and Russia (Budapest: Central European University Press), pp. 293–332. Madhok A. (1997). “Cost, value and foreign market entry mode: the transaction and the firm,” Strategic Management Journal, 18, 1, pp. 39–62. Marinov, Marin Alexandro and Svetla Trifonova Marinova (1999). “Foreign direct investment motives and marketing strategies in Central and Eastern Europe,” Journal of East-West Business. 5, 1/2, pp. 25–56. Meyer, Klaus (1998). Direct Investment in Economies in Transition (Aldershot: Elgar). —— (2001). “Institutions, transaction costs and entry mode choice in Eastern Europe,” Journal of International Business Studies, 32, 2, pp. 357–367. Nelson, C.R. and C.I. Plosser (1982). “Trends and random walks in macroeconomic time series,” Journal of Monetary Economics, 10, pp. 139–162. North, Douglass (1990). Institutions, Institutional Change and Economic Performance, (Cambridge, MA: Cambridge University Press). Organization for Economic Co-operation and Development (OECD) (1994). Assessing Investment Opportunities in Economies in Transition (Paris: OECD). Political Risk Services (2008). International Country Risk Guide. (Syracuse, New York: PRS) http://www.prsgroup.com/ICRG.aspx Rolfe, Robert, and Timothy Doupnik (1996). “Going east: western companies invest in East/Central Europe,” Multinational Business Review, 4, 2, pp. 1–12. Rumelt R.P., D.E. Schendel and D.J. Teece, eds. (1991). “Strategic management and economics,” Strategic Management Journal, Winter Special Issue, 12, pp. 5–29.

294 robert grosse and len j. trevino Scott, W.R. (2001). Institutions and Organizations (London: Sage Publications). Shenkar, O. and M.A. von Glinow (1994). “Paradoxes of organizational theory and research: using the case of China to illustrate national contingency,” Management Science, 40, pp. 56–71. Svetlicic, M. and M. Rojec (1994). “Foreign direct investment and the transformation of Central European economies,” Management International Review, 34, pp. 293–312. Trevino, Len J., John Daniels and Harvey Arbelaez (2002). “Market reform and FDI in Latin America: an empirical investigation,” Transnational Corporations, 11, 1, pp. 29–48. United Nations Conference on Trade and Development (UNCTAD) (2000). Bilateral Investment Treaties, 1959–1999 (Geneva: United Nations). http://www.unctad.org/ Templates/Page.asp?intItemID=1465 —— (1998 – WIR98)). World Investment Report (Geneva: United Nations). http://www. unctad.org/Templates/WebFlyer.asp?intItemID=2426&lang=1 —— (2001 – WIR01). World Investment Report (Geneva: United Nations). http://www. unctad.org/Templates/WebFlyer.asp?intItemID=2434&lang=1 Walker, G. and D. Weber (1984). “A transaction cost approach to make-or-buy decisions,” Administrative Science Quarterly, 29, pp. 373–391. Wei, Shang-Jin (2000). “How taxing is corruption on international investors?,” Review of Economics and Statistics. 82, 1, pp. 1–11. Williamson, Oliver E. (1975). Markets and Hierarchies: Analysis and Anti-trust Implications (New York: Free Press). —— (1985). The Economic Institutions of Capitalism. (New York: Free Press).

10. do investment agreements attract investment? evidence from latin america∗ kevin p. gallagher and melissa b.l. birch introduction The developing world needs more investment. Indeed, attracting investment— domestic or foreign—is increasingly a core goal of developing countries. Developing countries, especially in Latin America, have been looking to foreign direct investment (FDI) as a more stable way to balance their financial accounts. From a microeconomic perspective, the promise of FDI is in positive spillovers, whereby foreign presence creates forward and backward linkages to the domestic economy through supplier networks, and human and technological capital transfer through demonstration effects. One strategy used by developing countries to attract FDI has been to negotiate bilateral investment treaties (BITs). During the 1990s more FDI accrued to developing countries than in any other period in recorded history, and Latin America was no exception. What factors made this investment flow to Latin America and the Caribbean during the 1990s? Did BITs make a difference? What lessons can be drawn out for future trade agreements? The 1990s were a period of unprecedented increases in the level of FDI in the world economy, reaching $1.6 trillion in the year 2000. However, the lion’s share of that investment—70% of all FDI—stayed in developed countries. Of the FDI that did accrue to the developing world during the 1990s, almost 80% of it flowed to just ten countries. Indeed, China, Brazil, and Mexico received 58% of all FDI that flowed to the developing world in the 1990s (UNCTAD 2002). Table 1 shows that just ten countries received 76% of all the developing country FDI between 1990 and 2001. Five of those countries (Brazil, Mexico, Argentina, Bermuda, and Chile) are in Latin America and the Caribbean (LAC). Between 1990 and 2001 Brazil topped the list of total FDI recipients with $128.7 billion dollars in FDI. Mexico was a close second at $113.6 billion. The majority of the investment into Brazil was from Europe. In the Mexican case the majority of FDI comes from the United States.

∗ This chapter was reprinted with permission from the Journal of World Investment and Trade. The chapter was originally published as “Do investment agreements attract investment?—evidence from Latin America,” Journal of World Investment and Trade 7(6), 961 (2006). The authors would like to thank Charles Hernick for research assistance on this chapter.

296 kevin p. gallagher and melissa b.l. birch table 1. ten largest developing country recipients of fdi inflows (average annual inflows, 1990 to 2001) Country China Brazil Mexico Singapore Argentina Bermuda Malaysia Poland Chile South Korea Top 10 Total Total for Developing Countries Top 10 Share:

Inflows (U.S.$ billions) 45.6 12.9 11.4 7.2 6.8 4.5 4.4 4.1 3.5 3.2 103.5 136.7 76%

Source: UNCTAD 2002

As shown in Table 2, just five countries received 74% of the $527.5 billion of FDI in LAC between 1990 and 2001. In proportion to GDP, FDI in Latin America quintupled in the 1990s. Whereas FDI was at a steady 1% of GDP throughout most of the 1980s in LAC, by 2000 FDI had reached close to 6% of GDP. Until recently, BITs were not very uniform across countries, but in the past decade they have begun to include a few core elements: performance requirements, expropriation, and dispute settlement. Most BITs restrict the ability of developing countries to require that foreign firms adhere to performance requirements such as local content standards, technology transfer requirements, and so forth. Another key element is the notion of expropriation. BITs set clear rules regarding what constitutes an expropriation and what compensation is due if expropriation does indeed occur. BITs over the last twenty years vastly expand investor rights. In contrast with World Trade Organization (WTO) disputes—in which firms have to go to the government, which would then take a claim to another government—BITs now allow private investors themselves to file for damages against governments. These measures have become very controversial, with some claiming that new social and environmental regulations are being interpreted by clever legal teams for investors as new forms of expropriation (Moyers 2002). Others have argued that such measures constitute a moral hazard whereby firms have the incentive to file such claims and weaken social protection laws (Graham 2002). In the 1990s, nations in LAC signed over 300 BITs, over two-thirds of which were with developed countries (UNCTAD 2000a). Most recently, the elements

do investment agreements attract investment? 297

table 2. five largest fdi recipients in the americas (average annual inflows, 1990–2001, u.s.$ millions) Country Brazil Mexico Argentina Bermuda Chile Top 5 Total Americas Total Top 5 Share of Americas

Inflows 128.7 113.6 68.5 44.6 34.9 390.4 527.5 74%

Source: UNCTAD 2002

of BITs are now being placed inside what are called “free trade agreements,” which are really agreements that include a wide range of measures on such issues as agricultural subsidies, intellectual property rights, competition policies, and services. The United States and countries in the Western Hemisphere are now involved in (or have recently completed) the negotiation of numerous agreements, including the U.S.–Chile Free Trade Agreement, the Free Trade Area of the Americas (FTAA), the Central American Free Trade Agreement (CAFTA), and the Andean Free Trade Agreement. All have investment text with the core elements described above. This chapter attempts to identify the determinants of FDI into these and other countries in LAC during the period from 1980 to 2003. At this writing, the countries of the Americas are considering a hemisphere-wide trade agreement, the FTAA. In addition, the U.S. government is encouraging individual and sub-regional groups of nations to sign bilateral trade and investment treaties. Given that tariffs on goods to be sold in the United States are relatively low, the promise of increased foreign investment is the most sought-after prize that many nations in LAC hope will result from these new treaties. The chapter is organized into five parts. A literature review follows this introduction. Part B is a description of the methodology constructed and the data used for this analysis. Part C comprises a discussion of our results. The conclusion summarizes our work and draws out the chapter’s implications for ongoing trade policy negotiations. A. Literature Review 1. General studies There is a large and growing literature on the determinants of FDI in developing countries (for a comprehensive review see Lim 2001). A thorough review of that literature is beyond the scope of this chapter, but we will briefly discuss the literature’s major findings and review the smaller number of

298 kevin p. gallagher and melissa b.l. birch

studies specific to Latin America and the Caribbean. In the larger literature, there is a consensus that at least six general factors are key determinants of FDI location in the developing world: market size and quality, factor endowments and prices, macroeconomic stability, political stability, policy environment, and geographical and infrastructural determinants. We will address each of these in turn. Market size, (usually measured as GDP or, occasionally, as population) has long been identified in the general literature as a key determinant of FDI. In addition, potential market size (as measured by GDP growth rates) and market quality (GDP per capita) have also been found to affect the investment decisions of foreign firms (Singh and Jun 1995; Balasubramanyam 2001; Lim 2001). The importance of the market-related factors—particularly size—has been supported by several of the Latin America-specific studies (Tuman and Emmert 1999; Baumgarten and Hausman 2000; Trevino, Daniels et al. 2002; Vallejo and Aguilar 2002; ECLAC 2003). A recent study looked at the determinants of U.S. FDI in Latin America. That study also found market size to be the most significant determinant of FDI, in addition to exports and education (Tuman and Emmert 2004). Two kinds of factor endowments have been shown to be key drivers of FDI. First, the presence of natural resources may spur investment in extractive industries (Balasubramanyam 2001). Although Baumgarten and Hausman (2000) suggest that natural resources are not an important determinant in Latin America, the Economic Commission on Latin America and the Caribbean (ECLAC; 2003) suggests that such resource-seeking investment may be important in South America. Second, low wages may also draw FDI seeking low-skilled labor. Despite the profile of the wage issue in international debates, it has received relatively little attention in the Latin American context, although according to ECLAC it may be particularly important in Central America and the Caribbean. When it comes to labor resources, costs may not be the only concern—the quality of the labor force may also be important. Findings on this question have been mixed, with ECLAC (2003) and Nunnenkamp (1997) suggesting that human resource quality may be an FDI determinant, and Baumgarten and Hausman (2000) finding otherwise. Macroeconomic stability is commonly believed to be necessary for high levels of FDI, including in the Latin American context (Nunnenkamp 1997). Often, however, different studies find differing levels of substantiation for the importance of various macroeconomic variables. Debt, exchange rates, and inflation have all been proposed as determinants of FDI in Latin America, and have been supported by some studies and not by others (ECLAC 2003; Baumgarten and Hausman 2000; Tuman and Emmert 1999; Trevino, Daniels et al. 2002). One key study of the determinants of investment in LAC was conducted as part of large assessment by ECLAC. This study examined the determinants of

do investment agreements attract investment? 299

gross fixed investment between 1979 and 1995 in eight countries: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Mexico, and Peru. The authors examined gross fixed investment (as opposed to FDI) as a percentage of GDP as the dependent variable, and had results largely consistent with the literature discussed above. They found investment to be in large part a function of macroeconomic stability (such as variation in the exchange rate) and the extent to which each country implemented its reforms (Stallings and Peres 2000). Some studies have identified political stability as an important determinant of FDI, both in the general literature (Singh and Jun 1995; Balasubramanyam 2001) and in the Latin America-specific literature (Tuman and Emmert 1999; Baumgarten and Hausman 2000; ECLAC 2003). Other studies have not found political stability to be a significant determinant (Trevino, Daniels et al. 2002). The policy environment includes a number of elements that may affect FDI flows. Specifically, trade and investment policies, privatization, and the general business operating environment may be expected to impact FDI (Trevino, Daniels et al. 2002; ECLAC 2003). Restrictive investment and business policies include high tax rates, restrictions on capital repatriation, limitations on foreign participation in particular sectors, and a host of other potential measures. These can be expected to discourage FDI (Vallejo and Aguilar 2002; ECLAC 2003). Trade barriers might be expected to have a somewhat more mixed impact on FDI. Trade barriers have long been considered to encourage “tariff-jumping” FDI: investment by foreign companies seeking to bypass trade barriers in order to gain access to large, desirable markets. In contrast, as vertical integration has become an increasingly important strategy among multinational firms, we may expect restrictive trade policies to inhibit factor-seeking (particularly laborseeking) FDI, while less restrictive policies and a stronger export orientation may encourage such FDI (Singh and Jun 1995; Nunnenkamp 1997; Baumgarten and Hausman 2000; Balasubramanyam 2001; Lim 2001). Studies that do not find tariffs to be a significant determinant include Baumgarten and Hausman (2000) and Blonigen and Feenstra (1996), while Tuman and Emmert (1999) do not find an export orientation to be significant. Scholars have identified a number of factors relating to geography and infrastructure. In particular, distance from major markets and transportation and communication infrastructure seem likely to affect FDI flows. As in the case of trade barriers, however, they may affect factor-seeking and market-seeking FDI differently, with long distances and high transport costs potentially discouraging the former while encouraging the latter—although as with the other potential determinants, results have been mixed (Balasubramanyam 2001; Lim 2001; Vallejo and Aguilar 2002; ECLAC 2003). Finally, although this has not been included in most studies of FDI determinants, in many countries in the LAC region, including Brazil and Mexico, the rapid rise of FDI during the 1990s was associated with a rapid rise of mergers and acquisitions, rather than with new investment (see Table 3). Although acquisitions

1987 1988

1989

1990

27 6,274

1991

1992

1993

1994

1995

1996

1997

302 1,164

1,803

1,315 1,869

3,611

4,635

1998

1999

2000

2001

10,396 19,407

5,273

5,431

Argentina

0

60

Barbados

0

0

0

0

189

0

0

4

6

64

0

0

0

0

1

Belize

0

0

0

0

0

0

0

0

0

0

0

62

0

3

62

Bolivia

0

0

15

26

0

0

0

0

821

273

911

180

232

19

0

Brazil

196

287

2,217

158

174

624

367

1,761

1,761

6,536

12,064

29,376

9,357

23,013

7,003

Chile

0

38

260

434

338

517

276

891

717

2,044

2,427

1,595

8,361

2,929

2,830

Colombia

0

764

0

341

49

31

8

1,248

67

2,399

2,516

1,780

302

1,589

170

Costa Rica

0

0

64

3

0

0

1

17

96

27

28

2

71

21

0

Dominican Republic

0

0

0

0

0

0

0

0

0

46

0

28

673

464

0

Ecuador

0

0

0

0

0

49

0

44

35

105

27

79

214

153

6

El Salvador

0

0

0

0

0

0

0

0

40

0

41

978

0

0

168

Grenada

0

0

0

0

0

0

0

0

0

0

5

0

0

0

0

300 kevin p. gallagher and melissa b.l. birch

table 3. cross-border mergers and acquisitions by country of sale, 1987–2001 (u.s. $ millions)

Guatemala

0

0

0

3

3

0

29

0

0

26

30

582

101

13

121

Guyana

0

0

0

17

7

0

0

0

0

0

1

0

23

0

0

Haiti

0

0

0

0

0

0

0

0

0

0

0

2

0

0

0

Honduras

0

0

0

0

5

0

0

1

0

0

0

367

0

314

537

Mexico

1

54

395 2,326 10,961 1,864

1,913

719

719

1,428

7,927

3,001

859

Nicaragua

0

0

0

0

0

0

0

0

0

23

42

0

11

115

83

Panama

0

15

0

0

0

0

6

73

9

14

652

216

151

130

8

Paraguay

0

0

0

0

0

0

0

0

0

27

2

11

0

65

67

Peru

0

0

0

0

15

174

62

3,082

945

844

911

162

861

107

555

Suriname

–

–

–

–

–

–

–

–

0

0

0

0

0

0

3

Trinidad and Tobago

0

0

0

0

1

7

22

177

0

0

205

0

0

0

0

Uruguay

0

0

18

0

0

0

5

40

19

0

0

36

0

27

36

Venezuela

0

0

0

11

2,032

0

62

337

278

1,072

1,946

3,220

276

2,409

73

do investment agreements attract investment? 301

Sources: UNCTAD 2000b (1987–1994) and UNCTAD 2004 (1995–2001)

3,965 17,017

302 kevin p. gallagher and melissa b.l. birch

of private firms were widespread, many of these mergers and acquisitions were associated with privatization of state assets associated with the broad reform programs carried out in many Latin American countries during the 1980s and 1990s. In 1997, for example, countries receiving significant amounts of FDI in the process of privatization included Brazil, Colombia, and Venezuela, while in the same year, acquisition of private firms was more important in Mexico, Argentina, and Chile—countries that had carried out their major privatizations in previous years. Overall, during the period from 1988 to 1999, these six countries, plus Peru, carried out the most privatizations of state assets valued at U.S.$5 million or more (ECLAC 1998; MIGA 2005). 2. Studies on BITs To date, three studies have been conducted that evaluate the extent to which investment treaties affect investment flows. Despite the existence of several useful papers on the issue of investment determinants in Latin America and on the relationship between investment agreements and FDI, no study looks exclusively at the role of BITs in determining FDI in Latin America and the Caribbean. The first study to examine the role of BITs in attracting foreign investment was conducted in 1998 by the United Nations Conference on Trade and Development (UNCTAD). The study finds very limited evidence of a strong relationship between BITs and investment flows. First, UNCTAD conducts a regression analysis on the impact of BITs on three different sets of dependent variables—total FDI, total FDI stocks, and total FDI per unit of GDP—fusing time-series data from 14 home and 72 host countries. Looking at the years 1993 to 1995, the study finds BITs to be significantly correlated with investment flows in only one of the authors’ numerous regression runs (for total FDI flows). UNCTAD finds no relationship between BITs and FDI per unit of GDP or total FDI stocks. The study also performs a regression analysis using crosssectional data from 133 countries. The study finds that BITs do not appear to have a major impact on FDI. UNCTAD concludes that the impact of BITs is small and secondary to the effects of other determinants, especially market size (UNCTAD 1998). The World Bank (2003) examines whether BITs determined the levels of FDI from 20 OECD countries flowing to 31 developing countries from 1980 to 2000. In different fixed-effects models, the author examines dependent variables of total FDI flows to host countries, total FDI flows to host countries as a percentage of total FDI flows from OECD countries, and total FDI per unit of GDP. In only one case—the dependent variable FDI flows as a percentage of total FDI flows—does the author find a statistically significant relationship. The World Bank study concludes that BITs do not stimulate additional investment. It finds that market size and macroeconomic stability are the key drivers of foreign investment. The study stresses that BITs are no substitute for domestic institutions and that to the extent that they work to increase investment at all

do investment agreements attract investment? 303

in developing countries, they act as a complement to the existence of strong institutions (Hallward-Dreimeier 2003). A more recent study conducted by Tobin and Rose-Ackerman (2004) at Yale University looks at the relationship between BITs and investment flows in the world economy as a whole. In their fixed-effects regressions, the authors’ dependent variable is FDI as a percentage of world FDI. Examining FDI for 176 countries from 1975 to 2000, they find that the total number of BITs concluded by a country has no independent significant impact on investment flows into the country— except when such countries are seen to be politically risky. The authors conclude that a BIT agreement can be a signal that a country wants to get back on track. The authors also look at flows of FDI from the United States and BITs with the United States. Here they find a negative relationship: those countries with BITs with the United States received less FDI. The authors say that “overall, these results indicate that signing a BIT with the [United States] does not correspond to increased FDI flows” (Tobin and Rose-Ackerman 2004, 24). Another study has findings that conflict those of Tobin and Rose-Ackerman. An econometric analysis of the determinants of FDI in the global economy was conducted that measures the level of FDI from 1991 to 2000. The authors find that a U.S. BIT is significantly correlated with an increase in U.S. FDI (Salacuse and Sullivan 2005). Their findings may be different from those of Tobin and Rose-Ackerman because of their limited sample size: The study looks at flows to 31 countries. For Latin America, for instance, the sample only includes Argentina, Mexico, and Panama. Our study builds a model of determinants of FDI in the Americas by combining the key variables in the broader literature on FDI determinants in the hemisphere with these studies on BITs. We work to this end by examining a range of potential FDI determinants, testing not only many of the variables used in previous studies, but also the importance of BITs. Finally, our study will have a broader geographical and temporal coverage than previous studies (24 countries during the period from 1980 to 2003), and will examine FDI originating both from the United States and from the world as a whole. B. Data and Methodology Based on the literature discussed in the previous section, we develop an econometric model to identify the determinants of FDI in the hemisphere. We hypothesize that foreign investment in the Americas is a function of some combination of market size, growth, factor endowments, macroeconomic and political stability, and privatization policies. A general model for this function can be expressed as follows: INV = γ + λ + β1 BIT + β2 INFL + β3 EXCH + β4 GDP + β5 EXP + β6 LIT + β7 INCOME + β8 Growth + β9 PRIV + ε

where INV signifies the various dependent variable indicators for foreign investment, BIT is the variable for investment treaties, INFL is inflation,

304 kevin p. gallagher and melissa b.l. birch

GDP represents total market size, EXP are trade variables (total exports or exports-GDP ratios in various regressions), LIT is the literacy rate, INCOME is GDP per capita, GROWTH is the GDP growth rate, and PRIV is the number of privatization contracts. The various measures tested (and the sources) for these regressions are exhibited in Table 4. table 4. variables used in the regressions Dependent Variables Variable

Description

Units

Source

LNU.S.FDI

ln of U.S. FDI

Current U.S.$

LNTFDI

ln of net FDI from all source countries

Current U.S.$

Bureau of Economic Analysis World Development Indicators

Independent Variables Variable LNGDP

Description ln of GDP

Units Current U.S.$

LNEXCH

ln of U.S.-local currency exchange rate ln of annual inflation rate

Ratio

ln of total exports/GDP

%

Total Bilateral Investment Treaties signed U.S. BIT signed

# of treaties

LNINFL LNEXP_ GDP TSBIT U.S.BIT

LNLIT LNGDPGR LNGDPPC PRIV_5

ln of literacy rate among population aged 15+ ln of average annual growth rate ln of GDP per capita Privatizations worth at least U.S.$5 million

%

0 if no BIT signed with U.S. 1 if BIT signed with U.S. % % Constant 2000 U.S.$ # of privatizations

Source World Development Indicators World Development Indicators World Development Indicators World Development Indicators UNCTAD UNCTAD

World Development Indicators World Development Indicators World Development Indicators MIGA

do investment agreements attract investment? 305

table 5. countries included in regressions • • • • • • • • • • • •

Argentina Barbados Belize Bolivia Brazil Chile Colombia Costa Rica Dominican Republic Ecuador El Salvador Grenada

• • • • • • • • • • • •

Guatemala Guyana Haiti Honduras Mexico Nicaragua Panama Paraguay Peru Trinidad and Tobago Uruguay Venezuela

We look at two dependent variables: the total amount of FDI flows to a specific country, and the total amount of net inflows of U.S. FDI into a specific country. We perform fixed-effects regressions with panel data for investment between 1980 and 2003 for 24 countries in the region (for a list of countries included, see Table 5). For market size, GDP and GDP growth are the representative variables. For macroeconomic stability we choose the exchange rate and annual inflation rate. For trade policy orientation and trade in general we use exports as a percentage of GDP. Our policy environment variable is the prevalence of BITs. For the regressions examining total FDI, we use the total number of BITs signed by that country. Thus, holding other variables constant, we would expect to see increasing investment associated with an increasing number of BITs. In the case of the regressions looking at FDI from the United States, we use a dummy variable indicating whether or not the country has a BIT with the United States. C. Results Our analysis of the determinants of FDI into Latin America and the Caribbean arrives at findings similar to those of previous studies. We find market size, trade orientation, and macroeconomic stability to be the most important determinants of FDI in the hemisphere. We find that the total number of BITs that a country has signed does have an independent effect and positive effect on FDI flows. However, we also find that having a BIT with the United States does not independently attract U.S. FDI.1 1. First, however, we analyze the extent to which a fixed-effects estimation technique is appropriate. To test for this, we ran the regressions conducting estimation by fixed

306 kevin p. gallagher and melissa b.l. birch

The results of our regressions are presented in Table 6, including the results for two dependent variables: total flows of FDI into each country, and flows from the United States. Regressions for total FDI in the entire region show results that confirm some of the earlier studies: market size and export orientation are significant determinants of FDI. In addition, we found that privatization and the total number of BITs are also positive and significant. For U.S. FDI across the region as a whole, we found that, once again, market size and export orientation are key FDI determinants. However, unlike for total FDI, privatization does not have a significant impact on FDI from the United States, a BIT with the United States has no relationship with FDI flows, and there is a negative and significant relationship between literacy and FDI flows from the United States. Tables 6b and 6c report sub-regional variants of these regressions: Table 6b exhibits results for regressions for Mexico, Central America, and the Caribbean (Mesoamerica), while Table 6c reports results for the South American subregion. For Mesoamerica, the regressions clearly indicate that there are two main determinants of total FDI flows in the sub-region—market size (possibly reflecting the dominance of Mexico as a recipient of FDI) and exports as a percentage of GDP—confirming the importance of the use of some these countries as an

effects and by random effects. In both cases we found the fixed effects estimator to be the appropriate one.

table 6a. regression results: all countries TFDI

U.S. FDI

Variable Coefficient BIT LNEXP_GDP LNINFL LNEXCH LNGDP LNLIT LNGDPGR LNGDPPC PRIV_5 N Adj. R2

0.047 0.837 –0.077 0.017 1.398 –0.499 0.004 –0.889 0.028 336 0.703

T 4.89 3.52 –1.39 0.78 5.45 –0.36 0.08 –1.4 2.25 – –

Coefficient 0.291 0.383 –0.050 0.015 1.821 –4.036 0.071 –0.739 0.009 259 0.375

T 0.95 1.17 –0.67 0.58 4.98 –1.97 0.98 –0.81 0.63 – –

do investment agreements attract investment? 307

table 6b. regression results: mesoamerica TFDI

U.S. FDI

Variable Coefficient BIT LNEXP_GDP LNINFL LNEXCH LNGDP LNLIT LNGDPGR LNGDPPC PRIV_5 N Adj. R2

0.031 1.246 –0.118 –0.039 0.885 1.294 0.034 0.675 0.011 190 0.587

T 1.45 3.22 –1.66 –0.34 2.63 0.82 0.59 0.84 0.66 – –

Coefficient 0.212 0.778 –0.129 0.068 2.260 –6.589 0.120 –2.629 0.011 142 0.050

T 0.54 1.87 –1.19 1.32 4.63 –2.78 1.3 –1.96 0.56 – –

table 6c. regression results: south america TFDI

U.S. FDI

Variable Coefficient BIT LNEXP_GDP LNINFL LNEXCH LNGDP LNLIT LNGDPGR LNGDPPC PRIV_5 N Adj. R2

0.043 0.416 –0.036 –0.073 1.796 11.498 0.000 –2.488 0.054 146 0.685

T 3.01 1.02 –0.41 –2.29 4.02 2.76 0 –2.57 2.87 – –

Coefficient 0.118 –0.337 0.054 0.031 1.315 –4.240 –0.020 1.223 0.011 117 0.594

T 0.23 –0.57 0.48 0.7 2.18 –0.71 –0.16 0.91 0.45 – –

308 kevin p. gallagher and melissa b.l. birch

export platform. BITs are not significant, nor is privatization. For U.S. FDI, only the GDP variable shows a significant relationship. For South America, the only significant determinant of U.S. FDI is market size. Once again, a BIT with the United States is not associated with higher FDI flows. In the case of total FDI, market size is important, but so are literacy rates (a positive relationship, unlike for Mesoamerica), privatization and the total number of BITs. Export orientation is not significant. These results thus indicate that there is a substantial difference in the determinants of FDI between the two sub-regions, and that while an increase in the total number of BITs may be conducive to greater FDI in South America, this may not be the case for the countries in Mesoamerica. It seems likely that differences in the types of FDI that the two sub-regions attract make BITs less useful for attracting FDI to the latter than to the former. Overall, these findings are consistent with the work cited earlier. Numerous studies find market size and exports to be key determinants of FDI. Having a large domestic economy and the ability to export are very attractive to foreign investors. We also find a statistically significant relationship between literacy and FDI. Regarding Total Signed BITs, our study is consistent with the UNCTAD study cited earlier but not the World Bank study. Consistent with the Tobin and Rose-Ackerman (2004) study, we found no statistically significant relationship between U.S. BITs and levels of U.S. FDI. This is in contrast with the Salacuse and Sullivan (2005) study. The difference may arise because the Salacuse and Sullivan study only includes three Latin American countries and covers a shorter time horizon

conclusion We perform regressions to examine the determinants of FDI in Latin America and the Caribbean over 23 years. Although a number of countries are aggressively pursuing investment agreements with the United States in order to increase investment, based on past experience we find no evidence that signing such an agreement with the United States will bring increased investment. This is largely consistent with the previous work on the relationship between BITs and investment. Our analysis is also consistent with the broader literature on FDI determinants in Latin America and the Caribbean, which shows that the size of the domestic economy and the ability to serve as an export platform are key drivers of FDI inflows. If the findings in this study and the others on investment treaties and FDI are correct, they suggest that the additional benefits of an investment treaty should be juxtaposed with the costs of lifting performance requirements and of adopting expropriation rules that can be interpreted broadly. In the presence of imperfect competition and other market failures, it has been shown that well-designed

do investment agreements attract investment? 309

performance requirements on foreign investment serve as second-best policy arrangements that can increase welfare (Rodrik 1987). In the case of expropriation cases, nations may stand to pay large compensatory duties for claims that increased costs due to new social and environmental regulations are “tantamount to expropriation.” Additional costs may accrue to such countries if regulations designed to internalize externalities and distribute the benefits of aggregate increases in welfare are rolled back. If investment agreements do not attract additional investment to a developing country, the net benefits of additional investment agreements may be negative. Our analysis is not without limitations. Many of the more qualitative empirical studies in which researchers interviewed firm representatives who make actual decisions about firm location found political stability and the policy environment to be strong factors in addition to those outlined above. Those findings are not to be discounted—econometric analysis is limited in its ability to capture the relative importance of such factors. Our analysis is nonetheless comforting insofar as we confirm many of the factors that interviewed respondents in small samples have discussed with respect to Latin America. To our knowledge, there have no been qualitative studies that examine the specific importance of an investment agreement. Such a study would complement the quantitative analysis presented here, but is unfortunately beyond the scope of this chapter.

references Balasubramanyam, V. N. (2001). “Foreign direct investment in developing countries: determinants and impact,” Conference proceedings, OECD Global Forum on International Investment: New Horizons and Policy Challenges for Foreign Direct Investment in the 21st Century, Mexico City. Baumgarten, S. A. and A. Hausman (2000). “An analysis of U.S. foreign direct investment in Latin America,” Journal of Transnational Management Development, 5, 4, pp. 57–81. Economic Commission for Latin America and the Caribbean (ECLAC) (1998). Foreign Investment in Latin America and the Caribbean, 1998 (Santiago de Chile: ECLAC). —— (2003). Foreign Investment in Latin America and the Caribbean, 2002 (Santiago de Chile: ECLAC). —— (2005). Foreign Investment in Latin America and the Caribbean, 2004 (Santiago de Chile: ECLAC). Graham, E. (2002). Economic Issues Raised by Treatment of Takings Under NAFTA Chapter 11 (Washington, D.C.: Institute for International Economics). Hallward-Dreimeier, M. (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and it might bite,” World Bank Policy Research Working Paper No. 3121 (Washington, D.C.: World Bank). Lim, E.-G. (2001). “Determinants of, and the relation between, foreign direct investment and growth: a summary of the recent literature,” (Washington, D.C.: IMF).

310 kevin p. gallagher and melissa b.l. birch Multilateral Investment Guarantee Agency (MIGA), consulted (2005). “IPANet privatization transaction database,” World Bank website: http://rru.worldbank.org/ Themes/Privatization/. Nunnenkamp, P. (1997). “Foreign direct investment in Latin America in the era of globalized production,” Transnational Corporations, 6, 1, pp. 51–81. Singh, H. and K. W. Jun (1995). “Some new evidence on determinants of foreign direct investment in developing countries,” (Washington, D.C.: World Bank). Rodrik, D. (1987). “The economics of export-performance requirements,” Quarterly Journal of Economics, No. 102, pp. 633–650. Salcacuse, Jeswald and Nicholas Sullivan (2005). “Do BITS really work? An evaluation of bilateral investment treaties and their grand bargain,” Harvard International Law Journal, 46, 167–130. Stallings, B. and W. Perez (2000). Growth, Employment, and Equity: The Impacts of the Economic Reforms in Latin America and the Caribbean (Washington, D.C. and Santiago: Brookings/ECLAC). Trevino, L. J., J. D. Daniels, et al. (2002). “Market reform and FDI in Latin America: an empirical investigation,” Transnational Corporations, 11, 1, pp. 29–48. Tobin, J. and S. Rose-Ackerman (2004). “Foreign direct investment and the business environment in developing countries: the impact of bi-lateral investment treaties,” Working Paper No. 239 (New Haven: Yale Center for Law, Economics, and Policy). Tumman, J. P. and C. F. Emmert (1999). “Explaining Japanese foreign direct investment in Latin America, 1979–1992,” Social Science Quarterly, 80, 3, pp. 539–556. Tumman, J. P. and C. F. Emmert (2004). “The political economy of U.S. foreign direct investment in Latin America: a reappraisal,” Latin American Research Review, 39, 3, pp. 9–29. United Nations Conference on Trade and Development (UNCTAD) (1998). Bilateral Investment Treaties in the Mid-1990s (New York: United Nations). —— (2000a). Bilateral Investment Treaties, 1959–1999 (Geneva, UNCTAD). —— (2000b). World Investment Report 2000: Cross Border Mergers and Acquisitions and Development. (Geneva: UNCTAD). —— (2004). “World Investment Report Country Fact Sheets 2004,” UNCTAD website: http://www.unctad.org/. Vallejo G., H. and C. Aguilar (2002). “Integracion economica y atraccion de inversion extranjera directa: el caso de America Latina,” (CEDE: Universidad de los Andes).

11. the global bit s regime and the domestic environment for investment susan rose-ackerman introduction Under what conditions do bilateral investment treaties (BITs) stimulate foreign direct investment (FDI)? BITs are touted as a potent tool for attracting FDI, and recent empirical work suggests that potential host countries sign BITs in an effort to improve their attractiveness as sites for FDI. The proliferation of BITs is a result of competitive economic pressures among developing countries to capture a share of foreign investment (Elkins, Guzman, and Simon 2006). Yet the empirical basis for the claim that BITs stimulate FDI is weak. Studies, including a number included in this volume, come to conflicting conclusions. Salacuse and Sullivan (2004) and Neumayer and Spess (2005) found strong correlations between BITs and FDI flows on both a bilateral and a general basis. At the same time, using a different set of models and assumptions, HallwardDriemeier (2003) and Tobin and Rose-Ackerman (2005) found little evidence of this connection. The recent empirical and theoretical work by Tobin and RoseAckerman (2008) and Bubb and Rose-Ackerman (2007) summarized here suggests that one reason for the conflicting results is a failure to develop a convincing conceptual framework. Based on a more nuanced model of the relationship between BITs and FDI, we claim that BITs do indeed stimulate FDI inflows but under a more complex set of conditions than recognized in previous empirical work. There are two findings in my coauthored work. First, the impact of BITs on FDI depends not just on a country’s own supply of BITs but also on the emerging global BITs regime. One country’s ability to attract FDI is a function of the investment environment in other countries competing for a limited supply of funds. Building on Andrew Guzman’s (1998) article exploring this interaction, Ryan Bubb and I go beyond Guzman’s work to provide a formal model of the interaction that highlights some limitations of his argument. My work with Jennifer Tobin then provides an empirical test of the interconnection. We find that the marginal impact of a country’s own BITs on its ability to attract FDI falls as the global coverage of BITs grows because the host country must then compete with a growing number of more equally situated investment locations. Second, we show more convincingly than past work that BITs cannot entirely compensate for an otherwise weak investment environment. Their presence interacts with the political–economic environment to affect FDI flows. In Tobin and Rose-Ackerman (2008) we show that the marginal impact of BITs is greater

312 susan rose-ackerman

in countries that already have relatively effective legal regimes and favorable economic environments. This chapter proceeds as follows. Section A introduces the basic theoretical framework developed in Bubb and Rose-Ackerman (2007) and in Tobin and Rose-Ackerman (2008). Section B then summarizes the empirical finding of Tobin and Rose-Ackerman. Section C offers some thoughts on the rise of the BITs regime and its implications for the possibility of a Multilateral Investment Agreement. A. The impact of BITs on FDI To begin, first consider how BITs might affect FDI. Under one view, a treaty between a particular home country and a particular host country would stimulate FDI only from that home country. FDI from other countries might fall as a result of the more favorable treatment of investments from a particular source. Alternatively, the approach taken in my coauthored work stresses the interaction between domestic institutions and BITs. We posit that signing and ratifying a BIT with a wealthy country signals a commitment to treat foreign investors as well as or better than domestic investors. The strength of that signal is enhanced as the country ratifies more and more BITs. To see why this is a plausible claim, suppose that developing countries with good environments for investors have difficulty credibly signaling that fact through domestic publicity efforts. They need a mechanism that places an international imprimatur on their actions. Signing and ratifying BITs can provide that signal. A country that has many BITs in force is likely to find it worthwhile to strengthen domestic rule-of-law institutions that apply across the board. This claim assumes that there are economies of scale for governments in having uniform rules for investment. Rather than treat each FDI project as a special case, governments reform the background conditions for all investors. If investors believe that this connection holds, then a country can signal a favorable environment by signing BITs. However, other information about political risk is also considered by investors. If the information, provided, say, by commercial investment advisory services indicates high risk, BITs may cause investors favorably to update their views of this risk, but they cannot entirely overcome the negative impact of such reports. Notice that the signaling effect of BITs will only succeed if BITs actually do constrain developing countries from expropriating or harassing foreign investors. The treaties’ arbitration provisions are a key aspect of their effectiveness as a signal—a signal that becomes clearer and more credible as the number of BITs with developed countries increases. In the riskiest countries in the world, BITs cannot do much to stimulate additional FDI; they can only marginally affect the underlying high risk of investment, and such countries are unlikely to enforce arbitral decisions that go against them. For less risky countries, BITs signal that the country is committed to a favorable environment for investment, and the greater the number of BITs, the stronger the signal. This signal is credible because BITs give investors the ability

the global bit s regime and the domestic environment for investment 313

to threaten costly international arbitration if the treaty terms are not complied with. Such countries both have better domestic conditions and are more likely to comply with their international agreements. We hypothesize that the signaling function of BITs explains their popularity in the developing world. However, because BITs only allow investors to update their prior estimates of risk, we hypothesize that, all else equal, BITs will have the lowest marginal impact on FDI in the riskiest environments and will have a larger positive effect as political risk moderates. If the total number of BITs in force with developed countries helps to signal the quality of the domestic investment environment, then one would expect that a count of BITs with Organisation for Economic Co-operation and Development (OECD) countries interacts with domestic economic and political variables to enhance the productivity of FDI. Three types of interactions seem important. They are, first, the interaction between the number of BITs and economic conditions inside the country that influence FDI; second, the vast differences in the size and resource endowment of the countries trying to obtain a slice of the FDI pie; and third, the way the political environment may interact with BITs to influence FDI levels. Consider next the global BITs regime. Multinational enterprises (MNEs) can locate throughout the world and will select the locations that promise the best opportunities for profit. If very few countries have BITs, those that do ought to be especially attractive locations for investment so long as there are other positive reasons to invest. However, as more and more countries sign more and more BITs, the relative benefit of an extra BIT can be expected to fall as countries compete for FDI. As Guzman (1998) argued, if the coverage of BITs expands, the marginal benefit to any one country falls—the benefits are positive but not as great as when only a few states were willing to commit to BITs. Even though a country enhances the credibility of its legal and institutional environment, it may attract little additional investment if other potential investment sites also upgrade their environment and signal their behavior by ratifying BITs. True, the total amount of FDI directed toward developing and emerging economies is likely to rise with a rise in the count of BITs, but inter-state competition for investment will dampen the impact on any one country. Ryan Bubb and I (2007) developed a formal model of this possibility, the basic logic of which is easy to state. If no other developing country has signed BITs, then if one country signs such treaties, we assume that this will lead to a substantial inflow of FDI relative to a no-BIT situation. Because BITs provide benefits to foreign investors, however, the share of the total gains that goes to the developing country falls. If few other BITs exist, we assume that the increase in investment flows outweighs the reduced share of the gains. To take a simple example, suppose that without BITs the developing country negotiates a deal to receive 80% of the gains from investment and attracts FDI that produces $100 million in gross profits for a benefit of $80 million. With BITs, the risk of investment falls, and the country attracts FDI that generates $200 million in profits,

314 susan rose-ackerman

but the share of the gains falls to 60% for a benefit of $120 million. Here, the gain in volume offsets the fall in the share, and the developing country benefits from signing BITs. However, as more and more countries sign BITs, the marginal gain to signing BITs falls. This occurs because there is more competition for FDI from other countries with comparable protections for investors. Of course, the marginal benefit of staying out of the BITs regime also falls as more and more countries join the BITs regime. Hence, the coverage of BITs may continue to increase. Depending on the relative benefits of staying out or opting into the BITs regime, a equilibrium can occur either with all developing countries choosing to adopt BITs or at an intermediate level of coverage where the increase in the volume of investment is just balanced by the fall in share of gains going to host countries. Guzman (1998) concluded that the recent growth in the number of BITs is likely to have been harmful for developing countries as a group compared to a global investment environment governed by international customary law. Bubb and I show that this conclusion does not necessarily hold. A BITs regime might be superior for developing countries compared to a regime of customary international law even if the marginal gains from signing treaties fall as the global coverage of treaties increases (Bubb and Rose-Ackerman 2007). Of course, domestic economic and political conditions also affect the impact of BITs on FDI inflows. These conditions not only have a direct effect on FDI, but also interact with the number of BITs. In other words, taking the number of BITs as a signal of a state’s commitment to protect foreign investors, this signal enhances investors’ response to economic and political conditions. If investors expect their rate of return on investment to be higher and/or their risk to be lower the higher the number of BITs, then, given any set of underlying conditions, they are likely to invest proportionately more. BITs enhance the value of the other determinants of FDI. Although part of the impact of BITs in very weak states may be to reassure investors who know they face a risky institutional environment, Tobin and I posit that BITs’ main role is to enhance the value of other information about the institutional environment. BITs are not a suitable substitute for domestic institutions. Rather, for emerging economies that need to convince investors of their credible commitments to property rights and the rule of law, the number of BITs in force with wealthy countries sends a positive signal in interaction with information on other conditions. B. Empirical results My work with Jennifer Tobin seeks to explain real, net FDI inflows to low- and middle-income host countries between 1980 and 2003.1 These flows constitute between 20% and 40% of total FDI depending upon the year. We are not concerned with FDI between wealthy countries, most of which have strong domestic 1. This work is currently under review at a scholarly journal. Here, I summarize the conclusions and present one illustrative figure.

the global bit s regime and the domestic environment for investment 315

legal environments and do not sign BITs with each other. We average our data over five years from 1980 to 2003 (four years for the last period) to avoid the impact of year-to-year variation caused by the patterns of individual deals. The independent variable of most interest is a count of the BITs a country has in force with OECD countries because they are the treaties with the potential to have an impact on FDI flows during the time period of our study. Because we expect the BIT signal to operate with a lag, we use BIT data from the previous five year period to estimate FDI in any particular period. Our proxy for the political environment for investment is an index of political risk from the International Country Risk Guide (ICRG). This variable is based on institutional indicators compiled by private international investment risk services. The ICRG political risk index includes measures of the risk of expropriation, established mechanisms for dispute resolution, contract enforcement, government credibility, corruption in government, and the quality of the bureaucracy (Jun and Singh 1996; Wei 2000; Blonigen 2005). The next group of variables seeks to capture the economic determinants of FDI. First, better basic economic health in a country is likely to attract FDI and is captured by the log of GDP per capita (income) and the rate of growth of a country’s economy. Second, the country’s openness to trade is likely to have an impact on FDI, with high trade barriers attracting horizontal FDI (tariff-jumping), while more open economies attract vertical FDI. Finally, we proxy the absorptive capacity of the country in two ways. First, we include population as a measure of market size, given the level of per capita income. Second, the presence of natural resources in a country is expected to attract foreign investment regardless of other factors that would usually attract or discourage investors. To test our hypotheses we estimated three specifications. Our base model for foreign direct investment to low- and middle-income countries mirrors that of past studies, and seeks to explain real FDI flows to a given low-income country in a particular time period as a function of (a) the number of OECD BITs in force with that country, (b) the total number of BITs in the world with OECD countries, (c) an index of the political environment for investment, (d) per capita income, (e) population, (f ) economic growth, (g) natural resource trade, (h) openness to trade, (i) a random error, and (j) fixed country and time effects. Although this first model is useful for understanding the independent effect of BITs on FDI inflows, our theory leads us to believe that this effect is conditional. The next two specifications test the contingent relationship of BITs of a given country with world BITs and with the country’s political–economic environment. Our results are consistent with our theory concerning BITs although the results for other determinants of FDI differ from previous work. If BITs are considered as an independent determinant of FDI, we find that they have a positive and statistically significant effect. GDP per capita and population are the only other significant determinants, but in opposite directions—GDP is positively and population is negatively related to FDI totals. In past research, the investigations

316 susan rose-ackerman

have stopped at this point. Our theory, however, leads us to believe that BITs primarily affect FDI through their interaction with other factors. Thus, our next step was to interact BITs with the other domestic determinants of FDI and with the world total of BITs. Our results indicate that as the number of BITs in the world increases, signing an additional BIT has a decreasing marginal effect on FDI flows to the host country. Further, aside from GDP and openness, the individual interactive effects with BITs are significantly different from zero at least at the 90% confidence level. This lends support to our claim that the effect of BITs on FDI flows is conditional on the general political-economic environment for FDI. Additionally, the R-squared increases from 0.17 in the linear model to 0.27 in the fully interactive model. Thus, we are confident that our focus on the interaction between BITs and other factors provides an important window into their impact on FDI. Unfortunately, given the multiple interactions, we can not safely interpret the coefficient estimates. Many of the interaction terms are highly correlated with each other and with their underlying components. This high level of correlation results in large standard errors, making it difficult to understand the specific relationships between the independent variables and FDI. Despite the low significance levels, we can, nevertheless, conclude that the simple linear model is incorrectly specified—our second formulation explains much more of the variance. Having demonstrated that the linear specification is incorrect, we sought a more parsimonious specification that limits the problems of collinearity. To do this we estimated a model that included two indicators as well as interactions with world BITs and the political environment for investment. The first indicator groups together the economic variables that have been normalized for the size of the country, while the second includes the measures of absorptive capacity. All four interactions are significant at least at a 90% level of confidence and have the expected signs. To illustrate our results in a series of concrete cases, Figure 1 graphs the results for a sample of countries. We ranked countries according to the indicator that measures the economic environment, and randomly selected two countries from each quartile. The countries were: Sudan and Afghanistan (lowest quartile), Paraguay and Malawi (second lowest quartile), Senegal and Gambia (third quartile), and Malaysia and Chile (top quartile). We then estimated their expected net FDI flows based on their actual number of BITs, their index of the political environment for investment, the economic indicator, and the absorptive capacity indicator at the point in time matching the number of world BITs. For countries with good investment environments and strong political–economic environments (Malaysia and Chile, for example), an additional BIT always has a positive impact on estimated FDI flows, although the effect eventually decreases along with the number of BITs in the world. However, for countries with weaker political–economic and investment environments, this positive impact is increasingly smaller, with those in the lowest quartile (Sudan and Afghanistan, for example) reaching a point of zero impact well before the current number of existing BITs in the world.

the global bit s regime and the domestic environment for investment 317

figure 1. conditional impact of an additional bit on fdi at various levels of world bit s : 8 country estimates FDI (Billions of $) 45 Malaysia 40

Chile Georgia

35

Paraguay 30 Senegal I) 25

Malawi

20

Chile (1) Malaysia (1) Georgia (2) Senegal (2) Paraguay (3) Malawi (3) Sudan (4) Afghanistan (4)

15 10

Sudan

Afghanistan

5 0 38.8

63.6

93.4

133.2

192.6

347

654

905

World BITs

Source: Tobin and Rose-Ackerman (2008)

Hence, when there were relatively few BITs in the world, regardless of a country’s political environment for investment, level of economic development, or capacity to absorb FDI, signing an additional BIT significantly and positively affected FDI flows—and this impact was not much different across quartiles. For most countries, this positive impact increased for at least one to two periods before beginning a decline. In other words, when there were few BITs in the world, not only did signing a BIT send a positive signal to foreign investors, but the signal became more pronounced as investors took BITs into account. However, as BITs became more and more popular, their ability to serve as a signal of a superior investment environment faltered. In the current climate with over 1,000 BITs in force with OECD countries, the least risky and most developed of the developing countries gain much more from signing additional BITs than more risky countries. In short, countries outside the top quartile should not rush to sign BITs; rather they need to concentrate on improving their domestic investment environment through efforts to limit political risk and assure all investors a secure environment. Only then can BITs be an important tool for the encouragement of FDI. C. The History of BITs and the Failure to Negotiate a Multilateral Agreement on Investment To conclude, consider the implications of our results for the debate over the reasons for the growth of BITs and the failure to conclude a multilateral investment treaty. Bubb and I (2007) argue that in the 1960s, the newly decolonized nations had little to gain by respecting investments made in the past. Even if the

318 susan rose-ackerman

BITs regime provided higher payoffs from future investment than a non-BIT regime, the availability of bilateral treaties as a commitment mechanism may have created an incentive for poor countries to expropriate past investments. Their optimal strategy would be to expropriate past investments to extract value but then to commit to respecting property rights in the future. The availability of BITs as a credible commitment device going forward created a commitment problem for rich countries. Rich countries could not credibly commit to punish poor countries for expropriation by refusing to invest in the future or by refusing to sign BITs. A BIT provided a way for a country to make a clean break with the past and announce a newfound commitment to protecting the property of foreign investors. This explanation is consistent with the history of expropriations of foreign investors. As documented by Kobrin (1984) and Minor (1994), expropriations of foreign investors by developing countries increased throughout the 1960s and peaked shortly after the New International Economic Order resolutions were passed in the United Nations General Assembly in the early 1970s. Subsequently, developing countries rushed to sign BITs and expropriations became increasingly rare. Thus, contrary to Guzman’s (1998) suggestion, resistance by developing countries to a stringent compensation standard in customary international law does not imply that BITs are harmful to developing countries. This explanation also sheds some light on an empirical puzzle: the timing of the sudden increase in BIT signings that occurred in the 1990s. Although the first BIT was signed in 1959, by the end of 1989 there were only 385 BITs in the world economy. The number expanded rapidly during the nineties, and although the rate of growth has now fallen off, the total has continued to rise to over 2,500 at the end of 2006. In addition, 240 other types of agreements, mostly dealing with trade, also include trade chapters (UNCTAD 2007a, p. 15; 2007b, p. 2). The conceptual framework outlined above does not provide an explanation for the pattern of BIT signings over time. However, the alternative explanation offered here suggests that developing countries had an incentive to wait until they had extracted more value from investments left from the colonial period before entering into BITs. Almost all the early treaties were signed between Germany or Switzerland, on the one hand, and poor countries, mostly in Africa, on the other. Of the 56 BITs signed between 1959 and 1967, 45 were signed by either Germany or Switzerland. Neither was a major colonial power, and they signed BITs with countries that would not have had much German or Swiss capital to expropriate. Of course, as Bubb and I recognize, this is just a hypothesis that requires further historical analysis and empirical testing, but it is roughly consistent with the observed pattern. Newly independent states appear to have faced a tradeoff between expropriating the capital stock left over from the colonial period and providing guarantees designed to encourage new investment. Most BITs, however, cover past investments, so this suggestion needs to be tested against the experience of BIT negotiators who dealt with the issue of property expropriated

the global bit s regime and the domestic environment for investment 319

before BITs were signed. Did developed countries simply write off past actions by newly independent states as a fait accompli, or did they try to structure the BITs to permit compensation claims based on expropriations prior to the BIT?2 The rise of BITs, however, does not explain why the same countries that signed BITs were unable to conclude a multilateral agreement on investment (MAI). It should be possible to create an MAI that would produce larger aggregate benefits from investment than any set of bilateral arrangements, because providing a single uniform multilateral regime reduces the transaction costs of foreign investment.3 Furthermore, if the existing set of bilateral relationships does not provide legal protections for investment flows between every country pair, an MAI could improve on it by extending legal protection to those investment flows. It could, for example, cover investment flows between rich countries as well as any rich-poor flows not covered by existing BITs. The surplus from this simplification and rationalization of the legal framework protecting foreign investment must be divided between the rich and poor countries so that all prefer the outcome compared to the status quo. Terms in an MAI that would affect this division (and potentially also the size of the surplus) include the amount of flexibility poor countries retain in regulating foreign investment (e.g., requiring foreign investors to purchase local inputs) and any environmental or other standards imposed on investors. If the MAI must obtain unanimous support, it will be concluded only if the division of the resulting surplus is individually rational for each country, that is, if the payoffs under the MAI exceed the equilibrium payoffs of the BITs regime. Negotiating and concluding an MAI is a costly process. As the surplus achievable by an MAI narrows, it is less likely that countries will be willing to bear the transaction costs of creating an MAI. Thus as Bubb and I (2007) argued, an MAI is

2. Three other factors contributed to the rapid increase in BITs. One was the dissolution of the Soviet Union and the consequent emergence of formerly socialist states in Central and Eastern Europe. Those countries signed 455 BITs in the 1990s or 29% of all those signed in the decade (UNCTAD 2000, 4). Second, the United Nations Conference on Trade and Development pushed BITs as a way of strengthening investment cooperation between developing countries beginning in 1999 (UNCTAD 2000, 2–3), and this effort led to an increase in treaties between developing countries. Whether such BITs have actually increased cross-border investment flows remains to be seen. Investment inflows to very low income countries remain at very low levels. Finally, Santiago Montt (forthcoming 2009) has suggested that the increase in BITs reflected learning by the international investment community. A new type of treaty gradually became accepted and produced a bar specialized in the application of BITs. 3. An MAI could provide weakly larger aggregate benefits since at the very least it could simply reproduce the exact set of legal rules contained in a BITs regime. An MAI could likely improve on this complex regime by codifying these legal rules into a single code, as well as by improving on the substantive rules. The complexity of the current regime of BITs imposes costs on both investors and host countries (UNCTAD 2007b).

320 susan rose-ackerman

more likely to result from a status quo with few or no BITs, where the surplus is larger, than from a full bilateral treaty equilibrium, where some of the gains from a treaty are already realized. In the face of bargaining costs, the proliferation of BITs may well have made it more difficult to conclude an MAI by narrowing the surplus a multilateral framework could create. The failure of the 1998 OECD MAI thus appears, in part, attributable to the success of BITs. Given the strong protections provided by BITs, an MAI would have to offer protections to foreign investors at least as strong as those provided by BITs. And, indeed, the terms of the draft MAI largely followed the U.S. model BIT of the time, which provided stronger protections than most other BITs (Muchlinski 2000, p. 1041). However, some provisions in the U. S. model proved to be too strong for some OECD members. Without a large surplus to be gained from an MAI, given the success of BITs, and with perceived costs of opening new domestic sectors to foreign investment, some OECD members balked at forming an MAI. Under an MAI, developed states would be unable to target protections solely to inward foreign investors from net capital-importing nations. Despite their reciprocal terms, existing BITs largely cover one-way flows of investment between states party to them. The capital-exporting state desires protections for its nationals’ investments abroad, and the capital-importing state desires to encourage investment. However, some capital-exporting states that also import capital do not want to open up certain sensitive sectors of their economy to investment from abroad. BITs allow these countries to selectively grant rights that few will use in return for rights that protect their nationals’ foreign investments. But a multilateral agreement would necessarily include many large capital exporters within its ambit, and an MAI that provides rights to establish foreign investments in member states would potentially result in new and unwelcome investment flows into large developed states.

conclusion Thus, my research with Jennifer Tobin and Ryan Bubb has produced both theoretical and empirical advances. We have systematized Guzman’s (1997) seminal paper by showing how the growth of the global BITs regime has reduced the marginal gains that host countries obtain from BITs. Nevertheless, contrary to Guzman’s suggestion, the resulting FDI regime may still be better for developing countries that one based on customary international law. Although we cannot directly test that last empirical claim, Jennifer Tobin and I found an interaction between global BITs and the impact of host country BITs. Even if overall FDI flows to developing countries increase as total world BITs rise, we show that the marginal effect of a country’s own BITs on its own FDI decreases. Thus, if there are other costs associated with BITs, countries may be

the global bit s regime and the domestic environment for investment 321

less eager to sign such treaties over time. If BITs both limit the country’s profit per dollar of FDI and increase the volume of investment, a fall in the marginal impact of BITs on the volume of FDI will reduce the net benefits of BITs. Overall, BITs have a positive impact on FDI flows to developing countries but only in interaction with the political and economic environment for investment. Signing treaties with OECD countries generally appears to send a signal to foreign investors of a welcome investment environment, and that signal is stronger the greater the number of BITs in force. Other indications of the strength of the political environment for investment and of a better local economic environment are complements to BITs. Poor countries cannot bootstrap an aggressive program of signing BITs into a major increase in FDI. They cannot avoid the hard work of improving their own domestic environment for investment. In short, an individual host country’s BITs cannot be judged in isolation. The impact of BITs on a country’s FDI flows must be studied within the context of its political, economic and institutional environment and in the light of the worldwide BITs regime. Only with a broader understanding of the political–economic environment can we fully understand the impact of BIT programs on FDI flows.

references Blonigen, B. (2005). “A review of the empirical literature on FDI determinants,” Atlantic Economic Journal 33, 4, pp. 383–403. Bubb, Ryan and Susan Rose-Ackerman (2007). “BITs and bargains: strategic aspects of bilateral and multilateral regulation of foreign investment,” International Review of Law and Economics, 27, pp. 291–311. Elkins, Z., A. Guzman and B.Simon (2006). “Competing for capital: the diffusion of bilateral investment treaties, 1960–2000,” International Organization, 60, pp. 811–846. Guzmán, Andrew (1998). “Why developing countries sign treaties that hurt them: explaining the popularity of BITs,” Virginia Journal of International Law, 38, pp. 643–688. Hallward-Driemeier, M. (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and they could bite,” World Bank Policy Research Working Paper No. 3121 (Washington, D.C.: World Bank). Jun, K. and H. Singh (1996). “The determinants of foreign direct investment in developing countries,” Transnational Corporations, 5, 2, pp. 67–105. Kobrin, Stephen J. (1984). “Expropriations as an attempt to control foreign firms in LDCs: trends from 1960 to 1979,” International Studies Quarterly, 28, pp. 329–348. Minor, Michael S. (1994). “The demise of expropriation as an instrument of LDC policy, 1980–1992,” Journal of International Business Studies, 25, pp. 177–188. Montt, Santiago (forthcoming 2009). State liability in investment arbitration: Global constitutional and administrative law in the BIT generation. (Oxford, Hart Publishing). Muchlinski, Peter T. (2000). “The rise and fall of the multilateral agreement on investment: where now?” The International Lawyer, 34, pp. 1033–1053.

322 susan rose-ackerman Neumayer, E. and L. Spess (2005). “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33, 10, pp. 1567–1585. Salacuse, J. and N. Sullivan (2004). “Do BITs really work? An evaluation of bilateral investment treaties and their grand bargain,” Harvard International Law Journal, 46, 1, pp. 67–130. Tobin, Jennifer and Susan Rose-Ackerman (2005). “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Research Paper No. 293 (New Haven, CT, Yale Law School Center for Law, Economics and Public Policy). Tobin, J. and S. Rose-Ackerman (2008). “When BITs have some bite: the political economic environment for bilateral investment treaties,” (New Haven, CT: Yale University). United Nations Conference on Trade and Development (UNCTAD) (2007a). Bilateral Investment Treaties 1995–2007 (New York and Geneva, United Nations). —— (2007b). Development Implications of International Investment Agreements, IIA Monitor, No. 2 (United Nations: New York and Geneva). —— (2000). Bilateral Investment Treaties, 1959–1999. (New York and Geneva: United Nations). Wei, Shang-Jin (2000). “How taxing is corruption on international investors?” Review of Economics and Statistics, 82, 1, pp. 1–11.

12. the impact on foreign direct investment of bit s ∗ unctad introduction Bilateral investment treaties (BITs) are concluded by host countries to attract foreign direct investment (FDI). This chapter examines whether the conclusion of BITs does indeed contribute to an increase in FDI.1 The reasons why the conclusion of BITs should have a positive effect on FDI were discussed in greater detail in chapter 1; the most important reasons are that BITs strengthen the standards of protection and treatment of foreign investors, facilitate entry into a host country, and establish mechanisms for dispute settlement. Before examining the statistical analysis, a number of caveats need to be made. • FDI is a highly sophisticated international transaction, which involves (unlike, say, trade transactions) the engagement of often considerable assets abroad, a long-term commitment, and all the usual requirements for a successful investment project (such as good prospects for sustainable profitability and acceptable risk/profitability ratios). No single determinant explains FDI flows. Typically, many conditions and requirements must be met for an investment project to take place, and it is sometimes difficult to identify the conditions that play a decisive role. • To make matters more complicated, the relative importance of these conditions (and thus the importance of FDI determinants, including BITs) differs according to the type of investment (e.g., natural resources-seeking or market-seeking), the type of investor (e.g., services or manufacturing multinational enterprises [MNEs], small and medium-size or large MNEs), and the perspective from which they are viewed (that of the home or the host country). Nevertheless, the literature identifies a number of determinants as more important than others in influencing FDI flows (UNCTC, 1992). Some of theses determinants will be included in the analysis that follows ∗ This chapter was reprinted with permission from UNCTAD. The chapter was originally published in UNCTAD, Bilateral Investment Treaties in the Mid-1990s, chapter IV (New York and Geneva: United Nations, 1998). 1. The relationship could also work the other way around: existing FDI can stimulate BITs. Agencies responsible for international economic relations in developed countries report occasionally that they are encouraged by firms to conclude BITs with host countries in which they have already invested.

324 unctad

• The conclusion of BITs by host countries is only one element in a wide spectrum of policies dealing with FDI, and, as such, it is only one among many policy determinants. FDI, especially in developing countries, has often been governed by special investment codes. It is also influenced, in all countries, by internal regulations concerning the conduct of business, such as tax codes, subsidies, and foreign exchange regulations. Indeed, many of the policy changes in recent years aimed at attracting FDI—such as the liberalization of FDI policies and the relaxation of performance requirements—have taken place alongside the growth of BITs. However, policies themselves, especially liberalization policies, though they are undoubtedly a necessary precondition for FDI (since without them, FDI projects could not be undertaken), have only a slight influence on investment flows (UNCTC, 1991, pp. 59–60).2 • The convergence of the FDI policies of host countries and the promotion of FDI, including through BITs, somewhat neutralizes the advantage that host countries with more liberalized policies and more advanced promotional efforts once enjoyed over other countries, thus reducing the importance of policies, including BITs, as FDI determinants and increasing the importance of other factors. • Finally, the function of BITs can be fulfilled by other instruments, such as regional agreements. A. Theoretical Considerations: A Two-Stage Analysis The analysis that follows is based on the theoretical argument that, for the reasons outlined above, the conclusion of a BIT between two countries—a developed country and a developing country or a country in transition—may have an impact on FDI flows (or other measures of MNE investment) from a developed country (the home country) to a developing country or a country in transition (the host country) in the years following the conclusion of the BIT. There is no commonly accepted theory of the response lag of investors to policy changes, including to the conclusion of BITs. Some studies (e.g., UNCTC, 1991) that have traced the response of investors to changes in government policy announcements suggest a quick response of one to two years. Others (e.g., Kreinin, Plummer and Abe, 1997) suggest a longer lag of up to five years. There is evidence from the trade area that firms may anticipate policy changes, such as trade policy liberalization, and increase their exports before the changes occur. For example, even before the Treaty of Rome established the European Economic Community, trade among member countries had accelerated. To deal with this 2. It should be noted, however, that it is very difficult to measure policy changes in the area of FDI, including changes in the degree of liberalization and degree of openness to FDI. Therefore, analyses of the impact of liberalization and other policy changes on FDI flows are not very precise.

the impact on foreign direct investment of bit s 325

issue, different lags (including a zero lag and negative lags) will be examined to find the best statistical fit. To allow this, the behaviour of FDI flows will be examined during the five years before and the five years after the conclusion of a BIT. As regards the choice between the two possible dates of the conclusion of a BIT—the date of signing and the date of ratification—the former has been selected as an independent variable for the analysis. Because the great majority of BITs are ratified, it is reasonable to assume that, in the perception of investors, signing a BIT is the crucial action: once a BIT is signed, or expected to be signed, the market has absorbed it or begins to absorb it. This theoretical argument and the resulting selection of variables is obviously based on certain simplifying assumptions and, therefore, is not necessarily watertight. For example: • It cannot be ruled out that a BIT continues to have an impact on FDI flows as long as it is in force, and not just during the five years after its conclusion.3 BITs may also prevent declines in FDI by keeping existing investors in the host country. • The argument treats all BITs as equal, but in fact they may differ in scope and force, and thus in their importance to investors. • BITs with identical clauses may have a different “enforcement value” for MNEs from different home countries, depending on the strength of a corporation’s home government (Conklin and Lecraw, 1997). The hypothesis examines the relationship between BITs and FDI flows in the most direct way: by examining the FDI flows between the countries concerned. It will be tested in two stages: 1. Stage one In stage one of the analysis, the relationship is tested by using FDI data covering eleven years for individual BITs. Such data are available for 200 BITs signed between fourteen home countries and seventy-two host countries, which makes it possible to draw meaningful conclusions about the relationship, even though more than 1,100 BITs and many participating countries are excluded from the analysis. 2. Stage two In order to expand the coverage of BITs and countries, additional analysis is undertaken in stage two on the basis of data on total FDI flows and stocks of host countries and, for some dependent variables, on the total number of BITs. As these two variables are more easily available than information on

3. This may happen for a number of reasons. For example, a host country signing a BIT may not yet be able to offer the other locational advantages that are necessary to attract FDI; once it is able to offer these advantages, FDI may rise. Also, in a number of former centrally planned economies that already had BITs with home countries, the flow of FDI only became noticeable when these countries began the transition to a market-based economic system.

326 unctad

bilateral FDI flows, the number of host countries covered rises to 133, with a corresponding increase in the number of BITs covered. The modification of variables, by introducing total FDI stocks and flows and a cumulative number of BITs, also modifies the theoretical argument. For example, a possible relationship between the total number of BITs and total stock of FDI is based on the assumption that countries with a greater number of BITs receive more bilateral flows from home countries with which they have concluded the BITs than with countries that have a smaller number of BITs and that, consequently, these higher flows lead to larger stocks, either in absolute terms or relative to the size of their economies. By the same token, if total flows of FDI are taken as the dependent variable, the assumption is that countries with a greater number of BITs signed during the period under consideration receive more bilateral flows and that this is reflected in their total flows (in absolute or relative terms). In stage two, an additional question is raised: If the effect of BITs on FDI is weak, or nonexistent, what other variables have a greater influence on FDI? These variables might include host countries’ market size and growth, countryrisk indices, changes in exchange rates, inflation rates, and capital investment. Moreover, by relating total stock of FDI to the total number of BITs of host countries, stage two avoids one of the weaknesses of stage one, whereby consideration of the impact is limited to a period of eleven years. B. Stage One: Analysis of Time-Series Data for Individual Bilateral Investment Treaties 1. Data and methodology In this stage of the analysis, the focus is on changes in FDI flows between pairs of countries as a result of the signing of BITs. The analysis of FDI flows between pairs of countries over several years provides a time-based perspective that the cross-country analysis in stage two cannot give. Obtaining bilateral FDI data in connection with individual BITs is not easy. The 200 observations that were collected refer to the bilateral flows between seventytwo FDI-recipient countries and fourteen home countries from 1971 through 1994. They were obtained from a variety of sources, including the United Nations Conference on Trade and Development, the Organisation for Economic Co-operation and Development, and the governments or central banks of some developed countries. Four FDI indicators will be used as dependent variables: a. FDI: FDI flows between the pair of countries that have signed BITs (in millions of dollars). b. FDI/gross domestic product (GDP): The ratio of FDI over GDP of the host country. The FDI/GDP ratio corrects the imbalance created by large countries in a data set. It supplements an FDI growth index based on absolute figures by relating it to a host country’s economic growth.

the impact on foreign direct investment of bit s 327

c. FDI/inflow: The share of a home-country partner to a BIT in a host country’s total FDI inflow. For example, if Zambia and Japan sign a BIT and Japan’s share in Zambia’s FDI inflows increases, one can suggest that the BIT is associated with Japan’s increased share of FDI in Zambia. d. FDI/outflow: The share of a particular host country in a home country’s total FDI outflows. For example, if Mali and France sign a BIT and Mali’s share in France’s total FDI outflow increases, one can suggest that the BIT is associated with Mali’s increased share in France’s FDI outflows. Since FDI flows have been increasing generally over time for most countries, especially in the past decade, it may be argued that higher absolute FDI figures in the years following a BIT may only reflect the general rising trend, not specifically the impact of a BIT. The addition of the two share variables (which are based on relative rather than absolute figures) is intended to correct this bias. If either of these increases follows the signing of a BIT and is statistically significant, then one is on firmer ground in asserting that it is the BIT that is associated with the increase, rather than the general in FDI. For each pair of BIT countries, the data on these four variables were recorded for five years before and five years after the BIT was signed. For each pair, Year 0 is called the BIT year; the data go from year –5 to year +5, as follows: Year –5

Year 0

Year +5

BIT year Because the data for a single pair of countries generate a set of observations that is too small to allow for statistical testing, testing has to be done for the entire set of observations and for groups of countries. For each before/after BIT comparison, a t-test of differences in group means was conducted to test for significant differences. Although simple, this is a robust technique (even given a moderate amount of missing data). Because the hypothesis is that BITs increase FDI, a one-tailed t-test is appropriate. The main objective is to determine how the mean values of FDI in the years before a BIT was signed compare with the mean values of FDI in the years after the BIT was signed. However, as mentioned earlier, it is not known exactly when investors begin to react to the signing of a BIT: hence, to determine also any lags (or advance reactions) in the effects of BITs, the mean values of dependent variables for 200 BITs were grouped in various combinations of two time periods, each resulting from different possible splits of the eleven years covered by the analysis. This included not only combinations of data for the years before and after the BIT year (thus using the BIT year as a cut-off date) but also combinations

328 unctad

using other cut-off years, comparing, for example, the period between Year –4 and Year +1 with the period between Year +2 and Year +5. The comparisons thus involve periods of varying length before and after the cut-off year, with a minimum of two and a maximum of five years on either side. A total of fifty-four pairs of time periods (before and after combinations, using different cut-off points) were tested on the basis of the 200 observations. The 200 observations were also divided up by region (Africa; Central and Eastern Europe; South, East and Southeast Asia; Latin America and the Caribbean; and West Asia) to examine any regional effect. Similar tests were done for each region. Thus, for each set of tests, and for each of the four dependent variables indicated above, there is a maximum of fifty-four comparisons of group means. Two salient questions were asked in these tests: How many of the fifty-four tests in each category were statistically significant? And what time-lag for investors’ reaction to the signing of BITs can be deduced from the pattern of statistically significant results? 2. Results a. The overall pattern of significant t-tests. The results of the test examining the association between the signing of a BIT and FDI flows, presented in table 1 can be summarized as follows: • The results are not strong,4 but are consistent enough over all four variables FDI, FDI/GDP, FDI/inflow, and FDI/outflow) to suggest that BITs have an effect on FDI. Of the four variables, FDI/inflow and FDI/ outflow registered the largest number of statistically significant results. This suggests that BITs may serve, at the margin, to redirect the share of FDI from/to BIT signatories. • When all countries are taken together, the strongest results are obtained in the FDI/inflow category, where thirty-seven of the fifty-four tests were significant (Table 1). The negative sign for the difference appears because, as was expected, the FDI share before a BIT was lower than the FDI share after it. The consistent negative sign in this category indicates a noticeable, if weak, effect. • The fact that FDI/inflow and FDI/outflow produced more significant results than FDI alone or the FDI/GDP ratio is gratifying, since the share measures are better measures of the role of BITs than the other two variables.

4. One could argue that significance between the 0.05 and 0.10 level is somewhat weak. However, its use is common for many similar studies. Moreover, t-test of differences in group means are a robust and unambiguous enough technique that a significance level of 0.10 or more is acceptable.

the impact on foreign direct investment of bit s 329

table 1. comparison of means for fdi/inflow variable Test number

Mean FDI share for period

Mean FDI share for period

Difference

Pb

1

2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 2 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 3 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT

BIT year to 1 year after BIT BIT year to 2 years after BIT BIT year to 3 years after BIT 1 year after BIT to 2 years 1 year after BIT to 3 years 1 year after BIT to 4 years 2 years after BIT to 3 years 2 years after BIT to 4 years 2 years after BIT to 5 years BIT year to 1 year after BIT BIT year to 2 years after BIT BIT year to 3 years after BIT 1 year after BIT to 2 years 1 year after BIT to 3 years 1 year after BIT to 4 years 2 years after BIT to 3 years 2 years after BIT to 4 years BIT year to 1 year after BIT BIT year to 2 years after BIT

–7720.2

.033∗

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

–10992.5

.029∗

–13483.3

.050∗

–15413.1

.041∗

–17639.4

.065∗

–12680.6

.054∗

–21609.4

.061∗

–14304.8

.057∗

–9152.2

.083∗

–1830.7

.073∗

–8101.6

.048∗

–10600.5

.080∗

–12345.4

.064∗

–14627.9

.093∗

–9761.5

.095∗

–18054.9

.083∗

–10928.4

.097∗

–8812.8

.041∗

–11814.9

.024∗

Continued

330 unctad table 1. comparison of means for fdi/inflow variable (cont'd...) Test number

Mean FDI share for period

Mean FDI share for period

Difference

Pb

20

4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 4 years before BIT to 1 year before BIT 2 years before BIT to BIT year 2 years before BIT to BIT year 2 years before BIT to BIT year 2 years before BIT to BIT year 3 years before BIT to BIT year 3 years before BIT to BIT year 4 years before BIT to BIT year 4 years before BIT to BIT year 4 years before BIT to BIT year 4 years before BIT to BIT year 4 years before BIT to BIT year

BIT year to 3 years after BIT 1 year after BIT to 2 years 1 year after BIT to 3 years 1 year after BIT to 4 years 2 years after BIT to 3 years 2 years after BIT to 4 years 2 years after BIT to 5 years 1 year after BIT to 3 years 2 years after BIT to 3 years 2 years after BIT to 4 years 2 years after BIT to 5 years 2 years after BIT to 3 years 2 years after BIT to 4 years 1 year after BIT to 2 years 1 year after BIT to 3 years 2 years after BIT to 3 years 2 years after BIT to 4 years 2 years after BIT to 5 years

–14205.5

.042∗

–16240.7

.033∗

–18380.3

.055∗

–13556.0

.049∗

–22335.6

.047∗

–15254.9

.049∗

–10250.8

.097∗

–49261.3

.095∗

–19323.1

.046∗

–13094.7

.042∗

–8684.9

.073∗

–17985.2

.055∗

–11819.9

.060∗

–49583.2

.094∗

–50635.2

.087∗

–21321.0

.033∗

–15152.2

.032∗

–10775.1

.065∗

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Notes: 37 out of 54 comparisons were significant. ∗Significant at the 0.10 level (one-tailed t-test).

the impact on foreign direct investment of bit s 331

• In terms of specific regions, BITs signed by African countries appear to have more effect than BITs in other regions. In particular, the share of FDI inflows from a particular home country is more likely to be affected by the conclusion of BITs when the host country is an African country. This suggests that BITs are relatively more significant in redirecting FDI flows when the host countries are least developed or when they are perceived by investors as environments with a higher risk. However, this is only a hypothesis and would need to be tested further. • Central and Eastern Europe is another region where BITs appear to have an influence, as measured by some variables. In particular, when FDI is used as the dependent variable, half of the comparisons are significant, especially those with a greater lagged response. This is discussed further below. • In the case of South, East and Southeast Asia, BITs may be instrumental in redirecting the share of FDI outflows from home countries. b. Time-lag effects. By examining for which time periods comparisons are significant, one may deduce the lag response between BIT signing (Year 0) and FDI response. The data for the FDI/inflow variable for all countries reveal no clear pattern in the various significant results (Table 1). On the other hand, the test on the FDI variable for all countries shows significant results only in Year +2. In the case of Africa, the FDI/GDP variable exhibits a similar pattern: comparisons involving the period starting in Year +2 are the main ones that show significance. However, in the case of the FDI/inflow variable, again for Africa, the significant entries are in Year +1 or Year 0. Finally, in Central and Eastern Europe, the pattern of significant entries is again found in comparisons starting in Year +2. Because the results are rather mixed, only a tentative conclusion can be drawn, namely, that the response lag after the signing of a BIT may be as little as zero, but is more likely to be two years. B. Stage Two: Cross-Sectional Analysis In this stage, a cross-sectional analysis of the determinants of FDI, including BITs as an explanatory variable, was performed for 133 host countries. As most BITs were signed in the 1990s, the analysis was based on total FDI flows and stocks for the year 1995, with data for explanatory variable going back three years (1993 to 1995). As mentioned earlier, the objective is two-fold. The first goal is to test anew the relationship between BITs and FDI flows based on a larger sample of countries and BITs than it was possible to assemble for stage one. To the extent that some of the regressions will test the relationship between the total number of BITs signed by host countries and their total stock of FDI; this will give a maximum possible coverage of the BITs tested, at the cost, however, of using

332 unctad

a less perfect dependent variable than bilateral FDI flows. The second objective is to introduce other explanatory variables pertaining to the characteristics of host countries. Moreover, in relating total FDI flows in 1995 to the BITs signed between 1993 and 1995, one can double-check the time-lag effects.5 1. Selection of explanatory variables The effect of BITs on FDI did not turn out to be very significant in the stage-one analysis. Even if it had been significant, BITs may have a weaker influence on FDI flows than other FDI determinates, hence the need to introduce other independent variables. The literature on FDI determinants divides them broadly into firm-level determinates (see, for example, Pygal, 1981) and country-specific determinants, (see, for example, Lunn, 1983). The literature on country-specific determinants, in turn, is divided into explanations of patterns of outward FDI from one home country (see, for example, Scaperlanda and Balough, 1983) and cross-sectional analyses comparing inward FDI flows from all sources across a sample of host countries. The analysis here focuses on the latter type of variables, using common host-country– specific variables, such as market size and growth, exchange rate changes, inflation, capital formation and country risk, to explain inflows of FDI to host countries. a. Market size. The size of the host-country market is one of the most frequently used variables in the literature. It was found to be significant as an explanatory variable in a number of empirical studies (Kravis and Lipsey, 1982; Scaperlanda and Balough, 1983; UNCTC, 1991; UNCTAD, 1993; and Kreinin, Plummer, and Abe, 1997). A large market permits firms, including foreign firms, to achieve scale and scope economies; simply, it can accommodate more firms and more investments than small markets. The notion that FDI and market size should correlate positively is not automatic. If a country with a large market restricts or excludes FDI, as was until recently the case with large markets of the former centrally planned economies, it would be futile to test the FDI/market size correlation. Second, the argument of market size applies only to FDI oriented toward the local market and not to extractive, or export-platform–motivated investments intended for markets other than the country in question. Third, operationalizing the market-size measure with GDP or equivalent measures (as most studies do) leaves the analysis open to all the limitations of GDP as a surrogate for market size. Nevertheless, despite such caveats, market size is considered one of the most prominent determinants of FDI inflows in the literature. b. Market growth. Apart from market size per se, change in market size, typically measured by change in GDP, is another explanatory variable used in several studies (Lunn, 1983; Julius, 1990; Kreinin, Plummer, and Abe, 1997). As growth is a magnet for firms, including MNEs, a high growth rate in a host country

5. The results for another year (such as 1994) would be roughly comparable as far as the overall conclusions are concerned.

the impact on foreign direct investment of bit s 333

tends to stimulate investment by both domestic and foreign producers. Thus, FDI flows are hypothesized to respond to economic growth. Conversely, a slowdown in economic growth (or a decrease in GDP) is expected to slow down or reduce investment, including FDI. It has to be noted that the GDP growth variable has not always produced consistent results. c. Change in the exchange rate. In pure theory, with exchange rates indexed to purchasing power parity, the exchange rate variable should have no effect on FDI-flow motivations. However, persistent deviations of actual exchange rates from purchasing power parity are well known. More pertinently, a sharp devaluation creates an opportunity for foreign investors to buy assets in the country cheaply. On this argument alone, a devaluation should be followed by a subsequent rise in FDI inflows into the country. However, the true picture is much more complicated, because local currency devaluation also affects the future expected profit stream of foreign investors, as measured in their own currencies, in ways that are by no means uniform. Export-oriented investments benefit from this situation (but only so long as the currency remains relatively undervalued). The profits of local-market-oriented investments, as measured in the foreign investor’s currency, may suffer for a temporary or prolonged period, until price increases can be passed on to local customers, but this depends on relative inflation, macroeconomic demand conditions accompanying the devaluation, and the price elasticity for the product in question. Hence a uniform hypothesis, applicable to all countries, is not advisable. Moreover, sharp devaluations are sporadic: there is therefore no a priori reason to expect a change in the exchange rate variable to show up as significant, especially in a cross-sectional study covering one year at a time. Nevertheless, since some studies (e.g., Froot and Stien, 1991) have found devaluations to be significant in explaining FDI flows, it is worthwhile to include this variable in this study. d. Inflation. Once again, assuming the theory of purchasing power parity is working—that is, that relative inflation is reflected in continuous adjustments in the exchange rate—the inflation variable ought not, in theory, to be a significant explanation for FDI flows. However, high inflation may be perceived as reflecting macroeconomic instability resulting from mismanagement; moreover, by heightening economic uncertainty, it can by itself be a deterrent to investors. For this reason, a negative relationship can be hypothesized between inflation and FDI. e. Rate of capital formation. The rate of domestic capital formation can be used to explain FDI flows (UNCTAD, 1993). There should be a complementary relationship between FDI inflows and the rate of domestic investment as a proportion of GDP. According to UNCTAD, “The hypothesis is that economies or regions that invest a high proportion of their GNP in plant and equipment are likely to be attractive markets for foreign investors seeking to increase their participation through the acquisition of existing firms or the establishment of green field operations” (UNCTAD, 1993, p. 10). Thus, this variable is an

334 unctad

extension of the arguments on market size and growth. On the other hand, one could also think that FDI projects might be a substitute for domestic investment projects on the grounds of their higher efficiency—a substitution more likely to take place in smaller economies. Therefore, a negative association between the two may be expected. f. County risk. Both a priori reasoning, as well as the results of previous studies (e.g., Green and Cunningham, 1975; Schneider and Frey, 1985), indicate that, all other things being equal, the higher the perceived risk associated with a country, the lower the FDI flows to that country. Risk scores for a country are based on political, economic, or financial criteria that rating agencies try to apply uniformly across countries. Instead of the word “risk” (which can be misleading), such scores are often better regarded as indexes of the quality of a country’s investment climate. The operationalization of the county-risk variable remains rather varied, depending on the agency that compiles the comparative data on the risk profile of countries. Examples of agencies that supply such ratings include the Economist Intelligence Unit, Frost and Sullivan, the PRS Group, and Moody’s. Typically, risk scores are calibrated on a 0 to 100 scale, although some agencies use a letter format (A, B, C, etc.). In general, the hypothesis in previous studies is that country risk is negatively associated with FDI. g. Other variables. A few additional variables have been used occasionally in other studies, such as comparative labor rates, human capital (skills), and infrastructure development. For the most part, these variables are more relevant to certain industries, or to particular types of investments such as export-oriented FDI, and not to a study, such as this one, that encompasses a large number of countries and all sectors combined. On infrastructure, which is coming under increasing scrutiny as a possible bottleneck to FDI, there is unfortunately no large-scale comparative index that can be applied across a large group of host countries. However, some risk-rating agencies attempt to include infrastructure in their assessment of a country’s “economic” risk score. h. Summary of hypotheses. For each of the explanatory variables discussed above, a positive or negative sign in the parenthesis indicates its hypothesized relationship to FDI (box 1). FDI is expected to be positively associated with BITs, market size, market growth, and exchange rate devaluation; it is expected to be negatively associated with inflation and political risk.6 The relationship between FDI and capital investment is hypothesized to be bidirectional.

6. However, note that some rating agencies (such as the PRS Group) invert the scale, using a minimum of 1 for the most “risky” nation and a maximum of 100 points for the least risky. In effect, the scale then is one describing a favorable investment climate.

the impact on foreign direct investment of bit s 335

Box 12.1 Summary of Variables and their Expected Impact on FDI FDI = f [BITs (+); market size (+); market growth (+); devaluation of exchange rate (+); inflation (−); capital formation (?); country risk (−)]

The relationship between FDI and these variables is hypothesized to be stronger with a lag of one to two years (following the results of UNCTC, 1991), although other studies propose longer lags. Thus, for example, the values of independent variables in 1994 would be expected to have the strongest relationship with FDI in 1995. 2. Definitions of variables a. Dependent variables. Data on FDI have been operationalized in terms of FDI flows and FDI stocks. In addition to the absolute value of FDI flows and stocks, both variables were normalized by population and GDP. This was done to correct for the effect of large countries and to see how BITs affect not just FDI but FDI per unit of GDP (or FDI per capita) across countries. Finally, the dependent variable is expressed in terms of its growth rate to see if the growth rate is affected by BITs or other variables. Each of the dependent variables (Table 2) is regressed (one at a time) against the explanatory variables (Table 3).

table 2. list of dependent variables Variable (and abbreviation) A. FDI flows into host country Absolute flow (Flow) Flow per capita (Fper) Flow per $1,000 GDP (Fgdp$) Flows growth (Fgrow) B. FDI stock in host country Absolute stock (Stock) Stock per capita (Sper) Stock per $1,000 GDP (Sgdp$) Stock growth (Sgrow)

Unit

Millions of dollars Dollars Dollars Change over previous year (percentage) Millions of dollars Dollars Dollars Change over previous year (percentage)

336 unctad table 3. list of independent variables Explanatory variables (FDI host country)

Description

BITs

Number of BITs signed in a particular year by a host country Total number of BITs signed up to and including that year BITs per million of population BITs per billion of GDP in dollars GDP in local currency/exchange rate, in millions of dollars Change over previous year, as a percentage GDP/population Population, in millions Value of gross fixed capital formation, in billions of dollars Amount of local currency per dollar, average annual rate Change over previous year, as a percentage. Exchange rates in local currency units per dollar Annual change in consumer prices, as a percentage 100 = best investment rating; 1 = worst 100 = best investment rating; 1 = worst 100 = best investment rating; 1 = worst 100 = best investment rating; 1 = worst

Cumulative BITs BITs per capita BITs as a ratio of GDP GDP in dollars GDP growth GDP per capita in dollars Population Capital investment Exchange rate Change in the exchange rate Inflation rate Political rating for country Economic rating Financial rating Composite country rating

b. Independent variables. The principal explanation for FDI that is being tested is its link with BITs. These treaties are measured in four different ways: the number of BITs signed by a host country in a particular year; the total number a host country has signed up to and including that year; BITs per million of population; and BITs per billion dollars of GDP. Tests will reveal which of these measures are best able to explain FDI flows from a statistical point of view. At the same time, other explanatory variables will be tested in case the BIT variables fail to provide an adequate explanation. These variables include population and various GDP measures such as indexes of market size and growth. Measures for the GDP variable that will be used, one at a time, include the absolute size of GDP, GDP growth, and GDP per capita. Other FDI determinants include capital investment measured by the value of gross fixed capital formation in a host country and the inflation rate measured by consumer price index. Finally, investors’ perceptions of a country’s investment environment

the impact on foreign direct investment of bit s 337

(or, in other words, the risk rating for FDI) are shown on a scale ranging from 1 (worst environment, or highest risk) to 100 (best environment, or lowest risk), broken down for economic, political, and financial ratings, and also presented as a composite score combining these three criteria. In the statistical analysis, these indicators of investment risk are not used simultaneously, but instead as alternatives. 3. Methodology a. Multivariate regression Given that the hypotheses postulate a relationship between a dependent variable and multiple independent variables across countries, multiple linear regression is considered an appropriate statistical technique to use. Each of the dependent variables (Table 2) is individually regressed against the independent variables (Table 3). For all independent variables loaded at once—using the stringent criterion that an entire row of data is eliminated if any entry for a country is missing— only seventeen countries are left in the data pool. This is too small for any meaningful statistical analysis. To address this problem, statistical software packages have “missing value substitution” procedures.7 These procedures may involve, for example, replacing missing data with the mean value for a variable. However, this kind of substitution lowers the confidence one can place in the results of a study. An often better alternative to the problem of missing data is to use stepwise multiple regression, where independent variables are loaded in steps, or one at a time, starting, for instance, with the independent variable that has the strongest explanatory significance, followed by the independent variable that has the second-strongest explanatory significance, and so on. The statistical routine then stops at a point where only a subset of the most powerful of the explanatory variables are entered, and the rest are left out. Stepwise multiple regression, in a situation such as the one in this analysis, has three virtues: 1) Because not all the independent variables are entered, the problem of the missing data is reduced, and more cases are included in the analysis. 2) The technique serves to identify the subset of explanatory variables that is best able to explain statistically the dependent variable. In this analysis, it can serve to identify which of the nine or more explanatory variables best explains FDI flows (Table 3). 3) Most important for the purposes of this study, a stepwise regression that loaded the strongest independent variable first would tell whether BITs—as opposed to other FDI determinants—are the better explanation. That is to say, if the BIT variable is loaded first, this would indicate that, of all the variables, the BIT variable is the

7. For this analysis, the Statistical Package for the Social Sciences was used.

338 unctad

most statistically important.8 On the other hand, if the BIT variable is selected for entry in later steps, that would diminish its importance in relation to other FDI determinants, and if the BIT variable is not selected at all, then BITs could not be said to play an explanatory role for that particular regression run. It should be emphasized that even without using the stepwise regression technique there are several independent variables that should not, a priori, be entered together. That is to say, some of the variables in Table 3 are redundant and may be used as alternatives to each other, but not together. As one example, country investment ratings are broken into several subcategories, such as “political,” “economic,” and “financial.” Scores were obtained for each country on each of these subvariables. However, both in theory and in practice, it is difficult to disentangle political, financial, and economic risks. Such distinctions may provide some value to investors who wish to track a country’s performance over time, but in a study involving over a hundred countries, they may not be meaningful. Moreover, the subcategories are often strongly correlated. Hence they may be used as alternatives to each other, one at a time. Additional variations to the regressions involved the removal of three countries shown to be egregious outliers (China, Singapore, and the Hong Kong Special Administrative Region of China)9 and log transformation of independent variables (in case some exhibited a non-normal distribution). Other variables, such as independent variables with mean value substitution, and regression runs on a regional basis raised the total number of regression runs to 264. b. Euclidean distance. In addition to the regression technique, n-dimensional Euclidean distance, or “pattern analysis,” was used to test for the relationship between FDI and the profile of countries in terms of these independent variables. This technique involves identifying a small top percentile of the countries in terms of FDI (or outcome variables) and determining the profile of this desired group in terms of the mean values of the independent variables. Once the desired profile is determined, for each of the remaining countries (the sample countries) a distance measure is computed that measures the distance of each of these countries from the ideal profile in a multidimensional space. If the distance from the ideal profile is greater, then it is hypothesized that FDI will be lower;

8. However, this is a tentative deduction, since interaction effects combined with other variables and multicollinearity may sometimes cause a variable to be loaded in an early step even though it is weaker than the rest. 9. Statistical practice often recommends removal of clear outliers in order to improve the statistics and thus the significance of remaining variables. This requires no a priori theory. However, ex post, one knows from similar FDI studies that China often stands out as a gross outlier. The Hong Kong Special Administrative Region of China and Singapore also stand out in some studies because of the relatively small size of their populations.

the impact on foreign direct investment of bit s 339

in other words, the correlation between distance and FDI is supposed to be negative and significant. Distance is computed as a squared Euclidean distance. However, given the differences in the scales being used to measure the several independent variables (for instance, one variable could be in billions of dollars, while another is a percentage, or another a ratio), another measure, the Mahalanobis distance (Box 2), is more appropriate. The Mahalanobis procedure first standardizes each variable by subtracting its mean and dividing it by its standard deviation, which reduces each variable to a comparable scale.

Box 12.2. Mahalanobis Distance n

Distance=

∑ (X k=1

sk

− Xik)/s•d•ik

Where, k = variable n = number of variables s = sample group, and i = ideal group s.d. = standard deviation and x– = mean

4. Results a. Multivariate regression. For this stage of the analysis, eight different dependent variables referring to FDI in 1995 (Table 2) were computed and regressed with independent variables for three years: 1995, 1994 and 1993 (Table 3). This permitted a total of 192 regression runs, using a variety of techniques such as log-transformed variables and the removal of outlier countries. The salient results covering the stepwise regression method for all countries and untransformed variables on a normal basis are shown in Table 4. As discussed above, stepwise regression with forward inclusion is a technique that helps to identify the subset of explanatory variables that have the strongest significance, that is, explanatory power vis-à-vis the dependent variable. The technique enters variables one at a time starting with the strongest until the adjusted R2 stops growing or until minimum loading criteria can no longer be fulfilled to justify the loading of the rest. The subset of variables thus entered comprises the resultant regression equation. Table 4 shows the results for only three of the

340 unctad table 4. regression results 1. FDI flow (95) = Adj. R2 = 0.99

–18.91 (–0.08) F = 651.87

2. FDI flow (95) = Adj. R2 = 0.99

–0.63 (–0.003) F = 651.87

3. FDI flow (95) = Adj. R2 = 0. 99

–493.58 (–1.58) F = 661.53

4. FDI Stock (95) = Adj. R2 = 0. 95

+0.02 GDP$ (95) (5.53)∗∗

–45.89 Capital (95) (–3.94)∗∗

–

+0.02 GDP$ (94) (5.31)∗∗

–56.32 Capital (94) (–3.78)∗∗

–

+0.022 GDP$ (93) (5. 97)∗∗

–52.53 Capital (93) (–3.88∗∗)

+162.74 Bit (2.32)∗

52544.74 (1.99) F = 92.55

+25.16 Population (95) (27.42)∗∗ P = 0.00 +27.18 Population (94) (28.87)∗∗ P = 0.00 +25.63 Population (93) (24.97)∗∗ P =0.00 +0.43 GDP$ (95) (10.91)∗∗ P = 0.00

–963.01 Capital (95) (–6.79)∗∗

–

5. FDI Stock (95) = Adj. R2 = 0. 95 6. FDI Stock (95) = Adj. R2 = 0.96

–4527.93 (–1.61) F = 99.87 –3362.03 (–1.38) F = 128.71

+0.50 GDP$ (94) (10. 32)∗∗ P = 0.00 +0.49 GDP$(93) (11.20)∗∗ P =0.00

–1230.34 Capital (94) (–7.17)∗∗ –1187.80 Capital (93) (–7.26)∗∗

7. Fgdp$ (95) = Adj. R2 = 0.55

57.69 (2.94)˛∗ F = 10.66

8. Fgdp$ (95) = Adj. R2 = 0.72

50.21 (2.89)∗∗ F = 14.68

9. Fgdp$ (95) = Adj. R2 = 0.75

44.74 (3.90)∗∗ F = 16.55

–0.02 Population (95) (3.46)∗∗ P = 0.002 +0.03 Population (94) (3.32)∗∗ P = 0.002 +1.86 Bit (93) (2.81)∗ P = 0.001

–1.16 Economic risk (95) (–2.15)∗ –1.10 Economic risk (94) (–2.44)∗∗ –0.94 Financial risk (93) (–3.21)∗∗

–868. 54 Political rating (95) (–2.20)∗ +49.98 Population (94) (4.61)∗∗ +34.90 Population (93) (3.26)∗∗ –

–

–

–

+1.40 Bit (94) – (2.37)∗

+0.02 Population (93) (2.95)∗

–

Notes: The results represent stepwise regression and the variables are ordered in the sequence of highest contribution. In each regression result, the second row with numbers in parenthesis represents “t” values for coefficients. Significance levels ∗∗are at better than 0.01; ∗ are at better than 0.05. Regressions were tried with each of the eight dependent variables for three years (1995, 1994, and 1993) and only these nine regressions had large enough R2 and P values. All coefficients of independent variables are significant at the 0.05 level. For key to abbreviations of variables, see Table 2.

the impact on foreign direct investment of bit s 341

eight dependent variables—FDI flow, FDI stock, and FDI/GDP ratio—because the results for the other five dependent variables were patchy. In Table 4, the order of loading of the independent variables is from left to right. The overall conclusion is that indexes of a host country’s market size, such as population and GDP, are the leading determinants of FDI. It was only in equation 9 in Table 4 that the BIT variable was loaded first. Overall, the BIT variable appears only twice in table 4. In general, these results suggest that BITs play only a secondary—and minor—role in cross-sectional analysis comparing a large number of countries with each other. Looking at individual dependent variables, FDI flow (equations 1, 2, and 3) is consistently a function of population, GDP measured in dollars, and capital investment. When regressed with lagged variables, the BIT variable for 1993 partially explains FDI flow with a two-year lag. FDI stock (equations 4, 5, and 6) is also consistently a function of population, GDP measured in dollars, and capital investment. Political risk in a country also seems to explain FDI stock in some contexts. Results for Fgdp$ (FDI flow per $1,000 of GDP of the recipient country, in equations 7, 8 and 9) are not as strong. The equations’ adjusted R2 is lower, albeit highly significant. However in equation 9, BIT (93) is highly significant and by itself explains 50% of the variation of Fgdp$, a higher score than the population independent variable. Population and GDP, two indicators of market size, are positively and significantly associated with FDI flow and stock, as hypothesized. Capital investment is consistently, negatively, and significantly associated with FDI flow and stock. Two alternative hypotheses regarding this variable have been put forward, as the literature is silent or ambivalent on this issue. One hypothesis (UNCTAD, 1993, p. 10) suggests that a high rate of capital formation in a country—defined as gross fixed capital formation in billions of dollars for a particular year (table 3)—taken as an indicator of economic activity and investment, could attract FDI to a country. That is to say, FDI and capital formation are complementary and positively associated. The alternative hypothesis is that FDI is a substitute for domestic capital formation, which suggests a negative association between the variables. The findings lend some credence to the latter hypothesis. Political risk, while significant, is negatively correlated. Given the fact that the political risk factor was in increasing order of favorable conditions (a score of 1 being high risk and a score of 100 being low risk), the negative result does not support the hypothesis. The BIT variable is significant in two of the nine regression equations, indicating than an increase in BITs is associated with an increase in FDI flows, but with a lag of two years. As regards the magnitude of its effect on the dependent variable in question, in equation 3 (which is for many countries grouped together), each BIT in 1993 can be said to be associated with an incremental $162 million in FDI flows in 1995. However, this is a statistical abstraction,

342 unctad

in the sense of a fitted regression trend line for many countries, and not a policy conclusion. In the same vein, each BIT can be said to be associated with increasing the ratio of FDI to $1,000 of GDP of the host country by a factor of three. However, to reiterate the larger picture, BITs had a discernible effect in only two of the nine equations. In the other seven equations, the BIT variable failed to be loaded for lack of statistical significance—that is to say, BITs remain a very minor consideration overall. Foreign direct investment, and especially FDI flows, can fluctuate from year to year, and thus distort calculations based on flow data for one year. To make up for this, an additional twenty-seven regressions were made with dependent variables going back to 1993 and independent variables going back to 1991 (Table 5). The only meaningful dependent variables turned out to be FDI flow, FDI stock, and FDI/GDP. Other relative dependent variables did not produce meaningful relationships. As regards independent variables, the pattern of relationships that was revealed earlier for the dependent variables for 1995 was confirmed by these additional regressions for 1993 and 1994. GDP, population, and capital were the most dominant variables. A weak relationship was also revealed between political, economic, and financial risk and the BIT variable. As the extension of the dependent variables to two additional years has not made a difference for the results, further analysis was based, again, on the 1995 data for dependent variables and 1993–1995 data for independent variables. The data were divided into five regions: Africa; Central and Eastern Europe; East and Southeast Asia; Latin America and the Caribbean; and West Asia. Additional regression runs were performed on each region separately to see if there were any regional variations regarding the significance of BITs. BIT-related independent variables were frequently included in runs involving only one dependent variable, Flow (95), for all regions except West Asia. Moreover, because the division of data into regions resulted in incomplete data, a mean value substitution procedure had to be followed (since without one, no statistically significant results would have been obtained). However, not too much reliance can be placed in these results. The overall conclusion from Tables 4 and 5 (and annex tables 10 and 11) is that BITs play a minor and secondary role in a cross-country comparison of FDI determinants. In keeping with other studies (such as Scaperlanda and Balough, 1983; UNCTC, 1991; Kreinin, Plummer, and Abe, 1997), market size appears to be the leading determinant of FDI flows. b. N-dimensional pattern analysis. Pattern analysis measures the n-dimensional Euclidean distance for standardized variables, to a desired or ideal point. In this analysis, it measures the “distance” between a country and a desired subset of countries that exhibits high FDI flows. From this analysis one can deduce whether BITs are associated with high FDI flows.

table 5. additional regression results –9310.21 (–3.51)∗∗ F = 195.86

+30.12 Population (92) (19.79)∗∗ P = 0.000

+158.73 Political risk (92) (3.88)∗∗

–

–

2. FDI Flow (95) = Adj. R2 = 0.947

246.82 (0.680) F = 240.12

+27.75 Population (91) (17.207)∗∗ P = 0.000

+0.006813 GDP$ (91) (2.436)∗

–

–

3. FDI Flow (94) = Adj. R2 = 0.631

–1380.06 (–1.824)∗ F = 36.09

+14.95 Population (94) (6.459)∗∗ P = 0.000

+111.15 Bit (94) (3.251)∗∗

–

–

4. FDI Flow (94) = Adj. R2 = 0.720

–646.73 (–0.94) F = 39.54

+89.34 Capital (93) (4.99)∗∗ P = 0.000

+9.28 Population (93) (3.64)∗∗

–

–

5. FDI Flow (94) = Adj. R2 = 0.932

–621.40 (–1.38) F = 119.04

+25.84 Population (92) (12.99)∗∗ P = 0.000

+0.16 GDP per capita (92) (2.30)∗

+0.00658 GDP$ (92) (2.07)∗

–

Continued

the impact on foreign direct investment of bit s 343

1. FDI Flow (95) = Adj. R2 = 0.937

344 unctad

table 5. additional regression results (cont'd...) 6. FDI Flow (94) = Adj. R2 = 0.940

–280.25 (–0.76) F = 211.845

+24.74 Population (91) (15.12)∗∗ P = 0.000

+0.01099 GDP$ (91) (3.87)∗∗

–

–

7. FDI Flow (93) = Adj. R2 = 0.773

–620.08 (–1.19) F = 52.16

+0.02084 GDP$ (93) (5.80)∗∗ P = 0.000

+8.20 Population (93) (4.45)∗∗

–

–

8. FDI Flow (93) = Adj. R2 = 0.934

–6845.03 (–2.92)∗∗ F = 92.49

+22.14 Population (92) (12.50)∗∗ P = 0.000

+112.26 Political risk (92) (2.99)∗∗

+0.02192 GDP$ (92) (3.11)∗∗

–64.99 Capital (92) (–2.32)∗

9. FDI Flow (93) = Adj. R2 = 0.950

–3460.99 (–2.39)∗ F = 129.166

+21.26 Population (91) (16.10)∗∗ P = 0.000

+0.03181 GDP$ (91) (4.66)∗∗

–87.02 Capital (91) (–3.41)∗∗

+56.10 Composite risk (91) (2.32)∗

586.14 (0.16) F = 45.23

+0.16 GDP$ (92) (6.23)∗∗ P = 0.000

+41.38 Population (92) (2.49)∗

–

–

10. FDI Stock (95) = Adj. R2 = 0.773

–1715.11 (–0.51 ) F = 56.30

+0.18 GDP$ (91) (6.90)∗∗ P = 0.000

+48.39 Population (91) (3.20)∗∗

–

–

12. FDI Stock (94) = Adj. R2 = 0.537

–848.82 (–0.27) F = 48.54

+0.12 GDP$ (94) (6.97)∗∗ P = 0.000

–

–

–

13. FDI Stock (94) = Adj. R2 = 0.794

–43527.4 (–2.91)∗∗ F = 39.64

+0. 37 GDP$ (93) (6.18)∗∗ P = 0.000

–845 .17 Capital (93) (–3.83)∗∗

+1244. 82 Economic risk (93) 2.84)∗∗

–

14. FDI Stock (94) = Adj. R2 = 0.729

–12.27 (–0.00) F = 70.79

+0.17 GDP$ (92) (8.41)∗∗ P = 0.000

–

–

–

15. FDI Stock (94) = Adj. R2 = 0.747

–2249.36 (–0.69) F = 80.64

+0.19 GDP $(91) (8. 98)∗∗ P = 0.000

–

–

–

Continued

the impact on foreign direct investment of bit s 345

11. FDI Stock (95) = Adj. R2 = 0.804

346 unctad

table 5. additional regression results (cont'd...) 16. FDI Stock (93) = Adj. R2 = 0.799

–31993.9 (–2. 45)∗ F = 30. 85

+0. 37 GDP$ (93) (7.36)∗∗ P = 0.000

–909. 82 Capital (93) (–4. 97)∗∗

+1067.66 Economic risk (93) (2. 89)∗∗

–1659.86 Bit (93) –2.30)∗

17. FDI Stock (93) = Adj. R2 = 0.671

598.95 (0.18) F = 54.07

+0.14 GDP$ (92) (7.35)∗∗ P = 0.000

–

–

–

18. FDI Stock (93) = Adj. R2 = 0. 711

–1624.61 (–0.55) F = 67.41

+0.16 GDP $(91) (8.21)∗∗ P = 0.000

–

–

–

Notes: The results represent stepwise regression and the variables are ordered in the sequence of highest contribution. In each regression result, the second row with numbers in parenthesis represents “t” values for coefficients. Significance levels ∗∗are at better than 0.01; ∗ are at better than 0.05. Regressions were tried with each of the eight dependent variables for three years (1995, 1994, and 1993) and only these nine regressions had large enough R2 and P values. All coefficients of independent variables are significant at the 0.05 level. For key to abbreviations of variables, see Table 2.

the impact on foreign direct investment of bit s 347

table 6. pattern analysis correlation Variable

Distance measure with 1994 variables

Distance measure with 1995 variables

Flow (95) Stock (95) Flow/GDP$ (95) Fgdp$ (95) Growth (95)

0.39∗ 0.48∗ –0.14 –0.28 –0.05

0.58∗∗ 0.46∗ –0.38∗ –0.42∗ 0.38∗

Notes: ∗Correlation significant at the 0.05 level (one-tailed).

The pattern analysis results do not consistently support the hypothesis that a deviation from an ideal BIT profile will necessarily result in lower FDI levels. The results are not only mixed in direction, but also not very strong. This could be because some of the variables are correlated among themselves.

conclusions The time-series data analysis based on bilateral FDI flows between the BIT signatory countries shows that the influence of BITs on FDI is weak, especially in redirecting the share of FDI flowing from or to BIT signatory countries. In other words, following the signing of a BIT, it is more likely than not that the host country will marginally increase its share in the outward FDI of the home country; the same applies to the share of the home country in the FDI inflows of the host country. The effect, however, is usually small. On the question of the time lag—that is, how long it takes for the effect, if any, of signing a BIT to materialize—the analysis can only supply a very tentative conclusion, namely, that the response of foreign investors may be immediate but is more likely to occur in the three years following the signature of the BIT. In the cross-country comparison of FDI determinants, the overall conclusion is that BITs appear to play a minor and secondary role in influencing FDI flows.10

10. To quote a paper prepared by the Government of Germany, “BITs definitely are not the condition sine qua non for foreign investors’ decision” (Germany, 1997, p. 1). The same paper observes: “As many investors, however, postpone their investments until their establishment is protected by a BIT, the business community seems to be aware of additional benefits through these agreements” (p. 3). Similarly, according to an expert from China, “We are often consulted by foreign investors and our own overseas investors on BITs,

348 unctad

Other determinants of FDI flows, especially the size of a host country’s market, are more important; this finding supports the results of several previous studies. Moreover, since some two-thirds of BITs have been concluded in the 1990s, the distinctive influence of a BIT as a competitive signal to attract investment may have been eroded. Rather, BITs are increasingly regarded by foreign investors as a normal feature of the institutional structure introduced in the past decade.

especially when a large amount of investment and investment in some sensitive sectors such as natural resources, public utilities, are to be made” (China, 1997, pp. 5–6).

13. do bilateral investment treaties attract fdi? only a bit . . . and they could bite∗ mary hallward-driemeier “Even some of NAFTA’s strongest supporters say that clever and creative lawyers in all three countries are rapidly expanding the anti-expropriation clause in unanticipated ways.” —Business Week: April 1, 2002. “The Highest Court You’ve Never Heard Of” A Canadian trade lawyer gave the following assessment to Parliament regarding NAFTA’s Chapter 11: “They could be putting liquid plutonium in children’s food. If you ban it and the company making it is an American company, you have to pay compensation.” —Bill Moyers in “Trading Democracy”, PBS, Feb. 5, 2002. “Essentially, we’ve now seen a shift of the use of investment agreements as a shield to using them as a sword against government activity.” —Howard Mann, a lawyer with the International Institute for Sustainable Development, interview with Bill Moyers on “Trading Democracy” for PBS, Feb. 5, 2002. “NAFTA was not intended to provide foreign investors with blanket protection from this kind of disappointment, and nothing in its terms so provides.” —Robert Azinian, Kenneth Davitian and Ellen Baca v. The United Mexican States, Award, November 1, 1999, para. 83. “In these early days of NAFTA arbitration the scope and meaning of the various provisions of Chapter 11 is a matter both of uncertainty and of legitimate public interest.” Mondev International Lt. v. United States of America, — Award, October 11, 2002, para. 159.

∗

This chapter is an update of World Bank Policy Research Working Paper 3121. The author wishes to thank Richard Newfarmer, Pierre Sauve, Beata Smarzynska, the participants at the Columbia Law School Conference on Regulating FDI, George Washington University, and participants in the World Bank’s Economist Forum for their comments. The author is grateful to Sweta Bagai for research assistance. The views expressed here are those of the author and do not necessarily reflect those of the World Bank, its Executive Directors, or its member countries.

350 mary hallward-driemeier

introduction As foreign direct investment (FDI) has surged dramatically over the last two decades, more developing countries are competing to host these multinationals. In addition to negotiating firm-specific deals through tax incentives, subsidies, etc., countries have increasingly turned to signing bilateral investment treaties (BITs) as a way to entice foreign investors to their shores. Recent years have witnessed an explosion of such treaties. BITs are heralded by their proponents as an important means of attracting new FDI. Yet there has been little examination of whether these instruments actually affect the allocation of foreign investment. There has also been remarkably little attention paid to the implications of the strength of the rights bestowed to the investor and obligations assumed by the host country. Recent claims brought under such treaties are only now bringing to light the potential magnitude of the obligations assumed by the host countries.1 The potential prospect of large stake litigation makes it all the more important to assess the benefits of entering into such agreements. This chapter provides an empirical investigation of whether the benefits are being realized, whether a BIT can substitute for weak domestic property rights, and whether ratifying it results in a significant increase in FDI. A BIT could help attract investment by serving as a commitment device. It is hypothesized that countries with weak domestic property rights can increase their attractiveness as potential hosts by explicitly committing themselves to honoring the property rights of foreign investors. In particular, a BIT could be a commitment device to overcome dynamic inconsistency problems. Hosts would have an incentive to make those promises necessary to bring investors in, but once the sunk costs are made, the host then has the incentive to deliver only to the level that will keep the investor from leaving.2 The presence of the BIT, with its dispute resolution mechanisms and provisions for compensation in the case

1. In CME Ltd. v. Czech Republic, an award of $350 million was handed down; an amount that will stand as the Czech Republic’s appeal of the award was rejected by the Swedish Court of Appeal in May 2003. A claim for $450,000 in the case of The Loewen Group v. The United States of America was just dismissed on jurisdictional grounds after the Loewen Group was acquired by a U.S. interest after bankruptcy proceedings—and after more than four years in the arbitration process and a long, public debate on the merits of the case. Another high profile case arising under NAFTA is still pending, with claimants seeking $950 million in the case of Methanex v. The United States. Of course, even if the tribunals find in favor of the claimants, the size of the award will not necessarily be at the level the claimants seek, but clearly the sums involved are substantial. Non-fiduciary costs can also be substantial; for example, if certain proposals for reform are abandoned for fear of legal action. For more information on recent high profile cases, please see the appendix. 2. With the proliferation of BITs, another motivation for signing the treaty is the fear they the potential host will not be competitive as a location if they do not also offer similar protections.

do bilateral investment treaties attract fdi? 351

of expropriation, guards against host country actions that would adversely impact the profitability of the investment. The importance of property rights, and the quality of domestic institutions more broadly, have been recognized in studies on growth and investment.3 Investors care about the likelihood that they will be able to earn—and control— a return on their investments. The existing studies have tested for the effect of property rights using differences across countries at a given period in time. The measurement of the quality of property rights (or institutions) is based on qualitative assessments and does not vary too much over time. Turning the focus to BITs has some advantages over these earlier approaches. First, the effect of ratifying of a BIT provides a more specific test of the importance of property rights per se. Second, it also relies on changes over time rather than variations in the cross-section. Using time-series variation regarding a distinct change in the property rights of a group of investors provides a more direct test of whether this significantly affects investment. While it should be recognized that a BIT could be an important commitment device, the nature of the commitment can vary enormously depending on the terms of the BIT. Too much attention has been placed on whether or not a BIT exists than on the strength of the property rights actually being enshrined in these agreements. To date there is no discussion in the economic literature of whether the strength of the rights enshrined in a BIT would provide adverse incentives to potential investors, or provide insurance well beyond what domestic investors enjoy or that foreign investors would require—with consequences that could potentially have enormous impact on the feasible policy choices available to host governments. Such concerns have begun to be debated within legal circles,4 largely stemming from recent arbitration decisions and new cases of how rights in BITs are being exercised against the U.S. and Canada.5 This chapter

3. See Kaufmann, Kraay, and Mastruzzi (2006); Acemoglu, Johnson, and Robinson (2001); Stein and Daude (2001), Dollar and Kraay (2002); Rodrik, Subramanian, and Trebbi (2002); and Hallward-Driemeier (2002). 4. The issue is gaining some attention among legal scholars, but with the focus on the U.S. and Canada. For example, NAFTA’s regulatory takings is analyzed relative to the property rights protected in the Fifth Amendment of the U.S. Constitution, see Been and Beauvais (2003). 5. The most high profile examples involve disputes between the signatories of NAFTA. While NAFTA is not strictly a bilateral treaty, its Chapter 11 has language common to many BITs and highlights a number of relevant issues that apply more broadly to BITs’ signatories. Some of the cases under consideration demonstrate some of the unintended consequences of language commonly found in BITs that raises the distinct possibility that BITs can constrain policy choices on a broad set of issues from health to the environment, and open governments to substantial liabilities. For a brief description of some of the recent cases, please see the appendix. It should also be noted that some of the current cases that are grabbing media attention (e.g., Methanex’s suit against the U.S. for $970

352 mary hallward-driemeier

uses these cases to help motivate the issue more broadly, and takes the perspective of developing countries that represent the vast majority of host signatories of BITs. A. What is a BIT? BITs vary across countries, but they generally share similar features of defining foreign investment and laying out various principles regarding treatment, transfer of funds, expropriation, and mechanisms for dispute settlements. Because the central piece of a BIT is the assurance it gives investors regarding their property rights, it is important to look more closely at what these rights are. An examination of the language and growing legal caseload shows that not only do foreign investors secure additional property rights, but also that they could enjoy rights more substantial than many had anticipated. One common clause included in many BITs gives the investor the right to sue the host country if actions undertaken by the government are deemed to substantially expropriate the business of the firm. Two points should be highlighted: First, this right of an individual investor to sue the government is in itself an expansion of investor rights. In most cases, the government can claim sovereign immunity, leaving little recourse in the legal system. The remaining alternative is to seek the assistance of the investors’ own home country in gaining diplomatic protection. This may not be granted and makes the entire process a political one. Instead, with the investment treaty, the host country consents to a standing offer to arbitrate disputes covered by the treaty. Second, BITs outline the terms under which expropriation could be deemed lawful and compensation would be due. The exact wording of such clauses varies by signatory countries, but there is broad agreement on the thrust of the terms. Property can only be legally expropriated if it is for a public purpose, is done in a non-discriminatory way, compensation is paid, and the expropriation is done in accordance with due process of law. Of these conditions, the one with the largest consequences is the compensation clause. That there is some requirement for compensation is not controversial. What can be contentious are the terms of the compensation. Standards include “prompt, adequate and effective” or “payment of full value” or “just compensation.” This has been interpreted to mean the market value of the investment immediately prior to the expropriation being made public. Some statements are explicit (e.g., “the purpose of which shall be to place the investor in the same financial position as that in which the investor would have been if the expropriation or nationalization had not taken place.”

million due to California’s ban of MTBE) have not been settled. It is possible that as more cases are decided the prospect of expansive regulatory takings claims will not be upheld. Nonetheless, that such a case is in arbitration indicates that large suits that could limit feasible policy choices are at least a distinct possibility.

do bilateral investment treaties attract fdi? 353

China-Sweden BIT) while others leave the terms rather vague, creating challenges for courts and policy makers as they try to assess the impact of the BIT. The nationalizations that peaked in the 1970s provided many clear-cut cases of expropriation. Of greater concern more recently are “indirect expropriations,” “creeping” expropriation, or “regulatory takings” and whether they amount to a taking requiring compensation. These newer provisions on expropriation typically apply to actions by a country that “substantially impair the value of an investment.” There is no requirement that it be an isolated event or even that the country try to take ownership of the investment. Many BITs expressly state that expropriations include measures “tantamount” or “equivalent” to expropriation, or actions that would substantially impair the value of the investment.6 Rather than bringing the case in local courts (the quality and speed of which the foreign investors may not like) or seeking diplomatic protection, BITs usually specify dispute resolution mechanisms. One of the more popular options is to submit to binding arbitration through the International Centre for Settlement of Investment Disputes (ICSID), an affiliate agency of the World Bank. Two others are the International Chamber of Commerce and the United Nations Commission on International Trade Law (UNICTRAL). In these arbitration proceedings, three arbiters are selected—generally with each party selecting one and the forum selecting the third. These proceedings are not bound by precedents, are not necessarily obliged to be open to the public,7 and are not required to publish final decisions. The decisions have only limited avenues for appeal and cannot be amended by the domestic legal system or by the Supreme Court. The nature of the dispute resolution procedures can provide a great deal of leeway in how cases will be decided—with critics pointing out the danger that they could encourage investors to pursue their cases even if their merits are not all that strong. While expropriation cases have arisen from BITs over time, the caseload has been relatively small. In the last few years the numbers have jumped substantially. Having settled about 60 cases in four decades, ICSID now has over 40 cases currently pending. The increase in cases is partly a function of the increased number of BITs, and may also be a function of the publicity generated by cases brought under the North American Free Trade Agreement (NAFTA) Chapter 11.

6. See, e.g., BIT between Japan and Egypt, Article V: “expropriation, nationalization, restriction or any other measures, the effects of which would be tantamount to expropriation, nationalization or restriction.” France and Pakistan, Article 5: “measures of expropriation, nationalization or any other measures the effect of which would be direct or indirect dispossession” of an investment. See UNCTAD (1998), Chapter III for more detailed discussion of the provisions included in BITs. 7. Some countries do make documents available to the public. For example, the United States’ Freedom of Information Act mandates that documents be made available, however, this is not necessarily so for all countries.

354 mary hallward-driemeier

Critics worry that multinational corporations (MNCs) will use the provisions on regulatory takings and compensation as insurance against many risks the firms would otherwise have assumed themselves as part of the normal process of establishing and running a business. The terms of the treaty can be seen as essentially giving MNCs a property right by ensuring that those regulations that affect the MNC’s profitability will not change—and, if that gamble turns out to be wrong, that the MNC could be entitled to earn those profits anyway.8 How broadly the regulatory takings provision will be applied is still not determined, but the language of the treaty still offers greater property protection than is enjoyed by domestic investors.9 As the potential for legal recourse under BITs becomes more widely known, the importance of BITs in selecting a location for investment may become more important, and could lead to problems of moral hazard and adverse selection. If investors believe there is a chance for successful litigation against the host country, and that they are then protected from substantial amounts of risk, companies may not work as hard to make their firms successful, or may be attracted to locations where their legal cases could be made most strongly rather than moving for economic reasons. Those firms most likely to enter into business could be those most keen to pursue all legal recourses should the opportunity arise. Such cases may be rare, but the size of the claims in existing cases is large enough that negotiators should be careful in defining the terms surrounding expropriation and compensation clauses in future BITs, and in such agreements as the proposed expanded Free Trade Area of the Americas. The Azinian case provides an interesting example. On the one hand, the decision explicitly warns against the treaty being seen as a recourse against any poor outcome. A foreign investor entitled in principle to protection under NAFTA may enter into contractual relations with a public authority and may suffer a breach by that authority, and still not be in a position to state a claim under NAFTA. It is a fact of life everywhere that individuals may be disappointed in their dealings with public 8. In addition to the size of the awards and the constraints placed on policymakers, some American critics are concerned that Chapter Eleven is causing an “end run” around the Constitution and is decidedly anti-democratic. In support of their assertion, they argue that the terms and consequences of Chapter Eleven were never publicized or debated prior to signing; that there is no room for public comment or even public scrutiny of the arbitration procedures; and that there are limited mechanisms for appeal. Bill Moyers ran a special on PBS entitled “Trading Democracy” (Feb. 5, 2002), calling Chapter 11 the “Trojan horse of NAFTA” and the system of secret tribunals “a private court for capital.” A similar theme was sounded by Business Week in “The Highest Court You’ve Never Heard Of” (Business Week: April 1, 2002). It argued that decisions with widespread impact are and will be made by arbitration panels behind closed doors with no public accountability or recourse to the court system. 9. Been and Beauvais (2003).

do bilateral investment treaties attract fdi? 355

authorities, and disappointed yet again when national courts reject their complaints. NAFTA was not intended to provide foreign investors with blanket protection from this kind of disappointment, and nothing in its terms offers such security.10 On the other hand, considering the details of the case (some claims are dismissed as “preposterous”),11 the fact that the claimants even brought the case illustrates that they felt the treaty did give them a real possibility for relief. It should be noted that the rights secured in a BIT are reciprocal; investors from country A investing in country B are the same as those given to investors from country B investing in country A. Nonetheless, in practice there is usually tremendous asymmetry as almost all the FDI flows covered by BITs are, in fact, in one direction.12 It is precisely in those cases in which FDI flows in substantial amounts in both directions that countries have balked at ratifying BITs. It is striking that there is a dearth of such agreements between rich Organisation for Economic Co-operation and Development (OECD) countries. Rich OECD countries do participate in BITs, but almost exclusively with developing countries. It could be that in agreements between rich OECD countries there is no perceived need for a BIT to stimulate investment, as it is already substantial. Or, while OECD governments are keen to secure such rights for their companies overseas, they balk at granting such rights to MNCs within their own borders. B. Trends in BITs Since the first BIT was ratified in 1959, the number of BITs increased steadily through the 1980s. In the 1990s, the number boomed. In 1990 there were 470 treaties, by 2000 there were close to 2000 BITs (see Figure 1). Almost all the earlier treaties were ratified between rich OECD countries and developing countries (see Figures 2 and 3). With the fall of the Berlin Wall the former-Soviet republics entered the global market, and many East European countries ratified treaties with both OECD countries and with developing countries. The most recent rise in the number of BITs is attributable to agreements signed between developing countries. By 2000, half of all FDI flows from the OECD to developing countries were covered by a BIT. What is being tested in this chapter is whether this increase is simply due to the increased country coverage, or whether FDI flows are diverted to destinations covered by investment treaties. Clearly, a BIT is not a necessary condition to receive FDI. There are many source-host pairs with substantial FDI that do not have a BIT. Japan, the second largest source of FDI, has only 10. Azinian and others v. The United Mexican States, Award, November 1, 1999, para. 83. 11. Ibid., p. 7. 12. There are at least two cases, of the 120 filed before ICSID, in which the plaintiff is a developing country and the defendant is a developed country.

356 mary hallward-driemeier figure 1. surge in the number of bit s 2000 1500 1000 500 0 1950s

1960s

1970s

1980s

1990s

concluded four BITs. The U.S. does not have a BIT with China, its largest developing country destination. Brazil, one of the top receivers of FDI, has not ratified a single BIT. In addition, there are also numerous examples of countries that have concluded many BITs and yet have received only moderate inflows. SubSaharan Africa, for instance, has had difficulties in attracting FDI, though it has tried to improve the environment for FDI by entering into various agreements to protect the interests of investors. There are also examples such as Cuba, which does not have a BIT with either Canada or Mexico, its two biggest foreign investors. On the contrary, almost 60% of the countries it does have a BIT with actually have no foreign investment in Cuba.13 C. Other studies There is a growing literature on the importance of institutions and property rights. Most has been focused on the effects on long run growth rather than on FDI.14 Daude and Stein (1995) do look at the effect of institutions on FDI in a

figure 2. bit s concluded by developing countries 500 400

With Developed Countries

300

Between Developing Countries

200

With Transition Europe

100 0 1960s

1970s

1980s

1990s

13. See Perez-Lopez et al. (2001). 14. See Knack and Keefer (1995); Acemoglu, Johnson and Robinson (2001); Dollar and Kraay (2002); Rodrik, Subramanian, and Trebbi (2002).

do bilateral investment treaties attract fdi? 357

figure 3. top 25 countries in terms of the number of bit s concluded, 1 january 2000 140 124 120 95 92 92

100

94

90 84

77

80

72 63 60

62 60 47

62

59 53 52

53 52 52 45 45 44

43 43

40 20

lic ub

ze

ch

R

ep

ar

ia

ia an

Bu

lg

a si ni

om R

ub

a Tu

C

Ar

ge

nt

in

a

lic

t ep

ub

yp R

Eg

a, re

C

ux /L m Be

lg

iu

Ko

g

SA U

em

bo

ur

ly Ita

K U

G

er

m

an

y

0

DEVELOPED

DEVELOPING

EAST EUROPEAN/ TRANSITION

cross-section of both developed and developing countries, and find that institutions make a significant contribution to attracting FDI. Using panel data, Hallward-Driemeier (2002) looks at the effect of institutions on the allocation of FDI among developing countries and finds a weaker effect. These studies use broad measures of property rights, employing either International Country Risk Guide (ICRG) rankings or the Kaufmann, Kraay, Zoido-Lobaton (KKZ) indicator. The advantage of this study is that it looks at clear cases where property rights are explicitly strengthened to determine their importance. There are a couple of papers that have looked at other bilateral arrangements and their implications for FDI. Blonigen and Davies (2000) look at the role of tax treaties. Here there is a larger literature. They find that, contrary to expectations,

tax treaties can discourage FDI, and argue that the treaties can be used as devices to reduce tax evasion and not just as tools to simplify tax filings and avoid double

358 mary hallward-driemeier

taxation. Yeyati, Stein, and Daude (2002) look at the role of regional integration and the location of FDI to test whether greater access to larger markets attracts FDI. While they are almost exclusively looking at intra-OECD FDI flows, they find an important correlation between trade agreements and FDI. The role of BITs has received some discussion in law journals. There the focus has again been on the issue of providing a commitment device to overcome the dynamic inconsistency problem, or on the strategic concerns potential signatories face as other countries also consider signing such agreements.15 The question of whether the treaties actually do affect investment is not addressed. Within the economic literature, BITs have generated very little attention. UNCTAD (1998) sponsored one of the few BIT analyses. It studied the impact of 200 BITs on bilateral FDI data, examining years prior to and after their conclusion. It found a weak correlation between the signing of BITs and changes in FDI flows, but used minimal control variables in generating this result and did not control for the strong upward trend in FDI over time. Their cross-section analysis of 133 host countries in 1995 concluded that BITs do not play a primary role in increasing FDI, and that a larger number of BITs ratified by a host country would not necessarily bring higher inflows. While this cross-sectional result is interesting, the more rigorous test of the effect of BITs on FDI is to examine the impact of an investment treaty over time. This study looks at a panel dataset of bilateral FDI flows, augments the control variables included, and addresses a number of econometric issues not examined in UNCTAD’s earlier work. Since the working paper version of this chapter came out, two additional studies have been conducted. Tobin and Rose-Ackerman (2004) look at the total number of BITs signed and the total number signed with a high-income country. Looking at five-year averages between 1980–2000, they only find a significant impact for BITs under a random effects model. The random effects model is rejected by the Hausman test, but with fixed effects, the results are statistically insignificant. Neumayer and Spess (2004) use a longer panel, from 1972–2002. They also look at the total number of treaties signed and use aggregate FDI flows rather than bilateral flows. With less demanding requirements, they are able to include more developing countries in their study. They do find some evidence that countries that sign more BITs overall get higher aggregate flows of FDI—and that the effect is higher for countries with weaker institutions. This chapter provides a stronger test than the ones above by looking at the bilateral flows based on the bilateral treaties signed. Using the cumulative number of treaties signed makes it difficult to sort out the effects of the treaties 15. For dynamic inconsistency, see Vandevelde (1998), “Investment liberalization and economic development: the role of bilateral investment treaties,” Columbia Journal of Transnational Law; for strategic concerns, see Guzman (1998), “Why LDC’s sign treaties that hurt them; explaining the popularity of bilateral investment treaties,” Virginia Journal of International Law, 38, pp. 639–688.

do bilateral investment treaties attract fdi? 359

from the strong upward trend in FDI during the time periods covered, leading to upward biases in the effect of treaties. And the effect on aggregate FDI would only be there if one assumes that signing a treaty with any single country thereby protects the investment of a foreigner from any other country. It is certainly possible that ratifying a treaty might have a signaling effect. But as the standards of protection can be greater than what is afforded to domestic investors, and none of the legal protections would hold to the non-signatory investors, this is certainly not automatic. The effects are surely strongest for those investors that are actually covered, namely those involved in the bilateral relationship. Testing for the effects of ratifying the treaties themselves is the cleanest test of whether the treaty has the desired effect of stimulating flows. This is then tested in three ways: analyzing the bilateral level of FDI inflows, examining the share of the host’s inward FDI that comes from the source country, and studying the share of outward FDI from the source country that goes to the particular host country. D. Data This chapter focuses on the importance of BITs for FDI outflows from OECD countries to developing country hosts. This is because almost all but the most recent BITs are ratified between OECD countries and developing countries. Also, the vast majority of FDI inflows into developing countries originate from OECD countries. As the rationale for a host to ratify a BIT is most applicable for developing countries where property rights are generally weaker than in OECD countries, this focus facilitates testing the hypothesis that the strengthening of property rights significantly affects FDI flows. The chapter uses bilateral FDI outflows from 20 OECD countries to 31 developing countries.16 It covers the years 1980 to 2000, capturing the surge in the number of ratified BITs. The OECD is the source of over 85% of FDI flows to developing countries, so this chapter covers the vast majority of FDI to developing countries as well as FDI covered by BITs. With the increase in the number of BITs, the share of FDI to developing countries that is covered by treaties has grown tremendously. In 1980, the share of FDI under a treaty was less than 5%, while by 2000, it had grown to about 50% (see Figure 4). This increase in FDI by countries with BITs is largely explained by compositional shifts; as more country pairs ratify treaties, the amount of FDI flows covered increases. What remains to be seen is if the flow between host-source pairs changes significantly with the ratifying of a treaty. In addition to information on the date of ratification of BITs,17 the regressions control for the size of the source country, the size of the host country, the GDP 16. Eight other OECD countries, particularly those that more recently joined the OECD, do not report their FDI outflows and so are not included. 17. UNCTAD publishes both the signing dates of BITs and the dates the treaties were ratified. The distinction is important as the treaty only goes into effect once it is ratified,

360 mary hallward-driemeier figure 4. source of oecd-developing country pairs with a bit and the share of oced fdi they cover 60

Percent

50 40 30 20 10 0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 Share of FDI covered

Share of OECD-LDC country-pairs with a BIT

per capita of the host country, the host country’s macroeconomic stability (proxied by its inflation rate), the host country’s openness to trade (trade over GDP), and the gap in average years of education between the source and host pairs. These data come from the World Bank’s World Development Indicators, and the education variables from Barro and Lee. Different specifications were tested and these were the most consistent explanatory variables and are similar to those used in the location choice literature for MNCs. Recognizing that there could be other important time-invariant characteristics that are unobserved, the regressions are all run using fixed effects.18 Two dummy variables are also included. One dummy is included to capture the effects of the enormous political and economic changes in Eastern Europe and in the former Soviet Union in the 1990s relative to the 1980s. A number of these countries ratified BITs in the early 1990s, so the lack of a dummy could bias upwards the importance of the BIT that was actually due to the regime shifts. A second dummy is added for the NAFTA ratification. NAFTA is not strictly a bilateral investment treaty, but it shares similar language and so is included in the measure of investment treaties. But, unlike a BIT, the treaty was

and there are several cases in which “signed” treaties have never been ratified (e.g., Brazil has signed 13 BITs, but not ratified a single one). The chapter uses the date of ratification of the BIT in all the empirical work. 18. To check for robustness, the regressions were also run using host and source dummies and including host-source pair information on distance, colonial ties, shared language, etc. These geographic and political variables were strongly significant. The rest of the results were not significantly different from the fixed effects estimator, and so both sets are not reported here.

do bilateral investment treaties attract fdi? 361

largely a trade agreement, one that made Mexico a more attractive destination for investment as an export platform to the U.S. and Canada. Again, not controlling for the broader economic change would bias upwards the importance of a BIT that is really due to changes in trade policy. E. Hypotheses The importance of ratifying a BIT is tested for in a number of ways. What is of interest is the change in property rights introduced with the BIT. Thus, the tests rely on the variation over time rather than across countries. Including sourcehost pair fixed effects not only controls for other unobserved characteristics that could affect bilateral investment flows, it means that the significance of the BIT is only identified on changes over time.19 First, a dummy is included in a panel regression that takes the value of one once a BIT has been ratified between a pair of source-host countries. The significance of the coefficient on this variable is, then, a test of the importance of the treaty.20 Related is a test looking at the time horizon over which a BIT might attract additional FDI. One possibility is that there would be a window after the ratifying when FDI might increase. Some investors might delay their investment prior to the ratification, so that there would be a short spike with the ratification. Or, the publicity of the treaty could spark additional investment in the iperiod immediately after the ratification. Dummy variables capturing the three years postratification are included to test for the importance of a window. A related test is to look at a reduced sample of those countries that did ratify treaties during the sample period, and to compare the average FDI in the three year period after a ratification with the average FDI inflow in the three year period prior to the treaty. A third approach is to include a series of dummies: one for the year of ratification, another for each of the five years prior to ratification, and a third for post ratification to see if there are consistent patterns across country pairs. Including dummies for the years leading up to the ratification would also test for

19. The regressions were also run using separate source and host country fixed effects and including various source-host controls such as distance, common language, common border, and colonial links. The results are qualitatively the same. 20. This chapter does treat all BITs equally, when, in fact, there are some differences between them. The general point that BITs strengthen property rights holds across all of them. It is possible that there would be more of an effect if one looked only at those treaties with the strongest investor protections. Because such study would require reading and devising an index measure of several hundred BITs, it is beyond the scope of this chapter. However, if BITs are acting as a substitute for property rights, one would expect that the stronger clauses would be included in treaties with countries that have lower domestic property rights. That there is no evidence that these countries receive additional FDI after signing a BIT would indicate that the effort to classify individual BIT terms is unlikely to be fruitful.

362 mary hallward-driemeier

whether treaties came after increases in FDI. The results to all three tests are consistent, so only the third extension is reported. The hypotheses are tested using both the level of FDI received, and the amount of FDI normalized by the host country’s GDP. While the overall patterns would be expected to be similar, a few differences should be noted. It is well known that FDI in developing countries is concentrated in a few markets, however, these markets are large. If, instead, one looks at FDI compared to GDP, the ratios demonstrate much less variance than the levels. Also, the top recipients of levels of GDP are not among the top receivers once one looks at the ratios. In fact, a number of small countries have a higher ratio. Particularly as investment can be lumpy, a few large investment projects can represent a significant portion of a small economy. One difficulty with these approaches is that the FDI level rose substantially during this period, so dummies for the later period will be significant in part due to the trend in FDI. Adding a trend term can capture this effect, but another test is also developed. Regressing the level of FDI and the ratio of FDI to GDP address whether BITs increase the amount of FDI. A related question is whether BITs simply shift the destinations of the FDI among developing countries. To address this question, the amount of FDI a host receives is normalized by the total amount of FDI outflows from that source. Thus, the share of source X’s FDI to host Y is the dependent variable. The question is, then, whether the host receives a large share of X’s FDI with the conclusion of an investment treaty. BITs are often justified by developing countries as a signal that they will protect the property rights of the foreign investor, thereby strengthening their investment climates. But, the credibility of this signal is affected by the degree of corruption and the quality of the legal system of the host country. The existence of a BIT is thus combined with the quality of the legal system and the extent of corruption to see if the BITs signal is only valuable within a country with a certain level of overall property rights. F. Econometric Concerns It is possible that there is reverse causation: that the existence of extensive FDI flows means the source country has a larger incentive to conclude a BIT with the host country. Thus it is possible that FDI flows increase in the period prior to, or concurrent with, the ratifying of a BIT. This would imply that there is a positive feedback from FDI to the probability that a BIT is ratified. On the other hand, it is also possible that hosts that do not receive much FDI would be interested to sign a BIT as a way of increasing FDI. If this is correct, one would expect a negative feedback from FDI to the presence of a BIT. Which story dominates is an empirical question. This potential endogeneity of a BIT is addressed with the use of instrumental variables. The instrument used is the number of other BITs a host has entered into with countries other than the source country being considered. The willingness

do bilateral investment treaties attract fdi? 363

of a host to ratify a BIT, as measured by the number of outside BITs, should be correlated with the probability that it signs with this particular host country, but shouldn’t affect the amount of FDI that particular source country would send. Thus, when U.S. investors are considering investing in India, their decision would not be affected by whether India has ratified treaties with the UK or France, however, the fact that India has entered other treaties would be expected to influence their willingness to enter such a treaty with the U.S.21 One of the shortcomings of the data is that a great number of cells are left blank. The data comes from the source country, but they do not necessarily report all the FDI to each of the host countries. Thus, it is difficult to know if the blank represents a zero or simply a non-reported number. What is clear, however, is that the true value of the blank cells is less than the values that are reported. To deal with this issue, regressions are reported only using the data that is published. In addition, a number of rules were used to fill in blanks with zeros. Regressions were run using the different rules for missing values. The results remained consistent, so what is also reported is the more expansive inclusion of zeros. Blanks were filled in only for years after a source began reporting (e.g., some don’t report until 1985), and only if at least five other values are reported for that source for that year (e.g., the UK did not report any amounts in 1984, so none of these values were filled in as zeros). Following these rules nearly doubles the sample. It should be noted that a number of source-host pairs only have zeros (e.g., New Zealand-Czech Republic and Portugal-Thailand) and some of these pairs have BITs, although others do not.22 G. Results 21. It is possible that a U.S. MNC with a French subsidiary could invest in India via its French subsidiary rather than directly from the parent company so as to have the Indian plant covered by a BIT. The widespread use of such a practice would undermine the validity of the instrument, however, this possibility is one that is safeguarded against in most BITs. Not wanting to extend rights to investors that have only weak or tenuous links to the treaty partners, standards of nationality are spelled out in the treaties. These include “substantial ownership,” “ability to exercise decisive control,” and “principle place of business,” in addition to the location of incorporation (UNCTAD 1998, pp. 39–41). Furthermore, as a practical matter, if there were such flows they would be expected to bolster a finding that BITs attract FDI (which we don’t find in the data) and the actual correlation between FDI flows and the number of treaties the host has signed with other countries is 0.03. On the other hand, if the diversion of funds through third countries were common, the presence of additional alternative channels would then be expected to be negatively associated with FDI flows. 22. Another way to deal with the cutoff is to treat the sample as a truncated one; to replace the zero and negative observations with the lowest positive value in the dataset, and to estimate the regressions with a Tobit specification. The drawback with this approach is that neither fixed effects nor instruments can be incorporated, and the information on known negative flows is lost. It turns out that there are a significant number of negative flows between pairs with a treaty, and that losing this information influences the results.

364 mary hallward-driemeier

Table 1 illustrates the level of FDI flows. Column (1) reports the findings using the level of FDI for all the reported bilateral pairs using a fixed effects estimator to control for time invariant host, source, and host-source effects. Column (2 repeats the regression, using the augmented series that fills in missing amounts with zeros as discussed above. Including the additional zeros nearly doubles the sample size, and has little impact on the qualitative results while increasing the significance of the findings. The effect of the control variables are robust and of the expected sign: the larger the source and host countries, the larger the FDI flow. Flows are also higher for richer host countries. Macroeconomic instability discourages FDI. A host’s trade openness could be ambiguous if source countries are looking to jump tariffs. The negative finding would be consistent with that, but a more plausible explanation is that trade to GDP ratios are often higher for small countries. In which case, this measure is likely further evidence that larger FDI flows go to larger countries. The NAFTA dummy is large and significant, capturing the increase in FDI to Mexico with the implementation of this free trade deal. This is one of the few strong pieces of evidence that an investment treaty could stimulate investment, but, as it is tied to a trade agreement with the world’s largest market, it is hard to disentangle which effect really dominates. The coefficient on the BIT treaty is negative and not significant. Breaking down the effect of a BIT over the years preceding and following the ratification of a treaty illustrates that there is little positive association for a ten year window (see table 1, column 3). Only in the fifth year after the ratification is there a positive (and extremely weak) association. Controlling for the possible endogeneity of the decision to enter a BIT, columns four and five present the results from the IV estimation. The instrument is the number of BITs the host has entered into with other countries, a number positively correlated with the probability that it enters a BIT with the source, but should not be affecting the amount of FDI received from that source country. The results lead to a significant negative finding on the impact of ratifying a BIT. Assuming the instrument is valid, this implies there would otherwise be a positive feedback from larger investment flows encouraging the ratification of a BIT. Including the “missing zeros” still leads to a negative finding of a BIT, in which the coefficient’s fall corroborates the inference of the positive feedback in the non-IV regressions. The same set of regressions was repeated, this time looking at the ratio of FDI to host GDP (see Table 2). This normalization, however, leads to somewhat different interpretations. While larger countries get more FDI in absolute numbers, the ratio of FDI to GDP is highest for smaller countries. Now, the size of the source country is not significant and the size of the host is negative. Controlling for size, richer hosts do receive more, however. In these regressions the impact of a BIT is totally insignificant, even when instrumented for. Looking at the window around the ratification, there is weak evidence that the ratio

do bilateral investment treaties attract fdi? 365

table 1. levels of fdi flows

Source GDP Host GDP Host GDPPC Host Inflation Host Trade/GDP Skill gap E. Europe90s NAFTA BIT treaty Yr Ratify –5

(1) FDI Flow

(2) FDI Flow w/0s

(3) FDI Flow w/0s

(4) IV FDI flows

(5) IV FDI Flow w/0s

0.176

0.163

0.163

0.170

0.151

(13.79)∗∗

(23.34)∗∗

(23.27)∗∗

(12.74)∗∗

(18.10)∗∗

0.092

0.078

0.072

0.090

0.158

(4.37)∗∗

(7.50)∗∗

(6.94)∗∗

(4.19)∗∗

(8.71)∗∗

12.274

11.499

11.772

12.864

29.747

(1.80)+

(3.83)∗∗

(3.93)∗∗

(1.86)+

(6.39)∗∗

–6.193

–3.188

–3.271

–6.979

–6.813

(3.90)∗∗

(3.74)∗∗

(3.81)∗∗

(4.16)∗∗

(5.85)∗∗

–136.290

–46.329

–51.882

–166.602

–35.077

(2.55)∗

(1.81)+

(2.01)∗

(2.91)∗∗

(1.18)

11.703

7.634

7.928

16.159

25.171

(0.91)

(1.25)

(1.30)

(1.21)

(3.28)∗∗

–10.440

6.742

–7.186

22.878

182.407

(0.27)

(0.35)

(0.37)

(0.51)

(4.88)∗∗

256.311

196.005

198.304

227.505

97.975

(5.24)∗∗

(6.84)∗∗

(6.94)∗∗

(4.33)∗∗

(2.64)∗∗

–

–207.520

–101.320

(1.67)+

(1.90)∗∗

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–11.360

–11.615

(0.51)

(0.98)

–

–

–14.641 (0.67)

Yr Ratify –4

–

–

–13.718 (0.65)

Yr Ratify -3

–

–

–16.360 (0.80)

Yr Ratify –2

–

–

–25.177 (1.26)

Yr Ratify -1

–

–

–37.388 (1.91)+

Year Ratify

–

–

–40.503 (2.11)∗

Yr Ratify +1

–

–

–54.577 (2.86)∗∗

Yr Ratify +2

–

–

–31.512 (1.65)

Continued

366 mary hallward-driemeier table 1. levels of fdi flows (cont’d...) (1) FDI Flow

(2) FDI Flow w/0s

(3) FDI Flow w/0s

(4) IV

FDI flows

(5) IV FDI Flow w/0s

Yr Ratify +3

–

–

–

–

Yr Ratify +4

–

–

–

–

Yr Ratify +5

–

–

–

–

Constant

–162.401

–110.477

–17.467 (0.86) –4.025 (0.19) 2.760 (0.12) –106.870

–193.177

–229.021

(1.83)+

(3.41)∗∗

(3.30)∗∗

(2.72)∗∗

(6.04)∗∗

4261 434 0.16 – –

8153 537 0.13 – –

8153 537 0.13 – –

4261 434

8153 537

1390.30 0.00

1803.93 0.00

No. Obs. No. pairs R-squared Wald Chi2 Prob > Chi2

Absolute value of t-statistics in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1% Source-host country pairs included; year dummies not reported.

table 2. ratio of fdi to gdp

Source GDP Host GDP Host GDPPC Host Inflation Host Trade/GDP Skill gap

(1) Ratio

(2) Ratio w/0s

(3) Ratio w/0s

(4) IV Ratio

(5) IV Ratio w/0s

0.030 (0.71) –0.229

0.029 (1.52) –0.121

0.032 (1.66)+

–0.127

0.024 (0.32) –0.220

0.033 (1.64) –0.147

(2.74)∗∗

(3.17)∗∗

(3.35)∗∗

(1.84)+

(2.61)∗∗

0.184

0.101

0.106

(2.78)∗∗

(2.94)∗∗

–0.000 (0.63) –0.011 (0.51) –0.007 (1.43)

–0.000 (0.90) 0.002 (0.17) –0.003 (1.40)

–0.000 (1.00) –0.001 (0.09) –0.003 (1.36)

0.176 (1.57) –0.000 (0.52) –0.010 (0.41) –0.007 (1.42)

0.131

(2.25)∗

(2.19)∗

–0.000 (1.08) 0.003 (0.25) –0.003 (1.30)

do bilateral investment treaties attract fdi? 367

(1) Ratio

(2) Ratio w/0s

(3) Ratio w/0s

(4) IV Ratio

(5) IV Ratio w/0s

0.015 (1.50) 0.009 (0.81) 0.003 (0.67) –

0.010 (0.93) 0.009 (0.77)

Yr Ratify –5

0.020 (1.05) 0.007 (0.40) 0.004 (0.42) –

0.019 (0.99) 0.009 (0.38) 0.013 (0.14) –

0.018 (1.62) 0.006 (0.45) –0.020 (0.53) –

Yr Ratify –4

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

0.915 (1.29) 4261 434

0.253 (0.92) 8153 537

705.78 0.00

707.2 0.00

E. Europe90s NAFTA BIT treaty

–0.013 (1.54) –0.014 (1.76)+

Yr Ratify -3

–

–

–0.015 (1.95)+

Yr Ratify –2

–

–

–0.014 (1.77)+

Yr Ratify -1

–

–

–0.018 (2.37)∗

Year Ratify

–

–

–0.019 (2.56)∗

Yr Ratify +1

–

–

–0.023 (3.09)∗∗

Yr Ratify +2

–

–

Yr Ratify +3

–

–

Yr Ratify +4

–

–

Yr Ratify +5

–

–

Constant

0.879 (1.45) 4261 434 0.05 – –

0.241 (0.88) 8153 537 0.03 – –

No. Obs. No. pairs R-squared Wald Chi2 Prob > Chi2

–0.011 (1.55) –0.010 (1.30) –0.004 (0.47) –0.004 (0.44) 0.240 (0.87) 8153 537 0.03 – –

Absolute value of t-statistics in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1% Source-host country pairs included; year dummies not reported.

368 mary hallward-driemeier

of FDI to GDP rises—or at least loses the negative values—before the date of ratification. A final set of regressions looks at the FDI going to a particular host country as a share of the total FDI the source country sends. The results are reported in Table 3. Larger host countries do not necessarily get a larger share, although more developed ones do. Here one gets the single significant positive result that a BIT could increase FDI (see column 2), however, the result seems to come from the period five or more years after the ratification of the treaty. And, instrumenting for the ratification of the treaty reverses the sign on this coefficient. While these findings suggest that BITs do not serve to attract additional FDI, it is possible that this is due to BITs’ effects being obscured by other changes that are occurring between the two signatories over time. Such changes could include lowering trade barriers, increasing knowledge of conducting business in the host country, and following customers abroad. Nevertheless, these changes would likely work to increase the likelihood of investing overseas, so if the BIT variable is capturing some these effects, one would expect it to bias up the coefficient. One possible change that could work in the other direction is the ratification of a tax treaty. Blonigen and Davis (2002) find that the signing of a tax treaty could reduce FDI, and, if a tax treaty is entered into at the same time as a BIT, this could weaken the observed effect of the BIT. But their result stems from intra-OECD FDI flows; it is not clear whether their result would extend to OECD FDI into developing countries, particularly when so many now enjoy various numbers of tax holidays. In addition, there is little evidence that tax treaties and BITs are entered to at the same time. Tables 4, 5, and 6 report the results from testing the hypothesis that the quality of domestic institutions may be important in determining the effectiveness of a BIT in attracting FDI. The strong hypothesis is that BITs will be more effective in attracting FDI in weak institutional settings, acting as substitutes for strong domestic protection of property rights. A weaker hypothesis is that BITs would act as complements, reinforcing good domestic institutions. A follow-on from this is that if there is a threshold effect, a BIT would only serve to attract FDI once a certain level of domestic institutional quality was achieved. A positive interaction term on institutional quality and the ratification of a BIT would show that the two, domestic institutional quality and a BIT, are self-reinforcing, however, the sum of the coefficients will determine whether there is a threshold effect and the extent to which a BIT can be a substitute for weaker domestic institutions. The results show little evidence that a BIT can act as a substitute for weak domestic institutions. There is more evidence that a BIT can reinforce the quality of institutions, but there is a threshold effect that is reasonably high in most cases. In addition, most specifications show that the net impact of a BIT remains non-significant, or even negative.

do bilateral investment treaties attract fdi? 369

table 3. share of source countries’ fdi sent to host

Source GDP Host GDP Host GDPPC

(1) (2) (3) Share sent Share w/0s Share w/0s

(4) IV (5) IV Share sent Share w/0s

–0.007 (0.80) –0.025 (1.53) 0.025 (1.52) –0.001

0.007

0.007

0.032

0.012

(1.89)+

(1.86)+

(1.88)+

(2.81)∗∗

(5.39)∗∗

–0.016

–0.017

–0.079

–0.047

(2.03)∗

(2.21)∗

(2.90)∗∗

(3.90)∗∗

0.016

0.017

0.074

0.051

(2.15)∗

(2.35)∗

(2.85)∗∗

(4.04)∗∗

–0.000

–0.000

–0.001

–0.000

(4.52)∗∗

(4.45)∗∗

(5.42)∗∗

(5.47)∗∗

–0.007

–0.006

–0.006

–0.014

–0.005

(1.74)+

(2.97)∗∗

(2.93)∗∗

(2.61)∗∗

(2.33)∗

–0.000 (0.38) –0.001 (0.60) 0.001 (0.54) 0.002

–0.000 (0.32) –0.002 (0.78) 0.001 (0.47)

–0.002 (1.52) 0.002 (0.42) –0.008 (1.53) –0.057

0.000 (0.14) 0.001 (0.61) –0.003 (1.14) –0.026

(2.63)∗∗

(3.28)∗∗

Yr Ratify –5

–0.002 (1.57) –0.001 (0.33) 0.001 (0.22) 0.002 (1.43) –

–

–

Yr Ratify –4

–

–

–

–

Yr Ratify -3

–

–

–

–

Yr Ratify –2

–

–

–0.002 (0.93) –0.002 (1.07) –0.002 (1.16) –0.003

–

–

–

–

–

–

–

–

–

–

–

–

Host Inflation Host Trade/GDP Skill gap E. Europe90s NAFTA BIT treaty

(2.05)∗

–

(1.88)+

Yr Ratify -1

–

–

Year Ratify

–

–

Yr Ratify +1

–

–

Yr Ratify +2

–

–

Yr Ratify +3

–

–

–0.002 (1.27) 0.000 (0.03) 0.001 (0.61) 0.000 (0.22) 0.001 (0.34)

Continued

370 mary hallward-driemeier table 3. share of source countries’ fdi sent to host (cont’d...)

Yr Ratify +4 Yr Ratify +5 Constant

(1) (2) (3) Share sent Share w/0s Share w/0s

(4) IV (5) IV Share sent Share w/0s

– – – – 0.219

– – – – –0.011 (0.07) 4261 434

– – – – –0.012 (0.20) 8153 537

461.21 0.00

522.77 0.00

(1.81)+

No. Obs. No. pairs R-squared Wald Chi2 Prob > Chi2

4261 434 0.03 – –

– – – – –0.026 (0.47) 8153 537 0.02 – –

0.001 (0.35) 0.003 (1.73)+

–0.021 (0.38) 8153 537 0.02 – –

Absolute value of t-statistics in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1% Source-host country pairs included; year dummies not reported.

In Table 4, columns 1, 2, and 3 report the results from the Kaufmann, Kraay, Zoido-Lobaton (KKZ) indicator measure of the rule of law using the level of FDI, its share in GDP and the share of the source country’s FDI the host receives. The effect is insignificant for the level of FDI and the share of the source country’s FDI. However, it is significantly positive for the ratio of FDI to GDP, providing some evidence that a BIT can reinforce good quality domestic institutional quality. While this governance measure is the closest to the common justification given for signing a BIT, to test for robustness, other measures of institutions were also used. Table 4 also reports the results for corruption and Table 5 for regulatory quality and government effectiveness. These measures also provide more evidence of a positive interaction: that a BIT complements strong domestic institutions. But, for the level of FDI, in order for the interaction to offset the negative impact of the BIT, the quality of institutions would have to be strong—at least 1.2 standard deviations above the mean (e.g., around the level of Chile) and, for corruption, 2 standard deviations above the mean. Table 6 repeats the regressions using the ICRG measures of law and order and corruption. These measures have the strong advantage of including time variation in the quality of institutions. With country dummies included, it captures the effect of changes in institutional quality. For the ICRG rule of law measure, the interaction term is only significantly positive in one case. For corruption, the interactions are all positive, however, the net effect of a BIT is still

do bilateral investment treaties attract fdi? 371

table 4. interaction of bit and the rule of law and corruption (kkz) (1) Level of FDI

Source GDP Host GDP Host GDPPC

NAFTA

0.009

0.163

0.036

0.009

(2.22)∗

(22.85)∗∗

(1.85)+

(2.28)∗

0.091

–0.097

–0.022

0.094

–0.110

–0.022

(7.05)∗∗

(2.20)∗

(2.48)∗

(7.39)∗∗

(2.48)∗

(2.45)∗

19.309

13.487

0.069 (1.57) –0.000 (1.18) 0.001 (0.08) –0.004

(2.08)∗

(1.65)+

174.251

0.012 (0.99) 0.008 (0.77) –0.000 (0.00) 0.070

–3.760 –44.104

52.865 (2.11)∗

BIT

–124.365 (2.34)∗

BIT∗Rule of Law BIT∗Corruption Constant

–78.310 (0.57) – –190.700

(4.44)∗∗

–

0.023

10.956

0.081

0.022

(2.56)∗

(2.89)∗∗

(1.81)+

(2.48)∗

–0.000

–3.969

(4.88)∗∗

(4.31)∗∗

–0.001 (1.46) –0.000 (0.05) –0.004

–0.006

–46.566

(2.86)∗∗

(1.81)+

–0.000 (0.32) 0.000 (0.12) –0.001 (0.47) –0.005 (1.27) 0.004 (1.27) –

8.328 (1.31) 182.449

–0.000 (4.99)∗∗

–0.006 (2.93)∗∗

17.397 (0.67) –85.700 (1.60) –

0.011 (0.92) –0.003 (0.32) 0.028 (1.34) –

–0.000 (0.44) 0.000 (0.16) –0.002 (1.01) –0.003 (0.60) –

85.330

0.097

0.008

(1.90)+

(6.45)∗∗

(2.68)∗∗

–141.770

–0.027 (0.48) 8153 537 579.96 0.00

(6.09)∗∗

(1.84)+

8153 537

0.171 (0.62) 8153 537

–0.027 (0.48) 8153 537

8153 537

0.197 (0.71) 8153 537

1792.97 0.00

727.19 0.00

574.3 0.00

1809.56 0.00

745.98 0.00

(5.38)∗∗

Observations Number of source partner pairs Wald Chi2 Prob > Chi2

(6) Share of source FDI

0.033

(5.81)∗∗

E. Europe 90s

(5) FDI/ GDP

(1.71)+

(1.70)+

Skill gap

(4) Level of FDI

0.160

(4.06)∗∗

Host Trade/GDP

(3) Share of source FDI

(22.49)∗∗

(4.48)∗∗

Host Inflation

(2) FDI/ GDP

(4.27)∗∗

Absolute value of z-statistics in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1% Country pair fixed effects included; year dummies not reported.

372 mary hallward-driemeier table 5. interaction of bit and regulatory quality and government effectiveness (kkz)

Source GDP

(1) Level of FDI

(2) FDI/ GDP

(3) Share of source FDI

(4) Level of FDI

0.162

0.008

0.102

0.030 (1.56) –0.087

(6.68)∗∗

(1.96)∗

11.034

0.068 (1.53) –0.000 (1.24) 0.001 (0.10) –0.004

(22.76)∗∗

Host GDP Host GDPPC

(2.99)∗∗

Host Inflation

–4.070 (4.28)∗∗

Host Trade/GDP Skill gap NAFTA

–40.660 (1.57) 6.794 (1.04) 178.727

114.636

0.011 (0.91) 0.008 (0.80) –0.004 (0.20) 0.064

(1.69)+

(3.18)∗∗

–

–

(5.93)∗∗

E. Europe 90s BIT

16.032 (0.63) –136.134 (2.19)∗

BIT∗Regulatory Q Quality BIT∗Government Effectiveness Constant

–140.031

(6) Share of source FDI

0.162

0.035

0.009

(2.09)∗

(22.79)∗∗

(1.80)+

(2.17)∗

–0.022

0.092

–0.112

–0.022

(2.42)∗

(7.34)∗∗

(2.50)∗

(2.40)∗

0.023

11.694

0.078

0.022

(2.52)∗

(3.03)∗∗

(1.73)+

(2.44)∗

–0.000

–3.876

(4.59)∗∗

(4.23)∗∗

–0.000 (0.98) 0.000 (0.00) –0.005

–0.006

–46.230

(2.86)∗∗

(1.79)+

–0.000 (0.28) 0.000 (0.16) –0.001 (0.36) –0.004 (0.81) 0.000 (0.08) –

8.491 (1.32) 179.302

8153 537

0.105 (0.38) 8153 537

–0.024 (0.43) 8153 537

1808.41 0.00

719.73 0.00

573.26 0.00

(4.18)∗∗

Observations Number of source_partner Wald Chi2 Prob > Chi2

(1.81)+

(5) FDI/ GDP

(5.98)∗∗

30.390 (1.23) –110.332 (2.08)∗

– 59.957 (1.40) –144.729

(2.19)∗

0.009 (0.76) 0.005 (0.45) 0.016 (0.77) –

–0.006 (2.90)∗∗

–0.000 (0.46) 0.000 (0.14) –0.001 (0.58) –0.003 (0.77) –

8153 537

0.255 (0.92) 8153 537

0.004 (1.48) –0.023 (0.41) 8153 537

1806.33 0.00

745.19 0.00

575.87 0.00

(4.26)∗∗

Absolute value of z-statistics in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1% Country pair fixed effects included; year dummies not reported.

0.089

–0.000 (4.74)∗∗

(6.12)∗∗

do bilateral investment treaties attract fdi? 373

table 6. interaction with law and order and corruption (icrg) (1) (2) Level FDI FDI/ GDP

Source GDP

0.180

(3) Share of Source FDI

(4) (5) Level FDI FDI/ GDP

(6) Share of Source FDI 0.013

(6.16)∗∗

Host GDP

0.082

Host GDPPC

7.656

Host Inflation

–6.116

0.006 (0.28) –0.065 (1.30) 0.051 (1.07) –0.001

(5.09)∗∗

(1.82)+

Host Trade/GDP –70.514

–0.011 (0.87) –0.001 (0.28) 0.017 (1.42) 0.018 (1.59) –0.032

0.000 (0.06) –0.003 (1.27) –0.005 –0.023

–0.013 (1.04) –0.002 (0.60) 0.010 (5.41)∗∗ (0.82) –41.788 0.017 (1.43) (1.47) –251.702 –0.020

(2.52)∗

(2.27)∗

(2.63)∗∗

(1.67)+

(4.42)∗∗

–

–

–

–

–

–

–41.640

–0.012

–0.004

(4.35)∗∗

(3.35)∗∗

(5.62)∗∗

89.531

0.032

0.008

(3.76)∗∗

(3.67)∗∗

(5.01)∗∗

(19.46)∗∗ (5.92)∗∗ (1.70)+

(2.23)∗

Skill gap NAFTA

5.609 (0.79) 122.192 (3.77)∗∗

E. Europe 90s BIT Rule of Law BIT∗Rule of Law Corruption BIT∗Corruption Constant

–38.596 (1.31) –17.413 (0.13) –43.280

0.014

0.183

(3.23)∗∗

(19.55)∗∗

–0.027

0.089

(2.46)∗

(6.09)∗∗

0.028 –0.001

6.537 (1.45) –5.962

0.007 (0.33) –0.069 (1.38) 0.055 (1.14) –0.001

(6.68)∗∗

(4.88)∗∗

(1.71)+

–0.006

–41.056 (1.28) 10.019 (1.39) 175.104

(2.57)∗

(2.47)∗

(1.96)+

–0.011

–0.003

(4.94)∗∗

(3.39)∗∗

(5.22)∗∗

9.980 (0.41) –

0.005 (0.59) –

(2.04)∗

– 7.859 (0.15) 6952 537

Observations Number of source_partner Wald Chi2 1609.87 Prob > Chi2 0.00

–

0.004 – –

(2.97)∗∗

–0.023 (2.07)∗

0.027 (2.43)∗

–0.001 –0.005 (2.13)∗

0.001 (1.54) 0.000 (0.19) –0.004 (1.69)+

–0.032

0.348 (1.06) 6952 537

–0.070 (1.10) 6952 537

–10.741 (0.25) 6952 537

0.206 (0.63) 6952 537

–0.087 (1.37) 6952 537

671.76 0.00

615.24 0.00

1543.19 0.00

662.68 0.00

598.91 0.00

Absolute value of z-statistics in parentheses + significant at 10%; ∗ significant at 5%; ∗∗ significant at 1% Country pair fixed effects included; year dummies not reported.

374 mary hallward-driemeier

negative in five of the six cases. The threshold of institutional quality needed to make the net effect positive is over 3 standard deviations above the mean. Thus, the evidence suggests that BITs are more effective in settings of higher institutional quality and where institutions are already being strengthened, however, the net impact of a BIT remains insignificant or negative for the majority of countries. It is only those countries with the strongest domestic institutions that are shown to benefit, albeit slightly, from a BIT. As the majority of the developing countries signing BITs do not rank as having strong domestic institutions, the evidence does not show that a BIT can substitute for this weakness. The evidence to date does not support the hope of developing countries that a BIT can bypass weak property rights and ineffectual institutions. If host countries are committed to trying to attract more FDI, BITs have not provided an alternative to the need to implement broader reforms of domestic institutions.

conclusion Recent and pending cases of international investment disputes covered by investment treaties have raised concerns of the potential costs to host governments, both in terms of the size of potential awards and in the possible reduction of viable choices open to policy makers due to their adverse effects on foreign investors. Critics speculate that these cases will serve to encourage firms to look for ways to exploit the terms of the treaty as a lucrative way of doing business, seeking compensation for risks that they had not previously expected to be protected from. Given the increasing concern about the potential, and often unanticipated, costs of BITs, it is all the more important to examine whether BITs are delivering their expected benefits. If so, policy makers have the task of weighing the benefits and potential costs against one another. But, if there is little apparent benefit, the case to ratify new agreements—at least under terms that are extremely favorable to the investor—is harder to make. It is not that formalization of relations and treaties that protect against dynamic inconsistency problems should not be encouraged, just that the terms of these agreements and the strength of the rights given to investors should be scrutinized. An analysis of twenty years of bilateral FDI flows from the OECD to developing countries finds little evidence that BITs have stimulated additional investment. Those countries with weak domestic institutions, including protection of property, have not received significant additional benefits: a BIT has not acted as a substitute for broader domestic reform. Rather, those countries that are reforming and already have reasonably strong domestic institutions are most likely to gain from ratifying a treaty. That BITs act as more of a complement to than a substitute for domestic institutions means that those that are benefiting from them are arguably the least in need of a BIT to signal the quality of their property rights.

do bilateral investment treaties attract fdi? 375

It is possible that in a few years a different result will emerge. The publicity surrounding the investor protection cases being brought under NAFTA’s Chapter 11, and the cases being brought against Argentina as it dissolved its currency board, may make prospective investors more aware of the potential gains they would have under a BIT and insist on such terms. On the other hand, policymakers may take greater care to refine the expropriation and compensation clauses to ensure that the worst fears of the critics are not realized, and bring closer together the relative costs and benefits of BITs.

376 mary hallward-driemeier

appendix Recent cases on compensation of expropriation, highlighting regulatory takings Most of the recent publicized cases have arisen under NAFTA’s Chapter 11. While not strictly a bilateral agreement, the terms are the same as those used in many BITs. The cases below illustrate the types of obligations other signatory host countries could face. While cases like these have been brought by OECD multinationals in developing countries before, these are some of the first cases where MNCs have sued rich OECD host governments. The outcomes add insight into the reasons that OECD governments have refused to enter into other agreements that would give such rights to foreign companies operating within their borders, while at the same time wanting such rights for their own MNCs overseas. It should be noted that these cases have not all been settled, and that the prospect of expansive regulatory taking claims may not be upheld. Even so, the size of the suits and the potential constraints on policy choices should give host country signatories pause over the precise nature of the terms they agree to. Concerned about the possible health risks associated with a gasoline additive, methylcyclopentadienyl manganese tricarbonyl (MMT), Canada considered banning it (it was already effectively banned in the U.S.). Ethyl Corporation, an American company and the sole supplier of MMT in Canada, filed the first Chapter Eleven case. After instituting a ban, Canada’s parliament then reversed course, lifting the ban and paying Ethyl $13 million for damages incurred during the time the ban was in place. Avoiding the $200 million suit was not the only consideration, but the subject was widely discussed in the deliberations of the issue. The threat of another lawsuit also served to thwart a proposed health reform bill in Canada. Canada was proposing to increase the warnings on cigarette packaging. R.J. Reyolds and other tobacco firms threatened a lawsuit and the reform measure was dropped. Since the signing of NAFTA, only two new environmental regulations have been considered in Canada, and both have been challenged under Chapter Eleven. In the U.S., there is a case pending that will be extremely influential in determining the scope of such claims. The case regards another gasoline additive, methyl tertiary butyl ether (MTBE). Originally hailed as a means of improving air quality by enabling gas to burn more cleanly, it has since been discovered to have tainted the water supply and has been linked to cancer in laboratory animals. California decided in 1999 to ban the additive. Its maker, Methanex, a Canadian corporation is suing for $970 million in lost profits. Another high profile case was just resolved. The case involved the Loewen Group, a Canadian funerary home company. A Mississippi competitor had

do bilateral investment treaties attract fdi? 377

successfully brought Loewen Group to court on antitrust violations. Loewen group settled the case, agreeing to pay $150 million. Four years later, it sued the U.S. government claiming that it had been denied due process in the Mississippi courts (part of their claim was based on instructions and comments made to the jury that were characterized as anti-foreign and racially biased.), and sought close to $500 million in compensation. The case was registered four years ago and was just dismissed on jurisdictional grounds as the Loewen group had been bought by a U.S. interest. Another case that generated a lot of attention in the press is that of Metalclad, a U.S. waste disposal company that attempted to set up facilities in Mexico. Despite federal government assurances, local officials denied a building permit due to failures to clean up waste that was entering the water table and in response to intense protest from local residents. Metalclad sued and was awarded $16 million—a sum that had been reduced from the original amount sought due to the determination that expected profits would not have been that high.

references Acemoglu, Daron, Simon Johnson and James Robinson. (2001). “The colonial origins of comparative development: an empirical investigation,” American Economic Review, 91, 5, pp. 1369–1401. Been, Vicki and Joel Beauvais. (2003). “The Global Fifth Amendment: NAFTA’s Investment Protections and the Misguided Quest for an International Regulatory Takings Doctrine,” NYU Law Review, 78, 1, 30–133. Blonigen, B. and R. Davis. (2000). “The effects of bilateral tax treaties on U.S. FDI activity,” NBER Working Paper 7929 (Cambridge: National Bureau of Economic Research). Dollar, David and Aart Kraay. (2002). “Institutions, trade and growth,” Paper prepared for the Carnegie Rochester Conference on Public Policy. European Union. (1998). “Civil society consultation on trade and investment: report,” (Brussels: EU), available at: http://europa.eu.int/comm/trade/miti/ invest/1meeinne.htm. Guzman, A. (1998). “Why LDC’s sign treaties that hurt them; explaining the popularity of bilateral investment treaties,” Virginia Journal of International Law, 38, pp. 639–688. ICSID Review – Foreign Investment Law Journal. (1999). Case No. ARB(AF)/97/2, Robert Azinian, Kenneth Davitian and Ellen Baca v. The United Mexican States, Award, November 1. ——. (2002). Case No. ARB(AF)/99/2, Mondev International Lt. v. United States of America, Award, October 11, para. 159. Kaufmann, Dani, Aart Kraay and Massimo Mastruzzi. (2006). “Governance Matters V: Aggregate and Individual Governance Indicators 1996–2005.” World Bank mimeo. Knack, Stephen and Philip Keefer. (1995). “Institutions and economic performance: cross-country tests using alternative institutional measures,” Economics and Politics, 7, 3, pp. 207–227.

378 mary hallward-driemeier Moyers, Bill. (2002). “Trading democracy,” for PBS, February 5. Neumayer, Eric and Laura Spess. (2004). “Do bilateral investment treaties increase foreign direct investment to developing countries?” (London: London School of Economics). OECD. (2001). “Investment compact regional roundtable on bilateral investment treaties for the protection and promotion of foreign investment in South East Europe,” May 28–29 (Croatia: OECD). Perez-Lopez, Jorge and Matias Travieso-Diaz. (2001). “The Contribution of BITs to Cuba’s Foreign Investment Program.” Georgetown Journal of Law and Policy in International Business, 32, 3, pp. 529–570. Rodrik, Dani, Arvind Subramanian, and Francesco Trebbi. (2002). “Institutions rule: the primacy of institutions over geography and integration in economic development,” NBER Working Paper No. 9305. Stein, E. and C. Daude. (2001). “Institutions, integration, and the location of foreign direct investment,” (Washington, D.C.: Inter-American Development Bank), mimeo. Svensson, Jakob. (1998). “Investment, property rights and political instability: theory and evidence,” European Economic Review, 42, 7, pp. 1317–1341. Tobin, Jennifer and Susan Rose-Ackerman. (2004). “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Research Paper No. 293 (New Haven, CT: Yale Law School Center for Law, Economics and Public Policy). UNCTAD. (1997a). “Paper submitted by the expert from Japan to the Expert Meeting on Existing Agreements on Investment and their Development Dimensions,” May 28–30 (Geneva: United Nations). ——. (1997b). “Paper submitted by the expert from Lebanon to the Expert Meeting on Existing Agreements on Investment and Their Development Dimensions,” May 28–30 (Geneva, United Nations). ——. (1998). Bilateral Investment Treaties in the Mid-1990s (New York and Geneva: United Nations). Vandevelde, K. (1998). “Investment liberalization and economic development: the role of bilateral investment treaties,” Columbia Journal of Transnational Law (New York, Columbia University). World Bank. (2002). Global Economic Prospects (Washington, D.C.: World Bank). Yeyati, E., E. Stein and C. Daude. (2002). “Regional Integration and the Location of FDI,” (Washington, D.C.: Inter-American Development Bank), mimeo.

14. do bit s really work? revisiting the empirical link between investment treaties and foreign direct investment∗ jason yackee introduction Eric Neumayer and Laura Spess (2005) recently published in the journal World Development an article of great practical importance to the world’s capital-hungry developing countries. They presented the first published, peer-reviewed, methodologically sophisticated econometric analysis of the effects that bilateral investment treaties (BITs) might have on foreign direct investment (FDI) inflows. To skeptics of the utility of international law generally, and of the desirability of BITs specifically, the results were startling. The authors provided robust statistical evidence that developing countries that sign BITs with important capital-exporting countries enjoy potentially massive increases in FDI. A developing country signing large numbers of BITs might expect its FDI inflows to increase by up to 93%. As the authors concluded, [c]learly these are nonnegligible increases. . . . But whether the demonstrated benefits of signing BITs in the form of increased FDI inflows are higher than the substantial costs developing countries incur in negotiating, signing, concluding, and complying with the obligations typically contained in such treaties is impossible to tell. What we do know is that BITs fulfill their purpose, and those developing countries that have signed more BITs. . . . are likely to receive more FDI in return. (Id. at 1583). The article seems to put the lie to the assertion of Sornarajah (2004), a prominent critic of BITs, that the effect of BITs on FDI is an “untested hypothesis” and merely an “assumption,” and that “[s]tability and other factors have a greater influence on investment flows than do [BITs].” Or does it? In this brief chapter, I replicate, expand, and critique Neumayer and Spess’s study with the purpose of probing how much, in fact, we do “know” about the effects of BITs on FDI inflows. Given the very real potential that BITs impose high sovereignty costs, it becomes imperative to double-check strong

∗

The author would like to thank Eric Neumayer for providing very helpful comments on an earlier draft. This article is based upon work supported by the National Science Foundation under Grant No. 0418036.

380 jason yackee

claims that the treaties offer tremendous and potentially offsetting economic benefits. My overall aim here is not to criticize Neumayer and Spess’s work as shoddy (it certainly is not). It is instead simply to: bolster confidence that [their] research findings are not merely tied to a particular moment in time or to a particular way of defining a concept. . . . If a hypothesis, model, or theory is viewed as a plausible account of how some aspect of international relations works, then findings that falsify or challenge that account are an important part of the scientific process. Indeed, discovery of such falsifying or challenging evidence through replication is fundamental to the acquisition of reliable knowledge. (Bueno de Mesquita 2003). In other words, I attempt to take seriously Bueno de Mesquita’s point that “findings” do not begin to “take on the role of knowledge” unless they are “capable of being replicated using different data sets” and “different measures.” Note that this is a broad and ambitious view of what replication is, and what epistemological role it can and should play in our profession. “Duplication”—the verification that minor technical errors have not been made, performed by re-running an analysis using the same dataset and specification—can be useful as well, but it is not what I do here. My motivation is not just one of well-meaning double-checking. As I state below, the standard theoretical case for BITs—by which I mean claims that BITs should be expected to lead to large increases in investor confidence, and hence in FDI inflows—is wildly overdrawn. Simply put, the size of the effect of BITs on FDI that Neumayer and Spess report is so outsized that it should raise very serious flags of caution. A. A Skeptical View of BITs It has long been suggested that the primary problem facing would-be foreign investors is of effectively guaranteeing that the host country will not act opportunistically once the investment has been sunk. This problem has been described as one of “obsolescing bargain” in the business-school literature of the 1970s (Vernon 1971) and as one of “credible commitment” in the transactioncost-economics literature of the 1980s (Williamson 1996; Henisz 2002; North 1990). BITs, more than 2,000 of which have now been signed, are said to provide host countries with a particularly effective means of making credible, long-term guarantees to foreign investors that they will be well-treated by the host country throughout the lives of their investments. These treaty-based guarantees are said to be especially credible to the extent that they are backed up by a most important procedural guarantee: that investors shall have the enforceable right to unilaterally initiate binding international arbitration against the host country that breaches its substantive promises of favorable treatment. Many BITs, for example, guarantee investors access to the International Centre for the Settlement of Investment Disputes (ICSID), a World Bank-associated organ that

do bit s really work? revisiting the empirical link between investment 381

has emerged as the most important institutional enforcer of investment-treaty rights. ICSID arbitration can proceed even absent host country participation, and any resulting award can be legally enforced against host-country assets in domestic courts the world over. This standard story is reasonably well and good as far as it goes. Investors certainly aren’t clamoring against BITs, and it is quite reasonable to assume that, given the informed choice, the investor would rather invest protected by a BIT than otherwise. But it is the size of the apparent effect of BITs on investor confidence that Neumayer and Spess report that should raise concerns. It is theoretically implausible that the treaties, even if undeniably favorable to investors, would cause such dramatic changes in investment flows. This is so for three main reasons. First, potential investors seem to have little awareness or appreciation of specific BITs. The argument that BITs “cause” FDI typically assumes or implies that investors are aware of and decisively take into account the presence or absence of particular treaties at the preinvestment stage of the capital-allocation process. But there is extremely little qualitative evidence that foreign investors actually pay much attention to BITs when deciding where to invest. Indeed, there is suggestive evidence that historically investors have not paid much attention at all to the treaties or to international legal protections for their investments more generally. For example, a small survey of business executives conducted in 1976 found that only 16% of respondents were “familiar” with ICSID (Ryans and Baker 1976). Only 4 percent felt that ICSID arbitration provided “adequate safeguards.” These results led the authors to conclude that ICSID needed to mount a major promotional campaign. It is highly unlikely that investor awareness or appreciation of specific BITs was (or is) any higher. Perhaps even more revealing is the title of a recent practitioner-oriented publication, “Arbitration under Bilateral Investment Treaties: An often overlooked tool,” which suggests that additional promotional efforts may still be needed (Freyer et al. 1998), or a 2005 client memo by the white-shoe international law firm of Allen and Overy that referred to BITs as “these little-known treaties [that] may give you . . . rights you never knew you had” (Allen and Overy 2005). This lack of awareness should not surprise us, as multinational corporations have historically only haphazardly and imperfectly institutionalized more general “political risk assessment” procedures (Kobrin 1982). If investors have little awareness of BITs at the preinvestment stage, or alternatively, if they do have some limited awareness but only imperfectly integrate that awareness into their existing (and also imperfect) political risk assessment procedures, it is unlikely that the presence or absence of a treaty will decisively and directly affect the investment decision. Second, BITs are not necessary to resolve problems of credible commitment. Claims that BITs should be (and are) especially effective at inducing FDI implicitly and wrongly assume that BITs are the only potentially effective and actually used

382 jason yackee

legal means of credibly committing to treat investors favorably. In fact, investors have long preferred to secure their interests through investment contracts, which give the investor the opportunity to obtain deal-specific promises that are much more detailed and precise than the ambiguous one-size-fits-all promises typically provided in BITs (Fatouros 1962). Sophisticated investment contracts often include references to international law, they guarantee legal stability (e.g., that the host country won’t change relevant regulations in unfavorable ways postinvestment), and often contain enforceable international arbitration clauses. Such contracts are especially common, if not universal, in the natural resources and infrastructure-concession sectors. These are precisely the sectors in which the risk of bargain “obsolescence” is said to be greatest. Even high-tech manufacturing companies, like Intel, routinely insist on signing a written investment contract with a host country prior to making an important investment (Spar 1998). Other BIT alternatives include international investment insurance, usually provided and subsidized by the investor’s home government and which, unlike BITs, guarantees a full monetary recovery in the case of a breach of international law. The domestic laws of the host country itself can also be used to provide foreign investors with very favorable and very specific substantive terms of treatment, and even to provide binding consents to investor-initiated arbitration to safeguard those guarantees.1 There is very little reason to expect that BITs are any more effective at resolving the credible commitment problem than are investment contracts or investment insurance or well-crafted municipal investment laws. BITs add very little to what was already available to legally secure risky foreign investments. The host country’s decision to enter into a BIT tells us nothing about its use of other credible commitment devices. If BITs merely duplicate or supplant investor reliance on contracts, investment insurance, or favorable municipal legal promises, there is little reason to expect the treaties to induce massive new FDI inflows. Third, the “credible commitment” risk premium is objectively low. Scholarly estimates suggest that what might be called the “risk premium” caused by problems of credible commitment is not terribly high. Guasch (2004, p. 8), for instance, suggests that the risk of regulatory change for large-scale infrastructure projects— the kinds of projects that are likely to be most susceptible to problems of credible commitment, given their very high degree of asset specificity and the need to make the majority of the investments at the front end of the project lifespan— can add just “2–6 percentage points to the cost of capital,” “depending on the country and sector.” It is likely that the costs of political risk in more mobile sectors, like manufacturing, are far lower. No one argues that BITs can fully 1. For example, a recent survey has found that approximately 20 national foreign investment laws contain “generic consent provisions offering to submit disputes with investors to arbitration under the ICSID Convention” (Lovells 1999). See also Shihata and Parra (1999).

do bit s really work? revisiting the empirical link between investment 383

eliminate such risks or their associated costs. Assuming quite generously that BITs are uniquely able to reduce half of the risk, the cost-reducing benefit of BITs for the riskiest investment projects is just 1 to 3 percent—hardly enough to suggest entering into BITs will spur dramatic increases in new investment. And of course, infrastructure concessions are always covered by detailed investment contracts, making BITs largely redundant. Why might the costs of political risk tend to be so low, even for high-risk projects? Because in most cases—for example, those cases where a host country desires additional foreign investment in the future—host country reputational concerns will provide a very strong disincentive for the host country to treat foreign investors unfairly. Obsolescing bargaining theory, which underlies BIT optimism, unjustifiably discounts the risk-reducing role of reputational concerns by analyzing host country-foreign investor relations in the context of a single investment relationship (cf. Janeba 2002). In fact, when deciding where to invest foreign investors typically pay particular attention to the experiences of past and existing investors (Spar 1998, pp. 14–15). Unfavorable host country behavior is likely to have strong ripple effects beyond the investment immediately affected, as other existing investors withdraw from the host country, and as potential investors redraw their investment plans. Host countries certainly recognize this potentiality, and that recognition is undoubtedly behind the dramatic decline in instances of classical expropriation and nationalization over the past three decades (Minor 1994). Developing countries have almost universally decided that foreign investment should play an important long-term role in their economies. That decision powerfully, if informally, serves to significantly reduce political risk, leaving precious little risk for BITs themselves to further reduce. B. The Replication Model Neumayer and Spess present results for a number of generally similar model specifications. I focus here on replicating their Table 4 Model II, the results of which are representative of their other models. Their model takes the following basic form: FDI Sharei,t = BITsi,t + Riski,t +BITsi,t∗Riski,t + economic control variables,

where FDI Share is a logged measure of FDI inflows to developing country i at time t as a proportion of FDI inflows to all developing countries, BITs is a weighted count of the number of BITs that a given country has signed with the major capital-exporting countries (with the weight equal to the particular capital exporting country’s share of world FDI outflows),2 Risk is a 100-point composite 2. Like Neumayer and Spess, and with the exception of the model presented in Figure 3, I construct the BIT variable using dates of signature, without regard for whether a particular treaty ever actually enters into legal force. Using more theoretically appropriate dates of entry into force systematically worsens the results presented in Figures 1 and 2.

384 jason yackee table 1. fixed-effects (within) estimation results, natural log fdi share Independent variable

I. N&S results

II. Replication results

BITs (weighted count) ......................................... ICRG composite political risk .............................

0.035 (4.04)∗∗ 0.013 (2.00)∗ –0.0003 (2.04)∗ 1.916 (3.13)∗∗ –2.686 (5.13)∗∗ 2.268 (3.18)∗∗ –0.0001 (2.33)∗ 0.032 (2.70)∗∗ 0.216 (1.29) –0.049 (0.31) 1,346 91 1985–2001 0.07

–0.391 (2.64)∗∗ –0.005 (5.72)∗∗ 0.011 (4.68)∗∗ 0.421 (6.19)∗∗ –0.092 (1.98)∗ 0.222 (2.01)∗ –0.00001 (1.18)

BITs ∗ ICRG composite political risk .................. ln GDP p.c............................................................ ln population ....................................................... Economic growth ................................................ Inflation................................................................ Resource rents ..................................................... FTA count [Trade openness] ............................... WTO membership ............................................... Observations ....................................................... Countries ............................................................. Period ................................................................... R-squared (within) ...............................................

–0.002 (2.35)∗ [0.0003 (1.49)] –0.002 (0.11) 1,506 101 1985–2003 0.21

Notes: Absolute t-values in parentheses. Standard errors are robust. ∗indicates significance at the 0.05 level; ∗∗indicates significance at the 0.01 level. All variables are lagged one period.

measure of political risk provided by the PRS Group, and BIT∗Risk is a multiplicative product of the first two variables. Neumayer and Spess’s theory of BITs is conditional. They argue that if BITs work by reducing political risk, the effectiveness of the treaties at inducing new FDI flows should be lower for host countries that are already low-risk investment prospects. The product term models this intuition. Table 1, immediately below, compares Neumayer and Spess’s reported results with the results of my replication. My replication model is substantially identical to Neumayer and Spess’s, though it is not perfectly duplicative. Again, the point here is to examine the stability of Neumayer and Spess’s results “using different data sets” and “different measures” (Buena de Mesquita 2003). If their results do not hold up in the face of these very modest and justifiable extensions and changes, then we can and should have little confidence that they have isolated and identified a truly or universally causal relationship. Differences are five-fold.

do bit s really work? revisiting the empirical link between investment 385

First, I have extended the time period of the study from 2001 to 2003.3 Second, my BIT variable is a weighted count of BITs signed by capital-importing countries with the top 18 capital-exporting countries. I identify the 18 top capitalexporting countries by tabulating shares of world FDI outflows over the period of study.4 “Capital-importing countries” are those countries not among the top 18 capital-exporting countries. Neumayer and Spess’s BIT variable is a weighted count of signed BITs between “developing countries” and members of the OECD, a construction that assumes that OECD membership is a decent proxy for “capital-exporting.” Both methods produce a largely overlapping set of capital-exporting countries. Third, my BIT variable “counts” as BITs a number of broader commercial treaties that contain embedded BIT-equivalent investment-related chapters or provisions.5 Neumayer and Spess rely on a list of BITs published by the United Nations Conference on Trade and Development (UNCTAD) to construct their BIT variable. However, UNCTAD’s list includes only stand-alone investment treaties and ignores an earlier generation of so-called “friendship, commerce, and navigation” (FCNs) treaties (many of which remain in force today), and modern free trade agreements (FTAs), such as the North American Free Trade Agreement, that include important investment-related guarantees. I have supplemented UNCTAD’s list of BITs by counting FCNs and FTAs as “BITs” where the investment-related provisions substantively approximate a typical BIT.6 Fourth, I have replaced Neumayer and Spess’s separate count of FTAs with a standard measure of “trade openness,” exports plus imports divided by GDP, because I have already accounted for BIT-equivalent FTA agreements directly in the BIT variable. Neumayer and Spess claim that their FTA count variable is intended to measure “trade openness,” and they report in a footnote that they included an actual measure of trade openness, identical to my own, with no substantive impact on their reported results.

3. Neumayer and Spess report their sample as running from 1984 to 2001. Given the availability of the ICRG data and that they lag their independent variables, their study probably runs from 1985–2001. 4. The top 18 capital-exporting countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, Netherlands, Norway, Singapore, Spain, Sweden, Switzerland, the United Kingdom, and the United States. These countries have supplied between approximately 84% and 99% of annual world FDI flows over the past 30-some years. 5. I present a fuller critique of the use of UNCTAD’s list of BITs in Yackee (2008). 6. I (improperly) ignore here, as do Neumayer and Spess, the Energy Charter Treaty (a multilateral BIT between more than fifty mostly European countries), and international treaties that offer BIT-like promises to investors, such as applications to join the European Union, the European Convention on Human Rights, or various OECD investment “declarations.”

386 jason yackee

Finally, I added a constant “start” to all observations of the dependent variable (FDI Share) prior to logging it. In contrast, Neumayer and Spess transform negative values of FDI Share by recoding only the negative values as a positive number just greater than zero, a methodologically inappropriate means of dealing with undefined logs because it shifts coefficient estimates. Like Neumayer and Spess, I estimate the model using a standard fixed effects (within estimator) approach and robust standard errors and do not include year dummies. A table with summary statistics and data sources is provided in the Appendix. As to the control variables, we can briefly note that results are generally comparable in terms of significance and direction of effect. We are more interested in the performance of the BIT and risk variables, and their multiplicative interaction. These key explanatory variables are all statistically significant in the replication model (as they were in Neumayer and Spess’s original model). Note, however, that the results for the main explanatory variables—the BIT count, the International Country Risk Guide (ICRG) rating, and the interaction term— are statistically significant but oppositely signed from what Neumayer and Spess report. The substantive significance of this difference is not immediately obvious from the face of the table, as interpreting regression coefficients in the presence of a multiplicative interaction term poses certain subtleties (Braumoeller 2004; Brambor et al. 2005). It is far more meaningful to calculate the significance and magnitude of the marginal effects of changes of the value of one component of the interaction term on the dependent variable across various values of the other component. I do this in Figure 1, following the approach described by Brambor et al. (2005), and which illustrates the marginal effects of signing additional BITs on FDI share at varying levels of political risk. The marginal effects of the BIT variable are displayed along the y-axis, while the observed levels of political risk are displayed across the x-axis. The diagonal solid line represents the point estimate of the marginal effects, while the dotted lines around the marginal effects line illustrate the 95% confidence interval of the point estimation. The solid horizontal line is the x-axis at zero (“the zero line”). Where both the upper and lower bounds of the confidence interval are positive (above the zero line) or negative (below the zero line) the effect of BITs on FDI inflows is statistically significant in the direction indicated by the point estimate. Where the confidence interval straddles the zero line, we cannot reliably say whether or not BITs have positive or negative effects on FDI. Figure 1 illustrates how profoundly different the replication results are from those reported by Neumayer and Spess. Neumayer and Spess’s conditional theory of BITs predicts that BITs will decline in effectiveness as political risk decreases (and thus as values on the ICRG composite increase). They do not provide marginal effects figures, but they report that their results support their conditional theory. We would expect point estimates that decline as we move from left to right on the Figure. Instead, the replication suggests an opposite conditional

do bit s really work? revisiting the empirical link between investment 387

figure 1. marginal effect of signed bits on fdi inflows as political risk changes Marginal Effect of BITs on FDI Inflows

Dependent Variable: Log Share of LDC FDI Inflows 1 .5

0 −.5

−1 0

10

20

30

40

50

60

70

80

90

100

ICRG Composite (Higher Rating Means Less Risk) Marginal Effect of BITs 95% Confidence Interval Fixed Effects(Within Estimator), Robust SEs

relationship—BITs seem to become more effective as political risk decreases. This is illustrated by the upward-sloping point estimate line. Just as problematically, note that for high levels of political risk (from approximately 16 on the ICRG index to 45) the confidence intervals span the zero-line. This means that at these levels of political risk—which correspond to approximately 20% of sample observations—we are simply unable to say that BITs have any statistically significant marginal effect on FDI share.7 On the other hand, at the highest observed ICRG ratings (and thus at the lowest levels of political risk), it does appear that BITs have relatively large and statistically significant marginal effects on FDI share. For example, the highest observed ICRG rating is 91.5. The predicted marginal effect of the BIT variable at that level of risk is approximately 0.5. Substantively, this means that a one standard deviation increase in the BIT variable (itself roughly equivalent to signing a BIT with the United States) is predicted to increase FDI share by a factor of 1.14.

7. Results are worse in this regard if we cluster the standard errors by country, as we probably should—doing so makes the results statistically insignificant up to an ICRG risk rating of 55. Neumayer and Spess do not cluster in their analyses. On the desirability of clustering, see Williams 2000; Cameron et al. forthcoming; Bertrand et al. 2004; Kédzi 2004.

388 jason yackee figure 2. marginal effect of signed bits on fdi share, ols-pcse-ldv Marginal Effect of BITs on FDI Share

Dependent Variable: Log Share of LDC FDI Inflows

.2 .1 0 −.1 −.2 0

10

20

30

40

50

60

70

80

90

100

ICRG Composite (Higher Rating Means Less Risk) Marginal Effect of BITs 95% Confidence Interval OLS with Panel-Corrected SEs, Lagged DV, Country Fixed Effects

C. Sensitivity to Panel-Corrected Standard Errors With Lagged DV In what has become one of the most-cited methodological articles in the field of international political economy, Beck and Katz (1995) recommend estimating panel models using OLS and panel-corrected standard errors (PCSE). Elsewhere, Beck and Katz (1996) recommend including in the OLS-PCSE model a lagged dependent variable (LDV), which serves to capture dynamic tendencies in the causal relationship. Keele and Kelly (2006) have recently shown that this latter advice is usually well-taken. OLS-PCSE-LDV estimation has become very common in political science (Beck 2001), and PCSEs are especially useful in dealing with panel-level heteroskedasticity, which a Wald test suggests that we have here.8 Reestimating the replication model using a Beck and Katz estimation strategy substantially worsens results. As Figure 2 shows, the point estimates of the effects of BITs on FDI are mostly positive, but they are never statistically significant. (A traditional regression table of the Figure 2 results is presented in the 8. I include country-level fixed effects in the OLS-PCSE models, as recommended by Wilson and Butler (2007). Modeling the dynamic nature of FDI inflows as a first-order autoregressive process rather than through the inclusion of an LDV (see id.) produces results approximate to those presented in Figure 1.

do bit s really work? revisiting the empirical link between investment 389

Appendix). What is significant—and massively so, with a z-score of over 10—is the lagged dependent variable. This indicates that there is a strong dynamic relationship between present and past values of a country’s FDI share. The significance of the LDV should not be surprising. As Keele and Kelly (2006, 31) suggest, “[t]he preponderance of the evidence in both economics and political science is that many if not most cross-temporal processes are dynamic,” and there are strong reasons to believe that the processes that generate FDI are dynamic as well. For example, it is well recognized that foreign investment decisions often entail a “follow-the-leader” dynamic, in which an investment by one company spurs investments by others (Spar 1998, pp. 9, 14–15; Pennings 2005). Not surprisingly, previous empirical studies of the policy determinants of FDI inflows tend to control for the effects of past levels of inflows (Gastanga et al. 1998). Given strong theory and evidence that, in the present context, the past indeed matters, models of the effects of BITs on FDI that fail to control for past values of the dependent variable are likely to suffer from serious omitted variable bias. (In this regard, it is worth noting that simply adding an LDV to the basic replication model reported in Figure 1 renders the results statistically insignificant for nearly all observed values of risk). D. Sensitivity to the Exclusion of “Weak” BITs In the standard theoretical story of BITs and FDI, what makes BITs such potentially effective “credible commitment” devices is guaranteed investor access to international arbitration to resolve a wide range of potential investment disputes. One major problem for empirical BIT analysts is that not all BITs contain these binding arbitration provisions. A fair number of weak BITs have been signed (and many have entered into and remain in force) that offer investors no such access, or which offer it only for a very narrow class of potential investment disputes. If BITs theoretically matter because of their arbitration provisions, than we should be more likely to see a statistical relationship between FDI and BITs if we remove the weak (and thus theoretically irrelevant) treaties from our sample.9 Figure 3 illustrates the reestimation of the basic replication model of Figure 1 with a revised count of BITs that includes only strong treaties in the relevant independent variable. I count a BIT as “strong” if it contains the host country’s binding, enforceable preconsent to arbitrate a wide variety of investment

9. As Wälde (2005, p. 194), a distinguished practicing international lawyer and academic scholar of BITs notes, It is the ability to access a tribunal outside the sway of the Host State which is the principal advantage of a modern investment treaty. This advantage is much more significant than the applicability to the dispute of substantive international law rules. The remedy trumps in terms of practical effectiveness the definition of the right.

390 jason yackee

Marginal Effect of Strong BITs on FDI Share

figure 3. marginal effect of strong bits in force on fdi share as political risk changes Dependent Variable: Logged Share of LDC FDI Inflows .5 .25 0 −.25 −.5 −.75 −1 0

10

20

30

40

50

60

70

80

90

100

ICRG Composite (Higher Rating Means Less Risk) Marginal Effect of Strong BITs 95% Confidence Interval Fixed Effects (Within Estimator) and Robust Standard Errors

disputes at the investor’s initiative.10 If BITs have a consistently positive and statistically significant effect on FDI, we should be most likely to see it with this group of treaties. Instead, we see a variation on the same disappointing theme. With the “within” estimator, strong BITs are statistically significantly associated with greater FDI share only at the very lowest levels of political risk, corresponding to an ICRG score of over 80 and applying to only a small handful of the countries in our sample, and are significantly associated with lower FDI shares for the highest-risk countries. This latter result, in particular, is clearly counterintuitive, and indeed, if we cluster the fixed effects model by country, the Figure 3

10. The strong BIT variable includes only BITs that have entered into force. The model includes a “weak BIT” variable and its interaction term as separate controls, with the “weak BIT” variable constructed as a weighted count of all non-“strong” BITs. The methodology for classifying BITs as “strong” or “weak” is described in more detail in Yackee (2008). The coding exercise is only possible for treaties that have entered into force, as the texts of treaties that have only been signed (but which have not entered into force) are not widely available.

do bit s really work? revisiting the empirical link between investment 391

results become insignificant across all ranges of risk. It is perhaps not surprising, then, that when we use the Beck and Katz metholodgy the results are also uniformly nonsignificant.

conclusion Neumayer and Spess’s work on the empirical relationship between BITs and FDI is professional, nuanced, thought-provoking, and eminently worthy of emulation. Where they mainly err is in so strongly claiming that their work allows us to “know” that BITs work. The analysis above suggests that we actually know very little. Using a largely similar dataset and a slightly larger sample of years, I found no evidence that BITs and FDI share the kind of conditional relationship theorized (and identified) by Neumayer and Spess. Indeed, my basic replication results suggest an opposite conditional relationship, where BITs are statistically significant predictors of FDI share only for low-risk countries, and where the magnitude of that effect increases as risk decreases. The problem is that there is very little theoretical reason to expect BITs to only be effective in low-risk situations, or to become more effective as risk decreases. Perhaps it is not surprising, then, that the results prove sensitive to modeling choices, such as whether to estimate the model using OLS-PCSE with an LDV, or to control for differences in dispute settlement provisions. This latter nonfinding is highly revealing, as a positive relationship between BITs and FDI should be most evident in the case of strong treaties. All of this should not be taken to be an argument that we now “know” that BITs don’t meaningfully influence foreign investment decisions. I suspect that the presence or absence of a BIT is rarely, if ever, a particularly salient issue in most investment decisions, for reasons already discussed above. But I think that if we really want to prove that BITs do or do not matter, that they do or do not work as advertised, then we may want to consider whether large-n statistical studies of aggregate FDI flows are the best means of empirically addressing the question. One potentially promising solution is to return to less modern methodologies, in particular the in-depth case study or survey instruments. We currently have no direct evidence of the role that international legal considerations actually play in the investment decisions of major corporations. Are executives with decision-making power aware of BITs prior to making investment decisions (and not just after a dispute has arisen)? If so, do they view the treaties as salient in deciding whether to make an investment? These are essentially the questions that Neumayer and Spess mean to ask. I suggest in closing, but admittedly without doing any of the hard work required, that the answers would probably be more forthcoming if we simply asked—in a scientifically appropriate way, of course—those best positioned to know.

392 jason yackee table a1. summary statistics, basic replication model with data sources Variable

Obs.

Mean

Std. Dev. Min

Max

Log FDI share1 ...................................... BITs (weighted count)2......................... ICRG composite political risk3 ............. BITs ∗ ICRG composite political risk .......................................... Log GDP p.c1......................................... Log population1 .................................... Growth1 ................................................. Inflation1 ................................................ Resource rents4..................................... WTO Member5 ...................................... Trade Openness1 ..................................

1,506 1,506 1,506 1,506 – 1,506 1,506 1,506 1,506 1,506 1,506 1,506

–4.958 0.330 58.956 19.991 – 7.177 16.202 0.033 74.108 5.889 0.343 67.680

0.323 0.266 13.628 17.996 – 1.321 1.564 0.047 593.948 10.268 0.475 36.835

–2.748 0.912 91.5 72.522 – 10.318 20.964 0.259 13611.63 63.608 1 291.032

–7.285 0 14.25 0 – 4.533 12.361 –0.275 –29.173 –5.7E-05 0 0

Data sources: 1World Bank World Development Indicators 2006; 2[Present Author] 3 PRS Group International Country Risk Guide; 4World Bank Environment Department website; 5WTO website.

table a2. ols-pcse-ldv estimation results for figure 4, natural log fdi share Independent variable

Model results

BITs (weighted count) .................................................... ICRG composite political risk ....................................... BITs ∗ ICRG composite political risk ............................. ln GDP p.c....................................................................... ln population .................................................................. Economic growth............................................................ Inflation........................................................................... Resource rents ................................................................ Trade openness .............................................................. WTO membership .......................................................... Lagged DV (log FDI share) ............................................ Observations .................................................................. Countries ........................................................................ Period .............................................................................. R-squared ........................................................................

–0.035 (0.34) –0.001 (1.16) 0.002 (1.05) 0.089 (2.33)∗ 0.026 (0.81) 0.146 (2.12)∗ –1E-6 (0.25) 9E-5 (0.11) 2E-5 (0.13) –0.003 (0.34) 0.733 (10.83)∗∗ 1,506 101 1985-2003 0.86

Notes: Absolute z-values in parentheses. Model contains country dummies. ∗indicates significance at the 0.05 level; ∗∗indicates significance at the 0.01 level. All variables are lagged one period.

do bit s really work? revisiting the empirical link between investment 393

references Allen and Overy, LLP. (2005). “In focus: the rise of investment treaty arbitration,” November 7. Beck, Nathaniel (2001). “Time-series cross-section data: what have we learned in the past few years?” Annual Review of Political Science, 4, pp. 271–293. Beck, Nathaniel and Jonathan N. Katz (1995). “What to do (and not to do) with time-series cross-section data,” American Political Science Review, 89, pp. 634–647. Beck, Nathaniel and Jonathan N. Katz (1996). “Nuisance vs. substance: specifying and estimating time-series-cross-section models,” Political Analysis, 6, 1, pp. 1–36. Bertrand, Marianne et al. (2004). “How much should we trust differences-in-differences estimates?” Quarterly Journal of Economics, 199, 1, pp. 249–275. Brambor, Thomas et al. “Understanding interaction models: improving empirical analysis,” Political Analysis 14, 1, pp. 63–82. Braumoeller, Bear F. (2004). “Hypothesis testing and multiplicative interaction terms,” International Organization, 58, pp. 807–820. Bueno de Mesquita, Bruce (2003). “Getting firm on replication,” International Studies Perspectives 4, pp. 98–100. Cameron, Colin A. et al. “Bootstrap-based improvements for inference with clustered errors,” Review of Economics and Statistics, forthcoming. Elkins, Zachary et al. (2006). “Competing for capital: the diffusion of bilateral investment treaties, 1960–2000,” International Organization, 60, pp. 811–846. Fatouros, A. (1962). Government Guarantees to Foreign Investors. (New York: Columbia University). Freyer, D.H. et al. (1998). “Arbitration under bilateral investment treaties: an often overlooked Tool,” MEALEY’S International Arbitration Report. Gastanga, Victor M. et al. (1998). “Host country reforms and FDI inflows: how much difference do they make,” World Development, 26, pp. 1299–1314. Guasch, J. Luis. (2004). Granting and Renegotiating Infrastructure Concessions: Doing it Right (Washington, D.C.: World Bank). Guisinger, Stephen E. et al. (1985). Investment Incentives and Performance Requirements (New York: Praeger). Henisz, Witold Jerzy (2002). Politics and International Investment: Measuring Risks and Protecting Profits. (Northampton, MA: Edward Elgar). Janeba, Eckhard (2002). “Attracting FDI in a politically risky world,” International Economic Review, 43, 4, pp. 1127–1155. Keele, Luke & Nathan J. Kelly (2006). “Dynamic models for dynamic theories: the ins and outs of lagged dependent variables,” Political Analysis, 14, 2, pp. 186–205. Kézdi, Gabor (2004). “Robust standard error estimation in fixed-effects models,” Working Paper available at http://ideas.repec.org/e/pke76.html (2004). Kobrin, Stephen J. (1982) Managing Political Risk Assessment: Strategic Response to Environmental Change (Berkeley: University of California Press). Lovells LLP. (2005) “Client note: protecting investments overseas: bilateral investment treaties, foreign investment laws, and ICSID arbitration,” (July). Minor, Michael S. (1994). “The demise of expropriation as an instrument of LDC policy, 1980– 1992,” Journal of International Business Studies, 25, pp. 177–188. Neumayer, Eric and Laura Spess (2005). “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33, 10, pp. 1567–1585.

394 jason yackee North, Douglas C. (1990). Institutions, Institutional Change and Economic Performance (New York: Cambridge University Press). Parra, Antonio and Ibrahim F.I. Shihata (1999). “The experience of the international centre for settlement of investment disputes,” ICSID Rev.-Foreign Invest. L.J. 14, pp. 299–361. Pennings, Enrico (2005). “Learning from foreign investment by rival firms: theory and evidence,” Presented at the WZB Economics & Politics Seminar Series at Berlin, September 26. Ryans, John K. and James C. Baker (1976). “The international center for settlement of investment disputes (ICSID),” Journal of World Trade Law 10, 1, pp. 65–79. Sornarajah, M. (2004). The International Law of Foreign Investment (Cambridge: Cambridge University Press). Spar, Debora. (1998). “Attracting high technology investment: Intel’s Costa Rican plant,” Foreign Investment Advisory Service Occasional Paper No. 11, (April). Vernon, Raymond (1971). Sovereignty at Bay: The Multinational Spread of U.S. Enterprises (New York: Basic Books). Wälde, Thomas (2005). “The ‘Umbrella’ clause in investment arbitration: a comment on original intentions and recent cases,” Journal of World Investment & Trade, 6, pp. 183–236. Williams, Rick L. (2000). “A note on robust variance estimation for cluster-correlated data,” Biometrics, 56, pp. 645–646. Williamson, Oliver E. (1996). The Mechanisms of Governance. (New York: Oxford University Press). Wilson, Steven E. and Daniel M. Butler (2007). “A lot more to do: the sensitivity of time-series cross-section analyses to simple alternative specifications,” Political Analysis, 15, 2, pp. 101–123. Yackee, Jason Webb (2008). “Conceptual problems in the empirical study of bilateral investment treaties,” Brooklyn Journal of International Law, 33, 405–462.

15. bilateral investment treaties and foreign direct investment: correlation versus causation∗ emma aisbett introduction In the last fifteen years of the 20th century, foreign direct investment (FDI) growth outstripped all other global economic measures, with no fewer than ninety-four countries experiencing FDI growth rates in excess of 20% per year (UNCTAD 2001). At the same time, there was a broadening consensus regarding the benefits of FDI to host countries, particularly less industrialized ones. As a result, more than 1,100 national policy changes in favor of FDI were introduced worldwide between 1991 and 2000 (UNCTAD 2001). The rapid and concurrent expansion of FDI and policies to attract FDI make the question of causality between the two both interesting and challenging. This paper addresses this question for bilateral investment treaties (BITs). Specifically, I test whether participation in BITs leads to increased FDI inflows from the treaty partner countries. By the end of 2005, nearly 2,500 BITs had been signed, most of them after 1990 (UNCTAD 2006a). The stated intent of the treaties is the “reciprocal promotion and protection of investment” between signatory governments. Among other things, BITs specify rights to invest in accordance with the laws of the host, rights to freely transfer funds and assets minimum treatment standards, and protection from expropriation. The most economically important aspect of BITs, however, is the direct investor-state dispute mechanisms that allow investors to bring claims of treaty violations to arbitration tribunals outside of the host country. For example, if a host raises taxes levied on a foreign firm’s profits above levels agreed at the time of investment, the foreign investor may be able to take an expropriation claim to arbitration under the BIT. The incorporation of an international dispute resolution mechanism distinguishes BITs from domestic policy statements and makes them a potentially effective commitment device.1

∗ The author would like to thank Ann Harrison and Larry Karp for their excellent advice and support. This chapter has also benefited enormously from the comments and suggestions of Carol McAusland, Jenny Lanjouw, Ted Miguel, Jeff Perloff, Suzanne Scotchmer, and Brian Wright, as well as numerous seminar participants. 1. I discuss the evidence on the limits to the effectiveness of BITs are as a commitment device in Section A.

396 emma aisbett

Existing studies have found contradictory evidence regarding the impact of BITs on FDI. (Hallward-Driemeier 2003) and Tobin and (Rose-Ackerman 2004) find either no impact or a negative impact of BITs, while (Neumayer and Spess 2005) and (Salacuse and Sullivan 2004) find strong positive impacts. Neumayer and Spess attribute the different conclusions of the various studies to differences in sample, and in the case of (Hallward-Driemeier 2003), the inability of her methodology to capture signaling effects. I explain the different findings of these studies in terms of methodological issues, many of which are common to the broader literature on the impacts of FDI and trade promotion policies. I illustrate these general empirical issues using panel data of FDI outflows from Organisation for Economic Co-operation and Development (OECD) countries to twenty- eight low-and middle-income countries. I suggest several simple specification improvements to address these issues, and show that the omission of any one of these improvements (as is usually the case in the existing empirical FDI literature) can lead to serious errors in inference. One empirical issue turns out to be less of a concern than expected. I find that selection bias, which has received some attention in the recent trade and investment literature, is effectively eliminated by the inclusion of country-pair fixed effects in the specification (Razin, Rubinstein and Sadka, 2003), (Helpman, Melitz and Rubinstein, 2005), (Razin, Sadka and Tong, 2005). Although accounting for data-related empirical issues is important, the primary problem for researchers wishing to assess the impacts of policies to promote FDI is that policy adoption is endogenously determined. In the case of BITs, there is potential endogeneity due to both reverse causality and omitted variables. For example, increased FDI flows in one year may cause a BIT to be signed in the next, or an improvement in the investment climate of the host may cause a simultaneous increase in both FDI and BIT participation. I show the potential for both these forms of endogeneity by modeling a simple game between a host government deciding whether to participate in a BIT and a representative foreign firm deciding whether to invest in the host. The starting point for my empirical analysis of the impact of BITs is to test the robustness of the BIT indicator to the set of specification improvements discussed earlier. I find that the BIT indicator is positive and significant, even in my preferred (most conservative) specification. Robust positive correlation between BITs and FDI is one of the empirical predictions of my model; however, this prediction is based on a combination of the causal effect of BITs and the endogeneity of BIT formation. Thus, it would be a mistake to conclude from these results that BITs have a significant positive impact on bilateral FDI flows. Further evidence that the observed correlation between BITs and FDI flows is not predominantly attributable to a causal effect of BITs is provided by the magnitude of the BIT coefficient. The point estimate implies that BIT participation is associated with a greater than 50% increase in bilateral FDI flow. Not even the most enthusiastic proponent of BITs would feel comfortable attributing such an increase to the causal impact of BITs.

bilateral investment treaties and foreign direct investment 397

The standard approach to deal with the endogeneity of right-hand side variables is to use an instrumental variable. Unfortunately, a suitable instrument for BITs is not available; however, BITs themselves are exogenous ex post, that is, once in place, a BIT remains in place for a minimum of ten years. This characteristic of BITs means that it is possible to overcome the lack of an instrument if we can satisfactorily control for BIT adoption.2 The use of bilateral data provides a great deal of potential in this regard which has not been exploited by previous studies. I find that when I include either proxies for the underlying growth rate of bilateral FDI between countries3 or host- and source-year effects,4 the magnitude of the BIT coefficient drops and becomes statistically insignificant. These findings suggest that the strong correlation previously identified between BITs and FDI is substantially caused by the endogeneity of BIT adoption. There are two issues that need to be addressed before concluding from these results that BITs have no significant impact on FDI flows. The first is that the controls used to reduce the endogeneity bias may also have disguised a signaling effect of BITs. Signaling a safe investment climate has been suggested by other authors as potentially the primary function of BITs (Hallward-Driemeier 2003), (Neumayer and Spess 2005). The use of bilateral data allows me to explicitly test for a signaling effect by replacing the BIT dummy in my base specification with the number of BITs that the host has signed with other Organization for Economic Cooperation and Development (OECD) countries. If BITs have a signaling effect, participation by the host in BITs with other OECD countries should lead to an increase in FDI received. I find no evidence of such an effect. A second potential concern is that my efforts to control for endogeneity may have left too little variation in the data with which to identify the effects of BIT participation. It is not possible to completely rule this out; however, it is possible to show that much of the correlation identified in the base specification is due to the endogeneity of BITs. To do this, I use a graphical event study analysis to show that BITs are signed when FDI flows are already increasing. I also use Granger-type analysis of the relationship between BITs and expropriation risk to show that BITs are ratified when expropriation risk falls, but the ratification of BITs does not lead to further decreases in expropriation risk. The rest of this paper is organized as follows. I begin in Section A with a short overview of the basic theory and case evidence on the potential of BITs to function as a commitment device for the host. Section B presents a theoretical model of BIT function and the decision of a host country to participate in one. Section C

2. This approach is common in the program evaluation literature, where the identifying assumption is often referred to as the “ignorability of treatment” following Rosenbaum and Rubin (1983). 3. Specifically, I include individual linear time trends for each source-host country-pair. 4. Host-year effects are a set of dummies for each host in all but one year. Source-year effects are analogously defined.

398 emma aisbett

provides a summary of the data and general specification issues, which are addressed in detail in Appendix A. Section C also introduces and motivates my empirical approach to the endogeneity of BIT participation and shows that selection bias is not a concern in my specifications. Section D presents the results of the regression analysis, followed by the graphical event study of BIT participation. The final subsection of results shows that BITs do not attract investment from nonpartner countries through signaling. Finally, I conclude and suggest directions for future research. A. BITs in practice The primary economic function of BITs is to act as a commitment device for the host government. BITs are designed to achieve this function through direct investor-state dispute resolution mechanisms which are constituted outside of the host country. If an investor from the partner country feels that the host government has violated their rights under the BIT, for example by expropriating the investment, they may bring a compensation claim to an international tribunal for arbitration. Most arbitration cases are run by either the International Center for the Settlement of Investment Disputes (ICSID), which is a part of the World Bank Group, or under the United Nations Commission on International Trade Laws (UNCITRAL) Arbitration Rules. As of 2005, there have been 225 known arbitration cases brought by foreign investors against host governments under an international investment treaty (UNCTAD 2006b). Examples of host actions that have resulted in the investor bringing a compensation case under a BIT include the removal of tax breaks, failure to increase tariffs paid to the investor as agreed in contract, expropriation of land for incorporation into a national park, and denial of license renewal for a hazardous waste landfill.5 The need for an externally supported commitment device is motivated by the presence of sunk costs of investment which can lead to dynamic inconsistency of optimal policy for the host. Before the investor makes the investment, the host’s optimal policy is to promise good conditions such as low taxes. After the investment takes place and costs are sunk, the optimal policy for the host is to extract rents up to the value of the sunk costs, that is, to directly or indirectly expropriate the investment. The result is a classic holdup problem leading to underinvestment. BITs can solve the problem because they provide extranational arbitration of investor compensation claims and thereby help the host to credibly commit not to change its policy toward the investment. Although the basic commitment device view leads to clear predictions regarding the investment promoting impacts of BITs, there are a number of reasons that we may not observe the effect empirically. To begin with, host governments 5. All these cases can be found on the ISCID Web site, http://www.worldbank.org/ icsid/cases/awards.htm#awardARB011. The case numbers for the examples are in order ARB/95/3, ARB/97/3, ARB/96/1, and ARB(AF)00/2.

bilateral investment treaties and foreign direct investment 399

do not necessarily have a commitment problem in the absence of a BIT. Substantial literature is devoted to showing how investor-state commitment problems may be overcome in the presence of repeated interactions, reputation effects, or through the use of financial mechanisms such as upfront subsidies (Doyle and van Wijnbergen 1994), (Janeba 2002). There are also alternative legal mechanisms, which in some cases may be close substitutes for BITs as a means of protection from expropriation. For example, U.S. firms may stipulate in their contracts with host governments that disputes be referred to U.S. commercial courts (Pistor 2002). Finally, firms may purchase political risk insurance that is offered by private firms, source governments, host governments, and the Multilateral Investment Guarantee Agency of the World Bank group. Thus, the investment-promoting impact of BITs will depend on how efficient they are in comparison to a variety of alternative means of reducing transaction costs between investors and host governments. Independent of how efficient they are in comparison to alternative solutions to the holdup problem, it is also reasonable to question how effective BITs actually are as a commitment device. In particular, the power of investor-state arbitration is limited by the lack of a world government to enforce the decisions of the tribunals. The extranational arbitration process derives most of its power from raising the reputation costs of refusing to compensate an investor (Guzman 2005). Anecdotal evidence from the investor-state dispute case history suggests that the enhanced reputation effect is present. For example, more than half of the 110 completed cases listed on the ICSID Web site6 ended in settlement between the parties. Furthermore, both investors and defending countries invest significant sums in bringing and defending cases to arbitration tribunals. Both parties have average legal costs plus arbitration fees of around $1.5 million to $2.5 million (UNCTAD 2005). The confidence of investors in the mechanism is further suggested by the fact that despite the high expected legal costs, the rate of submission of disputes is rising rapidly. In 2005 alone, 50 of the total 226 investor-state cases brought under BITs were filed (UNCTAD 2006a). B. Modeling the Economic Function of BITs In this section, I model a host government’s decision whether to participate in a BIT and a foreign investor’s decision whether to invest in the host. The model is intentionally simple as its primary purpose is to motivate my empirical strategy. 1. Description and payoffs The model consists of a simple dynamic game between a monopolist foreign investor and a single potential host government. The potential host government must decide whether to sign a BIT with the investor’s home country in order to help attract the investor. In deciding, the host must weigh up the benefit of the potential new investment against the costs

6. Available at http://www.worldbank.org/icsid/cases/conclude.htm.

400 emma aisbett

of signing a BIT. The host has existing investments from the investor’s home country of magnitude S, on which it levies a tax of τ. If it signs a BIT with the home country, it must compensate both the new investor and any of the existing investors in the event that it expropriates their assets. In the first period, the host chooses whether or not to participate in the BIT and the tax rate, t, which the new investor must pay if they choose to invest. In the second period, the investor decides whether to make the investment. The activity will generate one unit of revenue and requires an investment of K < 1, which cannot be recovered. In the third period, the residual values r and ρ are revealed, which the host gains if it expropriates the new and existing investments respectively. The random variables r and ρ are drawn from a distribution g(•), which is known in advance by both host and investor. For simplicity, I assume that the distribution g(•) is uniform with a support between zero and an upper bound, R: g(•) U[0,R]. Based on the revealed residual values, the host decides whether or not to expropriate any of the investments in its territory. If the host does not expropriate the new investment, the new investor receives 1 − t − K and the host receives t. Similarly, if the host does not expropriate the existing stock of investment, it receives tax revenues Sτ. If the host does expropriate an investment, the payoffs depend on whether the host has ratified a BIT with the investors’ home country. If the host has not ratified a BIT with the home country and it expropriates an investment, it gains the residual value of the expropriated investment, r or ρ, respectively. If, on the other hand, the host has ratified a BIT with the home country and expropriates an investment, it must fully compensate the affected investor for its losses. Thus, if a BIT has been ratified and the host expropriates the new investment, it receives r − (1 − t), while the new investor receives 1 − t − K. If it expropriates the existing stock, it receives ρ − (1 − τ). Working backward from period 3, it is clear that if they have not signed a BIT, the host will expropriate the new investment whenever r > t. Thus, given r U[0,R], the probability of the host not expropriating is:

(1)

An analogous expression describes the probability of expropriating the existing stock of investment. The expected payoff for a host that receives the new investment despite not ratifying a BIT with the investor’s home country is given by:7

7. The following assumes t < R. If not, the host will never expropriate and is expected payoff is simply the tax revenue, t.

bilateral investment treaties and foreign direct investment 401

The payoff to a host that does not ratify the BIT and consequently does not receive the new investment is:

The new investor’s expected payoff to investment is:

The maximum expected investor payoff is at t = and is given by: (2)

I assume that the residual value of an investment is always less than the total revenue it generates when still in the hands of the original investor, so that R < 1. In this case, the full compensation requirement of a BIT will prevent the host from ever expropriating. Thus, the payoffs to a host that ratifies an investment agreement with the home country are: if the host receives the new investment, and if they do not. Because the investor is fully insured against expropriation, its expected payoff to investment is:

2. Host benefit to BIT ratification The host’s decision whether or not to ratify a BIT with the home country will depend on the trade-off between the benefit of extra investment gained and the cost of not expropriating the valuable assets of both existing and new investors. A host that knows it will receive the new investment without ratifying the BIT has no incentive to do so, as it means buying into a costly commitment device without getting any increased investment in return. From equation 2 we know that with full information about the host’s type (R), the new investor will invest in a host in the absence of a BIT if and only if the For a host with R > R, the decision to ratify a BIT with home host has depends on the expected payoff to nonratification and no new investment, compared with that for ratification and gaining new investment. Thus, the host the condition is: will ratify if and only if (3)

Once the BIT is signed, the host can set a maximum tax rate of 1 − K and still leave the investor indifferent between investing and not investing. I assume the

402 emma aisbett

investor invests in this case. Substituting this maximum tax into equation 3 and solving for the maximum R for which a host will be at least as well off after signing the BIT and gaining the new investment gives: (4)

Thus, the model allows us to identify three types of host: • Type A with low expected residual value of expropriated FDI, that is, with R < R, • Type B with medium expected residual value of expropriated FDI, – that is, with R < R < R , and • Type C with high expected residual value of expropriated FDI, that – is, with R < R . For Type A, the expected benefits and associated probability of expropriation are so low that they can induce investment without signing a BIT. For Type B, the incentive to expropriate is high enough that in the absence of the insurance provided by a BIT, investors will not invest regardless of the tax rate. For Type C, the benefits to expropriation are so high that committing to not expropriating by ratifying a BIT is not worthwhile, even if it does allow them to attract new investment. Thus, we would expect Type B hosts with intermediate probability of expropriation to be the most likely to participate in BITs. 3. Empirical implications of the model In order to draw empirical implications from the model, I first need to briefly translate a couple of important model parameters into empirical concepts. A key component of the model is the decision by a single representative investor whether to invest in the host country. The model is normalized by the amount of revenue that would be generated if the investor chooses to invest. I will refer to the empirical application of this concept as the “potential new investment” or the “investment response” to the BIT. The potential new investment usually appears in the model relative to the existing stock of investment (measured equivalently), S. The other important parameter that needs translating into an empirical concept is the upper bound of the distribution of possible residual values that the host obtains if it expropriates an investment, ‘R.’ The parameter R essentially defines the host’s type and, all other things being equal, the probability that the host will expropriate a given investment. In the discussion which follows, I will treat R as if it were a measure whether the host has a ‘good’ or ‘investor friendly’ investment climate, or alternatively whether the host is a high or low probability of expropriating investments. The size of the potential new investment relative to the existing stock and the host’s probability of expropriating are the two main determinants of the benefit to the host of participating in a BIT with the home country. Empirically, I equate increases in the expected benefit to the host of participating to increases

bilateral investment treaties and foreign direct investment 403

in the probability that the host will participate in a BIT. Thus, the key empirical implications of the model are that: • Conditional on the existing stock of investment, increasing the size of the potential new investment increases the probability that the host will participate in a BIT with the home country. • Changes in the host’s investment climate will change the probability that the host participates in BITs. The first implication highlights the potential for overestimating the magnitude of the effect of BITs on FDI if reverse causality is ignored—the bigger the likely increase in FDI, the more likely a BIT is to be ratified. Notice that we are not able to sign the second empirical implication. As I show on page 9, there is a range of expropriation risks over which the host’s benefits to BIT participation increase with decreasing risk, and a range over which they actually decrease. Since there is no way of determining which range a given country is in—particularly relative to investments from a given source—it is not possible to control for this implication by inclusion of a specific variable. The second implication highlights the need to address bias because of omitted variables about the investment climate in the host. I return to the issues of reverse causality and omitted variable bias and explain how my empirical approach addresses them in Section 4.1. Before that, however, I introduce the data and discuss econometric issues common to empirical analyses using this type of data. C. Data and Empirical Strategy There are a large number of econometric and data concerns that should be addressed by any researcher interested in empirically studying the short-run determinants of FDI. I address these issues in detail and derive my chosen regression specification in the appendix. The discussion of specification includes the advantages and disadvantages of different dependent variables for the regressions (e.g., FDI stocks, affiliate sales, log FDI flow) and motivates the choice of log bilateral FDI flow due to its relatively good time series properties and low susceptibility to influential data points. The appendix also discusses the choice of control variables. Based on empirical performance, I adopt a set of controls that combine some of those suggested by (Carr, Markusen and Maskus 2001) (namely share of trade in gross domestic product (GDP) for source- and host-, and the difference in factor endowments as proxied by differences in average years education), with others which are motivated both by the empirical trade literature and by recent theoretical FDI work by (Helpman, Melitz and Yeaple 2004) (namely log source and host GDP and population). Given the potential for omitted variable bias on my BIT coefficient, I also include a number of proxies in the specification. The proxy variables include

404 emma aisbett

country-pair fixed effects to control for unobservable determinants of the strength of the bilateral FDI relationship. Year effects are included to control for global shocks (such as business cycles) and trends in world FDI. I also reduce simultaneity by lagging the explanatory variables, and address heterogeneity by correcting the standard errors for clustering of residuals by country-pair. The resulting base specification is given in equation 5.8 infdi

(5)

where infdi is the log flow of FDI from OECD source country to low- or middle-income host country (a.k.a. bilateral FDI), BIT is a dummy which is zero in years before a BIT between host i and source j has been ratified, and 1 otherwise, Xkt, k = (i, j) is host (i) and source (j) log GDP, log population, and trade share in GDP, Zijt is the skill gap, the product of skill gap and GDP difference, and product of skill gap and host trade share in GDP, γij is a country-pair specific fixed effect, and ηt are year effects and εijt are idiosyncratic errors that I assume are clustered by country-pair. My main data source is an unbalanced panel of bilateral FDI outflows reported by 24 OECD member countries to 28 recipient low- and middle-income countries for the period 1980–99. Although the panel has a potential for 24×28=672 observations per year, missing data means that the actual observations per year is far less. In 1982—the first year for which I have a complete set of data (including all controls) —there are only 39 reporting country-pairs.9 Tables 1 and 2 provide summary statistics for both the FDI data and control variables at the start and end of the study period. Data are discussed further in the appendix. 1. Endogeneity of BIT adoption Although the general specification issues discussed in the appendix are all important for my analysis, the major econometric issue specific to my research question is the endogeneity of the decision to form a BIT. The model in Section 3 motivates the need to address both the potential for reverse causality (in the sense that increasing FDI flows increase the probability of a BIT being formed) and omitted variables (such as the host’s investment climate). Figures 1 and 2 supplement the theoretical motivation for attention to endogeneity issues with an empirical motivation. The figures also preview the conclusions of the regression analysis that will be presented in the next section. Figure 1 shows the strongly increasing trends in both total reported FDI flows and

8. Though all of these specification improvements should be routine when working with country-level panel data, they are remarkably inconsistently applied in the empirical FDI literature. This is a concern for many of the findings of the current literature, as is highlighted in the appendix by Tables 8 and 9, which show the impact of the stepwise addition of the specification improvements. 9. The potential selection bias introduced by the missing data is discussed in Section C.2.

bilateral investment treaties and foreign direct investment 405

table 1. summary statistics for 1982 Variable Bilateral FDI flow Lagged BIT ratification Host GDP (Mill. $US) Host GDP per capita (’000 $US) Host Population (Mill.) Host Trade Share in GDP (%) Source GDP (Mill. $US) Source GDP per capita (’000 $US) Source Population (Mill.) Source Trade Share in GDP (%) Skill Gap (Years Education) Sum of GDPs (Mill. $US) Squared Diff. GDPs Skill_gap∗GDP_diff. Host_trade∗Skill_gap2

Mean

Std. Dev.

Min.

Max.

N

191.987 0.051 0.233 3.559 126.334 37.136 0.672 10.085 66.038 46.936 3.074 0.904 0.396 -1.357 620.38

446.165 0.223 0.24 1.598 249.564 24.286 0.403 1.231 37.627 21.211 2.538 0.482 0.473 2.041 843.268

-36.87 0 0.046 0.908 11.147 14.29 0.057 7.239 5.123 28.65 -3.5 0.109 0 -7.678 7.558

2434.994 1 0.948 6.467 981.24 110.86 1.212 11.746 116.78 126.35 7.543 2.16 1.36 3.525 3722.36

39 39 39 39 39 39 39 39 39 39 39 39 39 39 39

Std. Dev.

Min.

Max.

N

12106 2 3.129 7.16 1227.2 221.54 5.528 20.647 267.74 141.7 7.767 8.657 30.448 18.812 4305.1

186 186 186 186 186 186 186 186 186 186 186 186 186 186 186

table 2. summary statistics for 1998 Variable Bilateral FDI flow Lagged BIT ratification Host GDP Mill. US Dollars Host GDP per capita Host Population (Mill.) Host Trade Share in GDP Source GDP Mill. US Dollars Source GDP per capita Source Population (Mill) Source Trade Share in GDP Skill Gap Sum of GDPs Squared Diff. GDPs Skill_gap∗GDP_diff. Host_trade ∗Skill_gap2

Mean

419.526 1323.204 −491.667 0.392 0.501 0 0.624 0.924 0.01 4.331 1.921 1.678 238.161 410.234 2.719 60.526 50.005 16.7 0.879 1.502 0.004 14.581 3.022 5.299 52.03 72.994 0.272 73.814 30.338 25.63 2.751 2.482 −5.041 1.503 1.724 0.081 3.297 7.636 0 9.561 −1.344 −42.035 646.864 721.078 0.324

406 emma aisbett figure 1. total reported fdi and total bit s ratified in the data Growth in BITs and Total Reported FDI 200

150 100000 100 50000 50

No. of Pairs with BITs Ratified

Total Reported FDI

150000

0

0 1980

1985

1990 year...

Total Reported FDI

1995

2000

No. of Pairs with BITs Ratified

figure 2. mean fdi for signing and non-signing pairs and total bit s ratified Growth in BITs and Mean FDI for Signing and Non-signing Pairs

300

200

200

150 100

100

50 0

0 1980

1985

1990 year...

Mean FDI BIT Pairs No. of Pairs with BITs Ratified

1995

2000

Mean FDI non-BIT Pairs

No. of Pairs with BITs Ratified

400

bilateral investment treaties and foreign direct investment 407

number of BIT ratified by country-pairs in the data. FDI and BITs are clearly highly correlated over time. Figure 2 suggests the importance of addressing the potential endogeneity of BIT formation rather than relying on simple correlations of the type depicted in Figure 1. Figure 2 shows three lines: the mean FDI flow between country pairs that never sign a BIT during the sample period; the mean FDI flow between country pairs that do sign a BIT during the sample period; and the number of BITs ratified by country pairs in my data. The first half of the graph shows that BITs begin to take off around 1985, a couple of years after mean FDI between signing pairs takes off. The second half of the graph shows that in the 1990s, when the rate of increase in BIT formation was highest, the mean FDI flow between signing pairs was growing slowly relative to both its own growth earlier in the period, and relative to flows between nonsigning pairs. Thus, the first half of Figure 2 is consistent with the explanation that BIT participation is driven by initial increases in bilateral FDI flows, and the second half of Figure 2 is consistent with the explanation that BIT participation is driven by variables that increase both FDI flows and BIT formation. In general, the preferred approach to addressing endogeneity is to use an exogenous instrument. In the case of BITs and FDI, however, it is very hard to identify a good instrument. Thankfully, BITs themselves are exogenous ex post. That is, once a BIT is in place, it cannot become more or less in place for at least ten years. This means that, as is commonly done in the program evaluation literature, I can address endogeneity by fully controlling for adoption of BITs. The three-dimensional (host, source, year) nature of the OECD FDI data allows me to construct three sets of controls for the adoption of BITs: host-year dummies, source-year dummies, and host-source (i.e., country-pair) time trends. Host-year dummies mean that there is a separate dummy variable for all but one host country and every year. Source-year dummies are analogously defined. The motivation for including these variables is to control for any unobserved or imperfectly observed features of the investment climate in host or source in each year. In particular, these dummies control for changes in exchange rates, changes in host domestic policies toward FDI, changes in host expropriation probability, elections, and so forth. Thus host-year dummies, in particular, address the concern that the coefficient on BIT ratification is driven by the omission of changes in host country investment climate, which lead to an increase in both FDI flows and BIT participation. Country-year dummies have also been recently recommended in the context of trade gravity models by (Baldwin and Taglioni 2006). The addition of country-pair specific time trends to the base specification helps to control for adoption of BITs driven by reverse causality from FDI flows. The model in Section B shows that, conditional on the existing stock of FDI, a higher bilateral flow of FDI will lead to a higher probability of BIT formation. This means that conditional on fixed effects (which control for the existing

408 emma aisbett

bilateral stock of FDI), country-pairs with higher bilateral FDI growth rates are more likely to form BITs. A final way to reduce the bias in the BIT coefficient because of the endogeneity of BIT adoption is to correct for autocorrelation. Although the presence of autocorrelation does not lead to inconsistent estimates under the standard assumption of strict exogeneity of right-hand side variables, it can exacerbate existing bias when strict exogeneity is violated. Intuitively, if there is feedback from higher FDI flows to BIT formation, a positive shock to FDI flows in one year increases the probability that a BIT is formed by the following year. Thus, if autocorrelation is not corrected for, the BIT dummy may capture the omitted effect of the serial correlation in the disturbance term. I test for first-order autocorrelation using the (Bhargava, Franzini and Narendranathan 1992) modified Durbin-Watson statistic for unbalanced panel data. I then correct for it by estimating the regression using (Baltagi and Wu’s 1999) feasible generalized least squares estimator for unbalanced panels with fixed effects in the presence of first order autocorrelation. Having established the presence of serial correlation, I report results for both the FGLS and the ordinary least squares with fixed effects for the remainder of the different specifications for addressing endogeneity and robustness checks. 2. Selection bias Before moving onto the results, there is one final potential econometric concern that needs to be addressed, that is, the large number of missing values in the OECD bilateral FDI dataset. There are two causes of missing observations in the data. The first is simply that different OECD members started reporting their FDI outflows at different times during the period of study. The first reporting year for each source country is reported in Table 10 in the appendix. Secondly, countries only report investment outflows greater than a particular size, where the threshold varies with the reporting country. Since bilateral FDI flows have been generally increasing since the 1980s, the number of reporting pairs has also increased. Figure 3 plots the increases in reporting source countries, reported host countries, and mean bilateral FDI over the study period. Nonrandom missing data is a potential concern for my estimation. Indeed, several papers have highlighted this issue for either FDI or trade data and demonstrated the importance of jointly estimating participation and flow equations to correct selection bias (Razin et al. 2003), (Helpman et al. 2005), (Razin et al. 2005). Joint estimation of participation and flow equations, however, requires exclusion restrictions on the flow equation in order to be well identified (Wooldridge 2002). In the absence of a structural model, I have little basis for exclusion restrictions. An alternative approach may be to treat the issue as one of truncation and use a Tobit framework; however, this would preclude the use of a fixed effects estimator, which is clearly unsatisfactory.10

10. The problem with this approach is that Tobit regressions with fixed effects are known to be biased. Using Monte Carlo studies, Greene (2003) shows that the coefficient

bilateral investment treaties and foreign direct investment 409

figure 3. mean fdi and number of source (host) countries reporting (reported) at least one bilateral fdi flow

300 30

25

200

20

Mean Bilateral FDI flow

Number of Reporting/Reported Countries

Growth in Reporting and Mean Flows Over Time 35

100

15 1980

1985

1990 year...

Source (Reporting) Countries

1995

2000

Host (Reported) Countries

Mean Bilateral FDI flow

The above discussion assumes, however, that selection bias is a problem for my estimates. It turns out that this is not the case, precisely because my preferred specification includes country-pair fixed effects.11 Table 3 reports the results of a test for selection bias in different specifications of the bilateral FDI relationship. The test is based on the method of (Nijman and Verbeek 1982) for a random effects context, applied to fixed effects as suggested by (Wooldridge 2002, p. 581). It simply involves the inclusion of a lagged indicator for missing FDI data in the

estimates have very low bias, especially for relatively long panels, such as the one used here; however, Greene found that the estimated variance is substantially biased, leading to an underestimation of standard errors. He concludes that when interested in a dummy variable such as a treatment effect, a random effects or pooled specification is preferred to the fixed effect model. Given the earlier analysis, the omission of fixed effects would be a serious limitation. 11. It would be more technically correct to say that attrition bias—that is, bias due to country pairs which appear in some but not all years—is not a problem. It remains true that the sample is not a random cross-section of country-pairs. To be precise, only 711 of the potential 1,334 potential country-pairs report flows for at least one year in the sample period; however, given the potential sample covered by the OECD dataset only ever included the largest source and recipient countries, this is nothing new. Rather, the findings should be qualified by acknowledging that they represent the effect of BITs on FDI from major source countries to their major recipient countries.

410 emma aisbett

regression equation. The test for selection bias amounts to a t-test of whether the coefficient on the lagged missing indicator is different from zero.12 The results of the test for selection bias in my preferred policy specification are shown in column 3 of Table 3. The lagged missing indicator is completely insignificant, suggesting that selection bias is not a problem in this specification. Columns 1 and 2 of Table 3 explain the apparent contradiction between my finding and the results of some other authors. Column 1 reports the results for a regression which includes year effects but no country or country-pair effects. This is consistent with the specification used by (Razin et al. 2003). Column 2 reports the results for a regression which includes year effects and individual host and source effects, but not country-pair effects. This is consistent with the specification used by (Helpman et al. 2005) and (Razin et al. 2005). In both cases the lagged missing indicator is significant at the 1% level, suggesting selection bias can be a real concern if country-pair fixed effects are not used. D. Results of analysis of bilateral FDI data This section presents the results of the empirical analysis of the relationship between FDI and BITs. I begin by examining the robustness of the BIT coefficient to the range of specifications improvements discussed in Appendix A. I find a strong positive correlation between FDI and BITs in all these specifications. This shows that my dataset is capable of reproducing the findings of (Neumayer and Spess 2005) and (Salacuse and Sullivan 2004), who both conclude that BITs have a strong positive impact on FDI. I then show that this finding is not robust to addressing the endogeneity of BIT participation by controlling for country-level characteristics at the time of ratification, or controlling for the underlying trends in bilateral FDI between country-pairs. I also show that correcting for autocorrelation using (Baltagi and Wu’s 1999) feasible generalized least squares estimator lowers the point estimate of the BIT dummy. The effect from correcting for autocorrelation is what we would expect if there is feedback from higher FDI in one year to increased probability of BIT formation in the next. The second results subsection supplements the regression analysis with graphical event-study analysis. Consistent with the regression findings, the graphical analysis shows that BITs occur during times of increasing bilateral FDI, but it shows no evidence that BITs cause an increase in FDI. In the final subsection, I construct a measure of BIT participation with OECD countries other than the source and find no evidence that BITs attract FDI by acting as a signal of a safe or productive investment environment. 1. Main results I begin the analysis of the relationship between BITs and FDI by showing the robustness of the BIT coefficient to a range of specification improvements that are commonly used in the literature to reduce omitted variable and simultaneity bias and correct for heteroskedasticity. These specification improvements are 12. The more sophisticated test suggested by Wooldridge (2002, p. 581) is not applicable here, as it also relies on exclusion restrictions from the flow equation.

bilateral investment treaties and foreign direct investment 411

table 3. a simple test shows that selection bias due to missing data is not a concern when country-pair fixed effects are used. log fdi is regressed on indicator of missing data plus other controls. missing data indicator is insignificant when country-pair fixed effects are used Coefficient

Labels

OLS

Source/ Host FE

Countrypair FE

fdiD

Missing Data Indicatora

lnJgdp

Source Log GDP

lnJpop

Source Log Population

lnIgdp

Host Log GDP

lnIpop

Host Log Population

edgap

Skill Gap

Itragdp

Host Trade Share in GDP

Jtragdp

Source Trade Share in GDP

edgapgdpdiff

Skill_gap∗GDP_diff.

Itragedgapsw

Host_trade∗Skill_gap2

Source/host effects Country-pair effects Year effects Observations

–

−1.505∗∗∗ (0.19) 8.512∗∗∗ (0.64) −8.054∗∗∗ (0.61) 0.776∗ (0.43) −0.681∗ (0.37) 0.247∗∗∗ (0.087) −0.00801 (0.0062) −0.00819 (0.0069) 0.0565∗∗∗ (0.021) 0.000165 (0.00019) No

–

No

No

Yes

– –

Yes 2317

Yes 2317

Yes 2317

−0.695∗∗∗ −0.0621 (0.14) (0.10) 0.828 1.957∗∗∗ (0.59) (0.56) −3.283 −7.354∗∗ (3.19) (3.25) 0.340 −0.293 (0.38) (0.38) −0.425 −1.828 (1.45) (1.76) −0.102 −0.289∗∗∗ (0.090) (0.083) −0.00953∗∗ −0.0136∗∗∗ (0.0046) (0.0045) 0.0136 0.0247∗∗ (0.011) (0.011) 0.0454∗∗ −0.0343 (0.020) (0.027) 0.0000795 0.000280∗∗ (0.000100) (0.00014) Yes No

a Indicator equals 1 if data for the country pair were missing in previous year, zero otherwise. Robust (clustered) standard errors in parentheses ∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1

discussed in detail in the appendix. The results of this exercise are presented in Table 4. The BIT dummy is robustly, economically and statistically significant. BITs, source GDP, and the host trade share in GDP are the only variables that remain significant and of the same sign across specifications. Furthermore, even in the most conservative specification (column 5), the magnitude of the coefficient suggests BITs are associated with an increase in bilateral FDI greater than 50%.

412 emma aisbett table 4. bit participation robustly correlated with fdi log bilateral fdi regressed on indicator of bit ratification plus other controls. specifications are increasingly conservative moving left to right Coefficient

Labels

(1) lnfdi

(2) lnfdi

(3) lnfdi

(4) lnfdi

(5) F.lnfdi

R

BIT ratification indicator

lnJgdp

Source Log GDP

lnJpop

Source Log Population

lnIgdp

Host Log GDP

lnIpop

Host Log Population

edgap

Skill Gap

Itragdp

Host Trade Share in GDP

0.799∗∗ (0.35) 8.156∗∗∗ (0.69) −7.267∗∗∗ (0.68) 0.958∗∗ (0.42) −0.969∗∗ (0.38) 0,305∗∗∗ (0.100) −0.0196∗∗∗ (0.0065)

0,717∗∗∗ (0,11) 2,272∗∗∗ (0,40) 2.936 (2.05) 0.599∗∗ (0,26) 2.161∗∗∗ (0.77) −0.226∗∗∗ (0.072) −0.00895∗∗∗ (0,0035)

0.467∗∗∗ (0.12) 1.861∗∗∗ (0.43) −4.152∗ (2.18) 0,217 (0.26) −1,470 (1.02) −0.220∗∗∗ (0,072) −0,00923∗∗∗ (0.0035)

0.467∗∗∗ (0,15) 1.861∗∗∗ (0,57) −4.152 (3,23) 0,217 (0,40) −1.470 (1.67) −0.220∗∗∗ (0.075) −0,00923∗∗ (0,0041)

0.444∗∗∗ (0.17) 1.861∗∗∗ (0.56) −6.905∗∗ (3.24) 0,279 (0.37) −0.684 (1.73) −0.283∗∗∗ (0,081) −0,0137∗∗∗ (0,0044)

Source Trade Share in GDP

edgapgdpdiff

Skill_gap∗GDP_di£

Itragedgapsw

Host_trade∗Skill_gap2

Constant

–

Country-pair FE Year effects Cluster robust errors Lagged controls Observations R-squared Number of country-pairs

– – – – – – –

0,0220∗∗ (0.0099) 0.0429∗ (0.024) 0.000202 (0.00021) 40.75∗∗∗ (3.97) No No No No 2098 0.51

0.00537 (0,0062) −0,0193 (0.022) 0,000285∗∗∗ (0.00011) −13,34∗ (7,39) Yes No No No 2208 0.28 281

0.000728 (0,0080) −0.0343 (0.022) 0.000252∗∗ (0,00011) 24.28∗∗∗ (8.74) Yes Yes No No 2208 0.32 281

0.000728 (0.011) −0,0343 (0,028) 0.000252∗ (0.00014) 24,28∗ (12.8) Yes Yes Yes No 2208 0,32 281

0.0234∗∗ (0.011) −0.0427 (0.026) 0,000276∗∗ (0.00014) 29.58∗∗ (13.8) Yes Yes Yes Yes 2317 0.35 287

Time-invariant controls included for column 1 but not reported are: number landlocked, number of islands, land border, colonial relationship and distance. Taken from Andrew Rose’s website and defined as in Rose (2004) Standard errors in parentheses∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1

bilateral investment treaties and foreign direct investment 413

Jtragdp

414 emma aisbett

The strong correlation of FDI and BITs,even when conditioning on all the factors usually applied in the literature, suggests two things. Firstly, BITs and FDI are closely related economic phenomena. That is, we would be surprised to see such robust correlation if BITs were merely a political or diplomatic tool with no economic policy relevance. Secondly, the coefficient on the BIT almost certainly does not represent the causal effect of BITs on FDI. Even the most enthusiastic proponent of BITs would not claim that they increase FDI flows by an average greater than 50%. Rather, the results in Table 4 are entirely consistent with the endogeneity of BITs predicted by my model in Section 3. I next demonstrate the effect of controlling for the endogeneity of BIT participation in a number of ways. Each of these controls takes as a base specification the regression reported in column 5 of Table 4, and given in equation 5. I first focus on reverse causality as a source of endogeneity of BITs. As discussed on page [IPR], two potential ways to reduce the bias on the estimated BIT coefficient caused by reverse causality are to correct for autocorrelation and to include country-pair specific time trends. I first confirm the presence of first order autocorrelation in the residuals from the base specification using Bhargava’s modified Durbin-Watson test for serial correlation in an unbalanced panel (Bhargava et al. 1992). This test easily rejects the null of no serial correlation at the 5% level.13 I, therefore, correct for first order autocorrelation using (Baltagi and Wu’s 1999) FGLS estimator for unbalanced panels with fixed effects. Columns 1–3 of Table 5 report respectively the OLS base specification (identical to column 5 of Table 4), the base specification estimated correcting for autocorrelation using FGLS, and the OLS base specification plus country-pair specific time trends.14 The results show that reducing the influence of reverse causality through either method reduces the estimated coefficient on BITs. In column 2, the coefficient remains barely significant at the 10% level, while in column 3, it is not significant even at that level. The insignificance of the BIT coefficient in column 3 needs to be understood in the context of the other coefficients in the regression. It is the case that none of the coefficients in column 3 are significant and of the theoretically predicted sign; however, the two other previously robust coefficients, source GDP and host trade share in GDP, are insignificant in column 3 of Table 5 because of a large increase in standard error, while the coefficient on the BIT has become insignificant mainly because the point estimate is less than half that in column 1. It is also worth noticing a number of other appealing features of the FGLS estimates. Firstly, the unrealistically large negative coefficient on source population is now much smaller and much more the magnitude we would expect. It 13. Based on comparison of the test statistic with the 5% significance tables in Bhargava et al., p. 537. 14. The autocorrelation coefficient for the results in column 2 is 0.36, and the countrypair time trends are jointly highly significant.

bilateral investment treaties and foreign direct investment 415

table 5. significance of bits not robust to controlling for country-pair trends from left to right: log bilateral fdi regressed on base specification; with correction for autocorrelation; and with addition of country-pair time trends Coefficient

Labels

(1) F.lnfdi

(2) F.lnfdi

(3) F.lnfdi

R

BIT ratification indicator

0.437∗∗∗

0.245∗

0.204

lnJgdp

Source Log GDP

lnJpop lnlgdp

Source Log Population Host Log GDP

lnIpop

Host Log Population

edgap

Skill Gap

Itragdp

Host Trade Share in GDP Source Trade Share in GDP Skill_gap∗GDP_diff. Host_trade∗ Skill_gap :l – – –

(0.15) 1.839∗∗∗ (0.55) -6.811∗∗ (3.25) 0.272 (0.38) -0.732 (1.76) -0.2S7∗∗∗ (0,082) -0.0135∗∗∗ (0.0044) 0.0235∗∗ (0.011) -0.0391 (0.027) 0.000282∗∗ (0.00014) Yes Yes No

(0.15) 1,030∗∗ (0.49) -0.0765 (0.49) 0.122 (0.34) 0.246 (0.34) 0,00180 (0.082) -0.0117∗∗ (0.0045) -0.00112 (0.00&1) -0.0119 (0.028) 0.0000445 (0,00014) Yes Yes Yes

(0.19) 1.854 (1.76) -9.521 (12.7) 0.fi24 (0.60) 12.96 (8.09) -0.343∗∗∗ (0.093) -0.0119 (0,0074) 0,0205 (0.013) -0.0363 (0.047) 0.000420∗∗∗ (0.00014) Yes Yes No

–

No

No

Yes

– –

2317 287

2030 258

2317 287

Jtragdp edgapgdpdiff Itragedgapsw Country-pair FE Year effects FGLS / autocorrelation Country-pair time trends Observations Number of country-pairs

Robust standard errors in parentheses ∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1

416 emma aisbett

now suggests that FDI is increasing in both the GDP and GDP per capita of the source. Secondly, two of the coefficients that were significant with the opposite sign to that predicted by the knowledge-capital model of FDI—that is, the skill gap and the interaction of the squared skill gap with host trade share—are now insignificant. In fact, the only three coefficients that are significant in the FGLS regression are the same three coefficients (BITs, source GDP and host trade share in GDP) that remained significant and of consistent sign across specifications in Table 4. Overall, the FGLS estimates are preferred to the OLS estimates. I therefore report the results of both estimators for the remaining regressions. The second source of endogeneity bias for BITs suggested by the model in Section B is omitted variables. As discussed earlier, the bilateral panel data I use provide a particularly useful means of controlling for omitted variables that are hard to measure. Specifically, I construct two sets of interaction terms: host-year dummies and source-year dummies. The results of including these interactions both individually and together are reported in Table 6. OLS estimates with errors corrected for clustering are in columns 1–3 and FGLS estimates correcting for autocorrelation are in columns 4–6. Columns 1 and 3 have host-year dummies in place of the year dummies in columns 1 and 2 of Table 5. The inclusion of host-year dummies causes the point estimate of the BIT effect to fall under the clustered error assumption and rise slightly for the autocorrelation case. Columns 2 and 4 have source-year dummies in place of the year dummies. In comparison to columns 1 and 2 of Table 5 the inclusion of source-year dummies decreases the point estimate of the BIT dummy. This suggests that omitted source country conditions are at least as important as omitted host country factors in biasing the BIT coefficient upward in the base specification. Finally, columns 3 and 6 show the results of including both host- and source-year dummy sets. The BIT coefficient reduces further and is now insignificantly different from zero at the 10% significance level for both OLS and FGLS estimators. Note that estimates in Table 6 are not reported for some variables because they are colinear with the sets of county-year dummies. Of course, the inclusion of such a large set of dummy variables could reduce the statistical significance of the BIT variable for two reasons: Either it simply reduces the degrees of freedom and power of the regression, or it effectively controls for important variables that were previously omitted. Two factors point to the latter interpretation. Firstly, a joint test of the significance of the dummy variables is highly significant. Secondly, the loss of significance of the BIT dummy is driven by a fall in magnitude of the point estimate. There is very little increase in the standard error of the BIT coefficient with the addition of the country-year interactions using either OLS or FGLS. 2. Graphical analysis An alternative means of illustrating the lack of evidence of causal impact of BITs is to literally illustrate the trends in FDI around the time of BIT ratification. I do this for both unconditional and conditional FDI flows in Figure 4.

Coefficient

Labels

(1)

F.lnfdi

(2) F.lnfdi

(3)

(4)

F.lnfdi

F.lnfdi

0,273∗ (0.15) –

0,210 (0.15) –

–

–

0.326 (0,34) −1.094 (1.80)

–

0.310∗ (0.16) 1.747∗∗∗ (0.56) −0.735 (0.69) –

–

–

–

0.116 (0.11)

R

BIT ratification indicator

0.396”∗ (0.16)

lnJgdp

Source Log GDP

1.719∗∗∗

lnJpop

Source Log Population

−S.S27∗

lnlgdp

Host Log GDP

(3.28) –

liilpop

Host Log Population

–

edgap

Skill Gap

−0,160

−0.274∗∗

Host Trade Share in GDP

(0.10) –

(0.12)

(0.61)

Itragdp

−0.0124∗∗

(0.0051)

–

(5) F.lnfdi

(6)

0,157 (0.15) –

0.108

–

–

0,440 (0,37) −0,285 (0.42)

–

−0,208∗∗

–

F.lnfdi

(0.17) –

–

(0.11) −0.0110∗∗ (0.0045)

–

Continued

bilateral investment treaties and foreign direct investment 417

table 6. significance of bits not robust to controlling for host-year and source-year effects log bilateral fdi regressed on base specification plus host-year and/or source-year effects. left columns ols, right column fgls correcting autocorrelation. non-reported coefficients are for variables colinear with country-year controls

Coefficient

Labels

(1)

F.lnfdi

Jtragdp

Source Trade Share in GDP

edgapgdpdiff

Skill_gap∗GDP_diff.

Itragedgapsw

Host_trade∗Slall_gap2

Country-pair FE Year effects FGLS/autocorrelation

Host-year effects Source-year effects Observations Number of country-pairs Robust standard errors in parentheses ∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1

(2) F.lnfdi

(3)

(4)

F.lnfdi

F.lnfdi

Yes Yes No Yes Yes 2317 287

0.00108 (0.0091) -0.0338 (0,033) -0.000326∗ (0.00018) Yes Yes Yes Yes No 2030 2S8

0,0214∗∗

(0,010) -0,0515 (0,037) 0.0000584 (0,00018) Yes Yes. No Yes No 2317 287

-0.0101 (0,035)

0.000332∗∗ (0.00015) Yes Yea No No Yes 2317 287

(5) F.lnfdi

(6)

–

–

0,0122

–

F.lnfdi

(0.041) 0.000225∗

–

(0,00014)

Yes

Yea

Yes

Yea

Yes No Yes 2030 258

Yea Yes Yes

2030 258

418 emma aisbett

table 6. significance of bits not robust to controlling for host-year and source-year effects log bilateral fdi regressed on base specification plus host-year and/or source-year effects. left columns ols, right column fgls correcting autocorrelation. non-reported coefficients are for variables colinear with country-year controls (cont’d...)

bilateral investment treaties and foreign direct investment 419

figure 4. event study graphs support the conclusion that correlation between bits and fdi is driven by endogeneity of bit adoption. change in conditional and unconditional log fdi round the time of bit ratification. top-left graph (unconditional fdi flows) highlights the endogeneity of bit adoption. other graphs show the impact of progressively adding controls for bit adoption FDI conditioned on base specification

Unconditional Mean Log FDI 3.5

Mean Ratified

Mean Log Bilateral FDI

.6 3

2.5

.4 .2 0 −2

2

1.5 −4

−2

0

2

−4

4

Years Since BIT Ratification

−2

0

2

4

Years Since BIT Ratification

FDI conditioned on base specification plus country-pair trends

FDI conditioned on base specification plus host and source year effects

.3 .2 .1 0 −.1

Mean Ratified

Mean Ratified

.4 .2 0

−.2 −.4 −4

−2

0

Years Since BIT Ratification

2

4

−4

−2

0

2

4

Years Since BIT Ratification

Graph 1 in Figure 4 plots the unconditional mean log bilateral FDI flow from three years prior to ratification through to three years post ratification of a BIT. The remaining graphs in Figure 4 are the corresponding plots for conditional FDI flows. In graph 2, log FDI is conditioned on the base specification given in equation 5 with the omission of the BIT dummy. In graph 3, the conditioning set is the base specification minus the BIT dummy, plus country-pair time trends (analogous to the results in column 3 of Table 5), while in graph 4, the set is the base specification minus BIT dummy plus host-year and source-year dummies (compare with column 3 of Table 6). The evidence presented in Figure 4 supports the conclusion that the positive and significant coefficient on BIT ratification in my base specification is due to not fully controlling for the endogeneity of BIT participation in that specification. The residual FDI in graph 4, in particular, is indistinguishable from white noise. 3. BITs as signaling device Previous authors have suggested that the main function of BITs in attracting FDI may be not to actually provide increased investor protections but to signal that the host already provides a low-risk investment environment (Hallward-Driemeier 2003), (Neumayer and Spess 2005). If this is the case, then BITs should increase FDI received from all sources, not just the BIT partner.

420 emma aisbett table 7. no evidence that bits increase fdi through signaling a good investment climate log bilateral fdi flow regressed on base specifications with indicator of ratification of bit between host and source replaced by number of bits ratified by host with oecd countries other than the source Coefficient

Labels

(1) F.Lnfdi

(2) F .Lnfdi

R1ImJ lnJgdp

BITs signed with other OECD partners Source Log GDP

lnJpop

Source Log Population

lnIgdp

Host Log GDP

lnIpop

Host Log Population

edgap

Skill Gap

Itragdp

Host Trade Share in GDP Source Trade Share in GDP Skill_gap∗GDP_diff.

0.0173 (0.016) 1.979∗∗∗ (0,56) −7.406∗∗ (3.28) 0,287 (0.38) −0.674 (2,02) −0.288∗∗∗ (0.083) −0.0143∗∗∗ (0.0047) 0.0254∗∗ (0.011) −0.0395 (0.027) 0.0002S4∗∗ (0.00014) Yes Yes No 2317 287

0.0177 (0.013) 1.099∗∗ (0.49) −0.124 (0.50) 0.108 (0.35) 0.273 (0,34) −0.00370 (0.083) −0.0120∗∗∗ (0.0046) −0.000135 (0.0091) −0.0139 (0.028) 0.0000418 (0,00014) Yes Yes Yes 2030 258

Jtragdp edgapgdpdiff Itraged gapsw Country-pair FE Year effects FGLS /autocorrelation Observations Number of country-pairs

Hast_trade∗ Skìll_gap2 – – – – –

Robust standard errors in parentheses ∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.

To test this possible explanation, I return to the base specification and replace the BIT dummy with the number of BITs that the host has signed with OECD countries other than the source. If BITs act as a signal of a safe investment climate, then source FDI should respond to participation by the host in BITs with other source countries. The results in Table 7 suggest that signaling is not the cause of the strong correlation between FDI and BITs in the base specification in Table 4. BITs ratified with other source countries are not a significant predictor of FDI at the 10% level.

bilateral investment treaties and foreign direct investment 421

conclusion Bilateral investment treaties are one of the most popular policy initiatives undertaken by low- and middle-income countries in the race to attract a larger share of global FDI. Like most such initiatives, BITs are not without costs. Resources are expended on the design and negotiation of BITs. When ratifying BITs, countries sacrifice policy flexibility and risk sizable fines and legal costs if they are sued by an investor. The experience of the United States and Canada under the BIT-like Chapter 11 of the North American Free Trade Agreement shows that even welldocumented actions undertaken by countries that are renowned for their investor protections, and undertaken to protect public health or the environment, may be subjected to claims by investors. Yet the number of BITs and similar agreements embedded in regional trade agreements continues to grow. Countries appear to believe that the FDI-promoting abilities of BITs outweigh these legal and policy costs. I find no evidence to support this belief. Furthermore, my results suggest that previous findings of a positive impact of BIT participation (Neumayer and Spess 2005) and (Salacuse and Sullivan) are almost certainly due to misspecification and insufficient attention paid to the endogeneity of BIT participation. Although this paper addresses a specific policy question, the empirical issues it addresses are relevant to the larger literature on the impacts of trade and FDI policy. Because of the relatively poor explanatory power of current theoretically motivated models of FDI, it is important that this literature consider carefully the influence of omitted variables. One advantage of using bilateral panel data is that country-pair fixed effects may be used to control for time-invariant variables affecting the bilateral FDI relationship. Though panel data has helped to reduce the omitted variables problem of earlier cross-section studies of FDI, it has also brought new challenges that have not always been fully appreciated. Many papers related to FDI are motivated by the observation that global FDI has grown rapidly over the last couple of decades, much more rapidly than common explanatory variables such as GDP and trade. It is ironic, therefore, that so many of these studies neglect to properly account for the time series properties of the data. My findings show the importance of the inclusion of year effects to remove common time trends if spurious correlation is to be avoided. I also argue that it is preferable to use FDI flows rather than FDI stocks as a dependent variable in order to reduce the degree of autocorrelation. Even when using log FDI flow as a dependent variable, I find significant autocorrelation and show that the use of feasible Generalized Least Squares to correct for this improves the estimates in qualitatively important ways. Correcting for autocorrelation is particularly important with endogenous right-hand side variables as autocorrelation may exacerbate endogeneity bias. Finally in terms of general methodological issues, I show that the selection bias resulting from the large number of nonrandom missing values in the bilateral FDI data is eliminated by the inclusion of country-pair fixed effects.

422 emma aisbett

Consideration of the above general specification issues in FDI regressions leads to my preferred base specification that includes country-pair fixed effects, year effects, lagging of dependent variables, and adjusting errors for clustering by country-pair. Using this specification, I find that BITs are positively and significantly correlated with FDI flows. This finding is consistent with those obtained by (Neumayer and Spess 2005), who apply a similar specification to aggregate host-country FDI inflows. My finding of a strong positive correlation shows that the difference between (Hallward-Driemeier 2003) finding of no effect and Neumayer and Spess’s finding of a strong effect are not due, as Neumayer and Spess suggest, to the former author’s use of the bilateral OECD data and associated restricted sample of countries. Instead Hallward-Driemeier’s results are likely to be driven by her use of levels FDI flows rather than log FDI. I show in Appendix A that when logs are not taken, FDI data is highly skewed and prone to influence by extreme observations. This initial finding of strong correlation between FDI and BITs does not, however, imply that BITs caused an increase in FDI. I use a simple model to show that BIT participation is endogenous and may be driven by omitted variables such as a change in the domestic policy environment of the host. My model also shows the potential for reverse causality, where a higher growth rate of FDI leads to increased probability of a BIT being formed. I find that controlling for either of these possibilities eliminates the statistically significant correlation between BIT participation and FDI flows. It is possible, however, that some of my attempts to deal with endogeneity obscured a potential signaling effect of BITs. Using the bilateral data, I am able to explicitly test for signaling by the inclusion of the number of BITs that the host has ratified with other OECD countries in the regression. If participation in BITs does signal a safe or productive investment environment, there should be an increase in bilateral FDI in response to ratification of treaties with other major source countries. I find no evidence of such an effect. Thus, the strong correlation between BITs and FDI in the base specification appears to be driven by the endogeneity, rather than either a direct or a signaling effect of the BITs. The major limitation of my analysis, which I would argue is common to studies of the short-run determinants of FDI, is that once spurious correlation and endogeneity are accounted for, the standard control variables have very little explanatory power. The concern, therefore, is that the finding of no effect of BIT participation on FDI flows is driven by data limitations. In light of this limitation, I do not conclude that BITs are ineffective. Instead, I conclude that there is no evidence of that BITs have an effect, and that previous findings in the literature of a positive impact of BITs were probably due to not properly accounting for the endogeneity of BIT participation and other specification issues. Aside from my finding of no significant impact once endogeneity is accounted for, there are a number of additional reasons to believe that the initial strong correlation between BITs and FDI was driven by the endogeneity of BITs. Firstly,

bilateral investment treaties and foreign direct investment 423

the magnitude of the BIT coefficient in the base specification implied that the ratification of a BIT brought on average an increase in bilateral FDI inflow of more than 50%. This figure is implausibly large. Secondly, the loss of significance of the BIT coefficient when additional controls were introduced to reduce endogeneity bias was caused primarily by a decrease in the magnitude of the point estimate, not by an increase in the standard errors. A large increase in standard errors would be expected if the loss of significance was driven by data limitations. Finally, I undertake a graphical event study that shows clearly that BITs are formed during times of increasing bilateral FDI flows, but it shows no evidence of an increase in flows after the BIT is ratified. There are a number of potential explanations for the apparent disconnect between the effort countries place on signing BITs and the lack of measurable response of investors to these efforts. The first is suggested by my model and related findings. The fact that BIT participation increases when expropriation risk has fallen and when FDI flows are already increasing will make the potentially small effect of the BITs difficult to identify within the larger changes. In a similar vein, it is possible that BITs are only of relevance for certain sectors, making their impact difficult to identify in aggregate data. Expropriation risk tends to be greatest in natural resource extractive industries, which are an example of vertical or factor seeking FDI. My results, and those of others (Blonigen, Davies and Head 2003), show that evidence of vertical FDI is hard to identify in aggregate data. Thus, future attempts to identify impacts of BITs may want to focus on bilateral FDI data disaggregated by sector or industry. It is also possible that there is no evidence of an investment response to BITs simply because there was none. It may be the case that while governments have always considered BITs economically significant, investors have not. Evidence that investors have been slow to trust BITs as a commitment device is provided by the very rapid increase in the number of investor-state arbitration cases being brought over the last few years, in a global climate that is generally continuing to become more investor-friendly (UNCTAD, 2006b). It appears that over time, as more disputes brought to arbitration under BITs have been completed and more settlements reached, confidence in the institution of investor-state arbitration may be growing. This means that the positive impacts of BITs on investor confidence have come long after many of the BITs were ratified. Finally, it is possible that the primary function of BITs is the propagation of good investor treatment norms. In this case, we would not expect to see positive FDI impacts associated with the ratification of any particular BIT. Major investment source countries, which have the most to gain by the global adoption of such a norm, may find their BIT promotion efforts more successful in countries that are competing heavily for FDI (which is consistent with the findings of (Elkins, Guzman and Simmons 2004), or to whom they are experiencing a growth in FDI outflow (which is consistent with my findings in Tables 5 and 6).

424 emma aisbett

These possibilities provide ample opportunity for future research on BITs. For now, I conclude that while BITs and FDI appear initially to be highly correlated, this finding is not robust to proper specification to account for the endogeneity of BITs and the time series properties of FDI data. This is important information for countries weighing the costs and benefits of beginning or expanding their participation in BITs and similar international investment agreements, as well as for other researchers undertaking international policy analysis.

bilateral investment treaties and foreign direct investment 425

appendix 1 Specification of bilateral FDI regressions

A.1. Choice of dependent variable Researchers wishing to analyze the impact of policies on bilateral FDI flows face an array of alternative empirical specifications, none of which has particularly good explanatory power, and none of which is tightly linked to theory. The theory of the determinants of FDI is at once too immature and too complex to be translated into structural econometric models (Blonigen 2005). The result is that there is very little consistency in the literature, even in terms of the appropriate dependent variable. It is common to find the use of affiliate sales, FDI stocks, or FDI flows (either in levels or in logs), FDI as a share of GDP, and FDI per capita. In light of the lack of strong indication from theory, it seems appropriate to base the choice of specification on statistical tests. This paper uses panel data on bilateral FDI flows from 29 OECD reporting countries to 46 partner countries over the period 1980–1998. The OECD database I use has both stocks and flows of inward and outward FDI. I use outward flows for the analysis because my interest is on the impact of BIT participation on lower income countries, not OECD countries. I use flows rather than stocks for two reasons. Firstly, there are significantly more missing values of the stocks. Secondly, stocks display a higher degree of autocorrelation and, for many countries, are far from stationary. We should be concerned about nonstationarity in FDI particularly because, as is often cited in empirical papers on this topic, world FDI has grown much more rapidly than either trade or GDP since the 1980s. This suggests that FDI is unlikely to be cointegrated with either of these controls, leading to the potential for spurious correlation with policy variables that are introduced over time. An example of the problem with using FDI stocks as the dependent variable is provided by (Blonigen and Davies 2004, p. 612) who find that: While the inclusion of fixed effects means residuals for any group of countries (such as rich ones) are zero on average, differing trends between groups may still remain. Specifically, over the time dimension of our sample, the rich countries average residuals become increasingly positive, while the poor countries average residuals grow increasingly negative. Having determined to use flows rather than stocks, the next choice is among levels, logs or some normalization of FDI flows. Logs have the disadvantage of losing zero and negative observations from the sample. The advantage of using logs is that the data are much less skewed than levels or normalized levels, meaning that when using logs, the results are less likely to be driven by a few influential data points. The severity of the skew in the data using levels of FDI flow or

426 emma aisbett Bilateral FDI Flows

Log Bilateral FDI Flows .15 Number of observations

Number of observations

.0015

.001

5.04-04

.1

.05

0

0 −5000

0

5000 10000 Bilateral FDI flow

15000

−15

5

10

1.5-04 Number of observations

Number of observations

−5 0 Log Bilateral FDI flow

Bilateral FDI Flow normalized by Host and Source GDP

Bilateral FDI Flow normalized by Host GDP 6.0-05

4.0-05

2.0-05

0

−10

−100000

0 100000 200000 300000 400000 Bilateral FDI normalized by Host GDP

1.0-04

5.0-04

0 −100000 0 100000 200000 Bilateral FDI Flow normalized by Product of Host and Source GDP

figure 5 histograms showing skewedness of alternative fdi measures

FDI normalized by host GDP, and the extent to which it is ameliorated by taking logs, is illustrated by the histograms in Figure 5. Though not reported here in order to save space, the histograms for alternative normalizations including FDI per capita and FDI normalized by the product of host and source GDPs look very similar to that for FDI normalized by host GDP. I use logs in the analysis to follow; however, for my sample the qualitative conclusions are essentially unchanged if levels or the share of FDI in GDP or population are used. This is to be contrasted with the results of (Hallward-Driemeier 2003), who finds no impact of BITs in specifications very similar to ones in which I do find a strong correlation between BITs and FDI. Hallward-Dreimeier uses a slightly different sub-sample of country-pairs to those I use. A possible explanation for the difference between her results and mine using FDI levels is that, due to the skew of the levels data, the results are highly sensitive to the sample used. A.2. General Specification for Bilateral FDI Having chosen a dependent variable, I next consider the appropriate set of controls for the regressions. The state of the art in theory-based empirical specifications for bilateral FDI is that proposed by (Carr et al. 2001) and applied to a similar policy question in two papers by (Blonigen and Davies 2002, 2004). Carr, Markusen and Maskus’s set of proposed controls includes the sum of host and source GDP, the squared difference between host and source GDP, the skill gap, the product of the difference in GDPs and the skill gap, trade costs for both

bilateral investment treaties and foreign direct investment 427

host and source, the square of the skill gap multiplied by the host trade costs, and a measure of the cost of FDI in the host. In their preferred specifications, Blonigen and Davies add country-pair fixed effects and in some cases rich country interaction terms. The importance of rich country interactions are highlighted by the finding of (Blonigen and Wang 2004) that the underlying factors that determine the location of FDI activity across countries vary systematically across least developed countries and developed countries in a way that is not captured by current empirical models of FDI. Since participation in BITs with OECD partners is mostly a lower income country phenomenon, I remove from the sample any recipient countries classified by the World Bank as high income.15 Summary statistics for these controls and others used in later regressions, as well as for bilateral FDI flows, are presented for the start and end years in Tables 1 and 2 respectively. The skill gap variable is proxied by the difference in average years education for adults over 25 years of age, taken from (Barro and Lee’s 2000) latest dataset. The specification of (Blonigen and Davies 2002, Column 2 of Table 5) is reproduced almost exactly in Column 2 of Table 8. Consistent with Blonigen and Davies I find that both the sum of host and source GDP and the square of the difference in GDPs are significant at the 1% level and have the expected sign (positive for the former and negative for the latter). In contrast to (Blonigen and Davies 2002), I find that several other variables are significant. Firstly, the skill gap is significant and has the “wrong” (i.e., negative) sign according to theory. This finding is consistent with (Blonigen et al. 2003) and (Blonigen and Davies 2004). The interaction of the skill gap with the GDP difference is also negative and significant, which is consistent with the predictions of the Carr, Markusen and Maskus knowledge-capital model. Finally, host trade share in GDP, which is my proxy for trade openness, is positive and significant. This would seem to support the dominance of export-oriented FDI over market-seeking FDI. The inconsistency between the results in column 2 and the theoretical predictions is a cause for concern; however, it is important to recall that the theory is one of long-run equilibrium FDI and is not designed for policy analysis. If we are interested in seeing how well the theory works at predicting long-run relationships, we may focus on the pooled ordinary least squares results in Column 1 of Table 8. Here, we see that both the sum of the GDPs and the skill gap have the anticipated (positive) sign and are significant at the 1% level. The GDP gap is positive and insignificant, but this is likely to be driven by high correlation with the skill gap. The time varying trade cost measures are also insignificant; however, truly exogenous measures of trade costs are provided by the geographical variables measuring the number of landlocked or island countries in the pair.

15. In regressions not reported here, high-income host countries were left in the sample and there was no qualitative impact on my results.

428 emma aisbett

table 8. limits of the knowledge capital specification for policy analysis Coefficient

Labels

(1) Infdi

(2) Infdi

(3) Lnfdi

(4) Lnfdi

(5) F.Lnfdi

IpJgdp

Sum of GDPs

ImJgdpsq

Squared Diff. GDPs

edga.p

Skill Gap

Itragdp

Host Trade Share in GDP

Jtragdp

Source Trade Share in GDP

edgapgdpdiff

Skill_gap∗GDP_difl.

Itragedgapsw

Host_trade∗Skill_gap3

lindi

No, Landlocked 0,1,2

1.101∗∗∗ (0.27) 0.00427 (0.064) 0.385∗∗∗ (0.11) -0.00020E (0,0074) 0.0147 (0.012) -0.000628 (0.029) -0.000230 (0.00024) 0.982∗∗ (0.50)

1.758∗∗∗ (0.16) -0.123∗∗∗ (0.034) -0.197∗∗∗ (0.076) 0.0156∗∗∗ (0,0032) 0.0114∗ (0.0065) -0,0792∗∗∗ (0.023) 0,000118 (0,00011) –

0.283 (0.17) -0.0288 (0,032) -0.141∗∗ (0,072) -0,00861∗∗∗ (0.0033) -0.000740 (0.0080) -0.0228 (0,023) 0.000165 (0,00011) –

0.283 (0,22) -0.0288 (0.042) -0.141∗ (0,074) -0,00861∗ (0,0044) -0,000740 (0.012) -0,0228 (0,028) 0.000165 (0.0001E) –

0.138 (0.2S) -0.0560 (0,044) -0.229∗∗∗ (0,078) -0,0121∗∗∗ (0,0046) 0,0197∗ (0.011) -0.0284 (0.028) 0.000217 (0.00014) –

No. Island 0,1,2

border

Land Bolder Dummy

colony

Colonial dummy

ldist

Log of Distance

Country-pair FE Year effects Cluster robust errors Lagged controls Observations R-squared Number of IJid

– – – – – – –

1.923∗∗∗ (0.49) -0.509 (0.81) 0.304 (0.74) -0.221 (0.35) – – – – 2098 0.29 –

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

–

Yes – – – 2208 0,18 281

Yes Yes – – 2208 0,30 281

Yes Yes Yes – 2208 0,30 281

Yes Yes Yes Yes 2317 0.33 287

Time-invariant controls in column 1 from Andrew Rose s website and defined as in Rose (2004) Standard errors in parentheses∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1

bilateral investment treaties and foreign direct investment 429

island

430 emma aisbett

table 9. performance of a gravity model alternative to the cmm knowledge capital specification Coefficient

Labels

(1) Lnfdi

(2) Lnfdi

(3) Lnfdi

(4) Lnfdi

(5) F.Lnfdi

lnJgdp

Source Log GDP

lnJpop

Source Log Population

lnIgdp

Host Log GDP

lnIpop

Host Log Population

edgap

Skill Gap

Itragdp

Host Trade Share in GDP

8.458∗∗∗ (0.67) −7.516∗∗∗ (0.66) 0.927∗∗ (0.42) −0.872∗∗ (0.38) 0.275∗∗∗ (0,098) −0.0165∗∗∗ (0,0063)

2,665∗∗∗ (0.40) 4.075∗∗ (2.06) 0.799∗∗∗ (0.26) 1.703∗∗ (0.77) −0.222∗∗∗ (0.072) −0,00930∗∗∗ (0.0035)

1.961∗∗∗ (0,43) 4.356∗∗ (2,19) 0.264 (0,26) −2.550∗∗ (0,99) −0,223∗∗∗ (0.072) −0.00914∗∗∗ (0,0035)

1,961∗∗∗ (0.59) −4.356 (3.26) 0.264 (0.40) −2.550 (1.69) −0.223∗∗∗ (0.075) -0.00914∗∗ (0.0042)

1.963∗∗∗ (0.56) −7.238∗∗ (3.24) 0.338 (0.38) −1.833 (1.76) −0.290∗∗∗ (0.083) −0.0137∗∗∗ (0.0045)

Source Trade Share in GDP

edgapgdpdiff

Skill_gap∗GDP_diE

Itragedgapsw

Host_trade∗Skill_gap2

Country-pair FE Year effects Cluster robust errors Lagged controls Observations R-squared Number of country-pairs

– – – – – – –

0.0248∗∗ (0,0099) 0.0501∗∗ (0,024) 0.000215 (0,00020) − – – – 2098 0.50 –

0.00920 (0.0063) −0.00200 (0.022) 0.000289∗∗∗ (0.00011) Yes – – – 2208 0,26 281

0.00292 (0,0080) −0,0263 (0.022) 0.000250∗∗ (0.00011) Yes Yes – – 2208 0,31 281

0.00292 (0.012) −0.0263 (0.027) 0,000250∗ (0.00015) Yes Yes Yes – 2208 0.31 281

0.0249∗∗ (0.011) −0,0336 (0.027) 0,000281∗∗ (0.00014) Yes Yes

Yes Yes 2317 0.34 287

Time-invariant controls included for column 1 but not reported are: number landlocked, number of islands, land border, colonial relationship and distance. Taken from Andrew Rose’s website and defined as in Rose (2004)Standard errors in parentheses∗∗∗ p<0.01, ∗∗ p<0.05, ∗ p<0.1

bilateral investment treaties and foreign direct investment 431

Jtragdp

432 emma aisbett

Both of these are significant and positive as predicted by a theory of market driven FDI. One interpretation of the results in column 2 of Table 8 and, therefore, of the results of Blonigen and Davies, is that the inclusion of country-pair fixed effects emphasizes the spurious correlation due to trends in both FDI and some of the control variables. Reference to Tables 1 and 2 shows that the significant coefficients in column 2 are all associated with the variables for which the mean changed the most between the start and end of the sample period. Further evidence that the results in column 2 are driven by spurious time-series correlation is provided by columns 3 and 4 of Table 8, which shows the impact of adding year effects to the regression and then additionally making the standard errors robust to clustering at the country-pair level. The only coefficient that remains significant in column 4 is the host share of trade in GDP, and it now has the opposite sign to that in column 2. Finally in Table 8, column 5 lags the explanatory variables to reduce simultanaiety bias. This is important given the large body of literature that claims to show that FDI drives growth. An alternative, and in some ways simpler, theory of FDI than the knowledge capital model has been proposed and tested by (Helpan, Melitz and Yeaple 2004). Their model is one of horizontal FDI in the context of monopolistic competition in differentiated products, with heterogeneous firms, and fixed costs of entry to the domestic market, additional fixed costs to exporting, and still higher fixed costs to FDI. FDI is driven by the desire to access foreign markets and avoid melting-iceberg trade costs. Helpan, Melitz and Yeaple develop and test this model with a focus on the cross-industry implications. As far as cross-country implications, they note only that the ratio of FDI to trade will be increasing in variable and fixed trade costs and decreasing in the fixed costs of engaging in FDI. These implications are standard to a model of horizontal FDI. It is easy to draw a number of other cross-country implications from the intuition of the Helpan, Melitz and Yeaple model. For example, FDI is aimed at supplying differentiated products, and the relative consumption of differentiated products tends to rise with income. Therefore, we would expect FDI to increase with per capita income of the host. Secondly, the most productive firms are the ones that engage in FDI. Since per capita income is a good measure of the average productivity of firms in a country, we may also expect bilateral FDI to increase with source per capita income. Similarly, for a given productivity distribution, a larger pool of firms implies a larger number of firms that will have productivity sufficiently great to be successful in FDI. To the extent that GDP is a measure of the number of firms in a country, we would also expect FDI to be increasing with the GDP of the host. Finally, the profit functions (Helpan, Melitz and Yeaple 2004, p. 302) suggest that profitability of FDI both in absolute terms and relative to exports is increasing with the size of the host market. Thus, we would expect bilateral FDI also to be increasing with the size of the host market.

bilateral investment treaties and foreign direct investment 433

table 10. first reporting year for source countries Country

Year

Australia Belgium France Denmark France Japan Netherlands Portugal Spain United Kingdom Austria Finland Germany Sweden Italy United States Canada New Zealand South Korea Norway Switzerland Iceland Poland Hungary Turkey

1980 1980 1980 1980 1980 1980 1980 1980 1980 1980 1981 1981 1981 1981 1982 1982 1983 1984 1985 1986 1986 1988 1993 1999 1999

Overall, some implications of the Helpan, Melitz and Yeaple model additional to those already in the Carr, Markusen and Maskus model are that the importance of host and source GDPs may not be symmetric, and that per capita incomes will play an important role. This suggests that the standard trade gravity model including the logs of GDP and income may be a good alternative to the sum of GDPs and squared difference in GDPs in the Carr, Markusen and Maskus specification. In order to avoid colinearity between GDP and GDP per capita in the log specification, I include a log population term together with log GDP. It is worth noting one further thing about the classic logarithmic gravity specification. When logs are taken, the ratio of the per capita GDPs is collinear with the product of the per capita incomes. The ratio of per capita incomes is a good

434 emma aisbett

proxy for the relative factor endowments that are important to vertical FDI. This means that the logarithmic gravity equation is flexible enough to accommodate vertically motivated FDI as well as horizontal. Table 9 shows the impact of the same stepwise refinements to the pooled ordinary least squares that are illustrated in Table 8 for the Carr, Markusen and Maskus model. The results are similar to Table 8 except that now two coefficients, source GDP and host trade share in GDP, are robust in sign and significance across specifications. The gravity model also has the advantage of showing the relative importance of source size and income compared to host characteristics. One concerning feature of the gravity model results in Column 5 of Table 9 is the fact that the magnitude of the negative coefficient on the source population is much larger than the magnitude of the GDP coefficient. This would imply that, conditional on a given GDP per capita, smaller source countries will have larger bilateral FDI flows. These coefficient estimates would suggest, for example, that Australia was a larger FDI source than the U.S. This is clearly not the case. Given the similar fit of the full Carr, Markusen and Maskus and gravity version, and the advantages of the gravity specification in terms of separating source and host effects, I will focus on the gravity specification in the analysis of the relationship between BITs and FDI. In the interests of space, the Carr, Markusen and Maskus results are not reported, as the qualitative conclusions are identical to those I find based on the gravity specification.

references Baldwin, R. and Taglioni, D. (2006). “Gravity for dummies and dummies for gravity equations,” NBER Working Paper, No. 12516. Baltagi, B. H. and P.X. Wu (1999). “Unequally spaced panel data regressions with AR(1)disturbances,” Econometric Theory, 15, pp. 814–823. Bhargava, A., L. Franzini and W. Narendranathan (1992). “Serial correlation and the fixed effects model,” Journal of Applied Econometrics, 7, pp. 243–257. Blonigen, B. A. (2005). “A review of the empirical literature on FDI determinants,” NBER Working Paper, No. 11299. Blonigen, B. and R. Davies (2004). “The effects of bilateral tax treaties on U.S. FDI activity,” International Tax and Public Finance, 11, pp. 601–622. Blonigen, B., R. Davies and K. Head (2003), “Estimating the knowledge-capital model of the multinational enterprise: Comment,” American Economic Review, 93, pp. 980–994. Blonigen, B. and M. Wang (2004). “Inappropriate pooling of wealthy and poor countries in empirical FDI studies,” NBER Working Paper, No. 10378. Carr, D., J. R. Markusen and K.E. Maskus (2001). “Estimating the knowledge-capital model of the multinational enterprise,” American Economic Review, 91, pp. 693–708. Doyle, C. and S. van Wijnbergen (1994). “Taxation of foreign multinationals: a sequential bargaining approach to tax holidays,” International Tax and Public Finance, 1, pp. 211–225.

bilateral investment treaties and foreign direct investment 435

Elkins, Z., A. Guzman and B. Simmons (2004). “Competing for capital: The diffusion ofbilateral investment treaties, 1960–2000,” UC Berkeley Public Law Research Paper, No. 578961. Guzman, A. (2005). “The design of international agreements,” European Journal of International Law, 16, 4, pp. 579–612. Hallward-Driemeier, M. (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and they could bite.” World Bank Policy Research Paper, WPS 3121. Helpman, E., M. Melitz and Y. Rubinstein (2005). “Trading partners and trading volumes,” Working Paper. Helpman, E., M. Melitz and S. Yeaple (2004). “Export versus FDI with heterogeneous firms,” American Economic Review, 91, 1, pp. 300–316. Janeba, E. (2002). “Attracting FDI in a politically risky world,” International Economic Review, 43, 4, pp. 1127–1155. Neumayer, E. and L. Spess (2005). “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33, 10, pp. 1567–1585. Nijman, T. and M. Verbeek (1982). “Nonresponse in panel data: the impact on estimates of a life cycle consumption function,” The Review of Economic Studies, 49, 4, pp. 533–549. Pistor, K. (2002). “The standardization of law and its effect on developing economies,” The American Journal of Comparative Law, 50, 1, pp. 97–130. Razin, A., Y. Rubinstein and E. Sadka (2003). “Which countries export FDI, and how much?” NBER Working Paper, No. 10145. Razin, A., E. Sadka, E and H. Tong. (2005). “Bilateral FDI flows: Threshold barriers and productivity shocks,” NBER Working Paper, No. 11639. Rose, A. (2001). “The determinants of foreign direct investment: sensitivity analyses of cross-country regressions,” Kyklos, 54, pp. 89–114. Rosenbaum, P. and D. Rubin (1983). “The central role of the propensity score in observational studies for causal effects,” Biometrika, 70, pp. 41–55. Salacuse, J. and N. Sullivan (2004). “Do BITs really work? an evaluation of bilateral investment treaties and their grand bargain,” Harvard International Law Journal, 46, 1. Tobin, J. and S. Rose-Ackerman (2004). “Foreign direct investment and the business environment in developing countries: The impact of bilateral investment treaties,” Yale Law School Center for Law, Economics and Public Policy Research Paper, No. 293. UNCTAD (2001). World Investment Report 2001: Promoting Linkages. New York and Geneva: United Nations. UNCTAD (2005), “Recent developments in international investment agreements,” IIA Monitor, No. 2. UNCTAD (2006a). “Developments in international investment agreements in 2005,” IIA Monitor, No. 2. UNCTAD (2006b). World Investment Report 2006: FDI from Developing and Transition Economies: Implications for Development. New York and Geneva: United Nations. Wooldridge, J. M. (2002). “Econometric Analysis of Cross Section and Panel Data,” The MIT Press, p. 34.

This page intentionally left blank

16. why do developing countries sign bit s ?∗ deborah l. swenson introduction Developing countries often compete for foreign investment with the hope that foreign direct investment (FDI) will bring a wide range of economic benefits.1 These benefits include increased levels of investment and economic activity, worker training, well-paid jobs and technology transfers that enhance the productivity of local firms.2 In addition, foreign investment may be viewed as a particularly attractive means of increasing developing country investment stocks since foreign investment is much less likely than other financial flows to leave the host country if the host experiences a financial crisis.3 Although multinational firms may have negative effects on host country markets, such as

∗ This chapter was originally published as “Why Do Developing Countries Sign BITs?,” 12 U.C. Davis Journal of International Law & Policy 131 (2005), copyright 2005 by The Regents of the University of California. All rights reserved. The chapter is reprinted with permission from the publisher. The author would like to thank Marina Nazarova for excellent research assistance, and the International Relations Program at UC Davis for research funding. 1. See Gordon Hanson, Should Countries Promote Foreign Direct Investment?, UN Doc. UNCTAD/GDS/MDPB/G24/9 (G-24 Discussion Paper No. 9, 2001), available at http:// www.unctad.org/en/docs/pogdsmdpbg24d9.en.pdf. Hanson discusses recent trends in investment promotion by developing countries. Hanson’s comparison of country investment promotion efforts uncovers a wide range in the techniques countries employ. While some countries offer exemption from corporate or value-added tax obligations, others offer exemptions from import duties, export processing zones, or subsidy packages tailored for multinational investors. 2. See Aitken et al., Spillovers, Foreign Investment, and Export Behavior, 43 J. of Int’l Econ. 103 (1997) (evidence on spillovers to domestic firms); Aitken et al., Wages and Foreign Ownership: A Comparative Study of Mexico, Venezuela and the United States, 40 J. of Int’l Econ. 345 (1996) (evidence on labor market effects). See also Giorgio Barba Navaretti & Anthony J. Venables, Multinational Firms in the World Economy (2004) for an overview of the extensive work on these topics. 3. See Robert E. Lipsey, The Role of Foreign Direct Investment in International Capital Flows (Nat’l Bureau of Econ. Res., Working Paper No. 7094, 2000) (providing evidence that FDI is much less likely to leave developing countries in time of financial crisis, when compared with portfolio capital investments).

438 deborah l. swenson

intensified competition with local firms, policy makers generally assume that FDI is beneficial on net.4 While countries may choose from a number of policy instruments as a means of increasing foreign investment inflows, the use of bilateral investment treaties (BITs) became especially common in the 1990s.5 By the beginning of the year 2000, 173 countries had signed at least one BIT and the total number of treaties had risen to 1857 from a base of 385 treaties a decade earlier.6 An increase in treaty signings by previous signers, as well as an increase in the number of participating countries, drove this expansion.7 Of the 173 countries that had signed treaties by 2000, seventy-one had not signed a single treaty before 1990.8 While Germany was the most frequent participant, having signed 124 separate agreements by the beginning of 2000, the average country had twenty-one separate BITs by that time.9 BITs generally contain provisions that touch on a common set of investment issues.10 After defining investment, BITs typically discuss the application of national and most-favored nation treatment to foreign investments.11 They may also include measures related to transparency of national laws, performance requirements, or the movement of foreign personnel.12 However, while BITs usually address a common set of topics, the content of

4. See UN Conference on Trade and Development, World Investment Report 2003: FDI Policies for Development: National and International Perspectives, UN Doc. UNCTAD/ WIR/2003 (Sept. 4, 2003), available at http://www.unctad.org/en/docs/wir2003_en.pdf [hereinafter WIR-2003] for a review of recent trends in this area. See also Brian J. Aitken & Ann E. Harrison, Do Domestic Firms Benefit From Direct Foreign Investment? Evidence From Venezuela, 89 Am. Econ. Rev. 605 (1999) and Beata Smarzynska Javorcik, Does Foreign Direct Investment Increase the Productivity of Domestic Firms? In Search of Spillovers Through Backward Linkages, 94 Am. Econ. Rev. 605 (2004) for analyses of domestic firms in Venezuela and Lithuania that show how the presence of multinational firms has multiple and conflicting effects on the performance of domestic firms. 5. See UN Conference on Trade and Development, Bilateral Investment Treaties 1959– 1999, UN Doc. UNCTAD/ITE/IIA/2 (Dec. 15, 2000), available at http://www.unctad.org/ en/docs/poiteiiad2.en.pdf [hereinafter BITs] (chronicling the acceleration in signing of BITs). 6. Author’s calculation based on treaty rosters included in BITs, supra note 5. 7. See WIR-2003, supra note 4, at 89 (claiming that the increase in the 1990s was fueled by treaty signing between developing countries). 8. Author’s calculation based on roster of treaty signings included in BITs, supra note 5. 9. Author’s calculation based on treaty rosters included in BITs, supra note 5. 10. See WIR-2003, supra note 4, at 89–91, which outlines the general elements contained in BITs. The later parts of the same chapter discuss investment liberalization accomplished through regional or multilateral means. 11. See WIR-2003, supra note 4, §§ A–B in chapter 4 (discussing the details on these questions). 12. See WIR-2003, supra note 4, at 89.

why do developing countries sign bit s? 439

BITs often differs significantly.13 Nonetheless, the element of BITs that may be of greatest interest to foreign investors is a county’s agreement regarding its obligations if a dispute arises in the future.14 While some domestic reforms may coincide with the interests of foreign investors, it may be difficult for host countries to persuade investors that domestic reforms will be implemented as promised.15 In particular, investors may be especially concerned about the permanence or strength of domestic reforms implemented in countries that have a higher level of perceived risk or endemic corruption.16 The dispute-settlement procedures contained in BITs may alleviate these concerns, since the dispute settlement provisions codify the forum and treatment of any future disputes.17 As a result, BITs are capable of providing a commitment mechanism that helps to reduce the amount of uncertainty foreign investors believe they face in a particular host country.18 The recent surge in foreign investment may certainly help to explain the increased motivation for countries to use BITs as a means of providing international investors with an attractive investment environment.19 As Figure 1.A shows, the amount of foreign investment in developing countries grew dramatically during the 1990s. More importantly, as the data in Figure 1.B demonstrate, foreign investment flows were increasingly large when compared with expenditures on gross capital formation in their developing country hosts. Compared with gross capital formation, FDI between 1970 and 1992 averaged a mere 3%.20

13. See UN Conference on Trade and Development, Bilateral Investment Treaties in the Mid-1990s, UN Doc. UNCTAD/ITE/IIT/7 (Jan. 7, 1999) for a detailed description of the measures included in BITs and the variations across BITs (noting that the content of BITs tends to vary even for treaties signed by a particular country with different partners). 14. See WIR-2003, supra note 4, at 114, which notes that most investors would receive protections through the host countries’ general laws of the land. However, investors often desire an internationalized approach. Here, the common form involves recourse to arbitration. 15. Eric Neumayer & Laura Spess, Do Bilateral Investment Treaties Increase Foreign Direct Investment to Developing Countries? (Nov. 2004) (unpublished manuscript, on file with author) [hereinafter Neumayer & Spess]. 16. Mary Hallward-Driemeier, Do Bilateral Investment Treaties Attract Foreign Direct Investment? Only a Bit, and They Could Bite (World Bank Policy Research, Working Paper Series No. WPS 3121, 2003), available at http://econ.worldbank.org/files/29143_wps3121. pdf [hereinafter Hallward-Driemeier] 17. See Neumayer & Spess, supra note 15. 18. See Kenneth J. Vandevelde, Investment Liberalization and Economic Development: The Role of Bilateral Investment Treaties, 36 Colum. J. Transnat’l L. 501 (1998) (discussing the use of BITs as a commitment mechanism). 19. See WIR-2003, supra note 4, at 89. 20. Author’s calculations based on data from Annex Table B.5. Inward and Outward FDI Flows as a Percentage of Gross Fixed Capital Formation, by Region and Economy, 1992–2003, in UN Conference on Trade and Development, World Investment Report 2004: The Shift Towards Services, UN Doc. No. UNCTAD/WIR/2004 (Sept. 22, 2004), available

440 deborah l. swenson

In contrast, FDI between 1993 and 2003 averaged 10.8% of gross capital formation.21 As a result, it is clear that FDI flows made a non-negligible contribution to the level of investment in developing countries.22 Surprisingly, many analyses exploring the economic effects of BIT signing have generally come to the rather discouraging conclusion that BITs are not associated with large increases in foreign investment. Hallward-Driemeier concludes, for example, that BIT signing only appears to have elevated FDI if one examines the share of a source country’s FDI attracted by a host.23 Further, she finds that such effects were only apparent five years after a treaty was signed.24 Hallward-Driemeier finds that other measures, such as FDI levels or FDI relative to GDP, were, if anything, negatively correlated with BITs.25 Tobin and RoseAckerman also come to the same disappointing conclusion: BITs do not appear to increase foreign investment flows, nor improve the characteristics of the local investment environment in signatory countries.26 In contrast, when Salacuse and Sullivan, and Neumayer and Spess use larger sets of countries to investigate these questions, they find a positive association between the number of BITs signed and the foreign investment received by a country.27 Neumayer and Spess also discover that the apparent boost provided by a BIT is bigger in countries that were characterized by greater risk, and hence likely to benefit more from the decision to sign a BIT.28

at http://www.unctad.org/en/docs/wir2004_en.pdf and UN Conference on Trade and Development, WIR Annex Tables Key Data, Inward FDI Flows as a Percentage of Gross Fixed Capital Formation, by Host Region and Economy, 1970–2004, http://www.unctad. org/Templates/Page.asp?intItemID=3277&lang=1 (data file downloaded Feb. 2004) [hereinafter WIR-2004]. 21. Author’s calculation based on data from WIR-2004, supra note 20. 22. Even if FDI took the form of acquisitions, the payments made by foreign acquirers would have increased financial resources in the developing country host, thus freeing domestic resources that could be channeled towards increased domestic investments in the host economy. 23. Hallward-Driemeier, supra note 16. 24. Id. 25. Id. 26. Jennifer Tobin & Susan Rose-Ackerman, Foreign Direct Investment and the Business Environment in Developing Countries: the Impact of Bilateral Investment Treaties, (Yale Law and Economics Research Paper No. 293, 2005), available at http://www.law.yale.edu/outside/html/faculty/sroseack/FDI_BITs_may02.pdf [hereinafter Tobin & Rose-Ackerman]. 27. Jeswald Salacuse & Nicholas Sullivan, Do BITs Really Work? An Evaluation of Bilateral Investment Treaties and Their Grand Bargain, 46 Harv. Int’l L.J. 67, (2005) [hereinafter Salacuse & Sullivan]. 28. Neumayer & Spess, supra note 15.

why do developing countries sign bit s ? 441

figure 1a 29 Foreign Direct Investment in Developing Countries ($Millions) 300000 250000 200000 150000 100000 50000 03 20

01 20

99 19

97 19

95 19

93 19

91 19

89 19

87 19

19

85

0

figure 1b 30 Foreign Investment as a Percentage of Developing Country Gross Capital Formation 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0

03 20

01 20

99 19

97 19

95 19

93 19

91 19

89 19

87 19

19

85

0.0

The fragility of the results in the previous literature on BITs leaves one to question whether BIT signing is in the best interest of a developing country. Since BITs effectively cause developing countries to relinquish some of their property rights to foreign investors,31 one may reasonably ask whether the countries receive an adequate level of benefits in return for their participation. To be certain, the recent surge in the signing of BITs represents a form of investment liberalization since the expansion of investment protections is designed to facilitate increased globalization through international investment.32 29. Data source is WIR-2004, supra note 20. 30. Id. 31. See, e.g., Bhagirath Lal Das, A Critical Analysis of the Proposed Investment Treaty in WTO (July 2003), Global Policy Forum, http://www.globalpolicy.org/socecon/bwiwto/wto/2003/07critical.htm (last visited Oct. 16, 2005) [hereinafter Das]. 32. See WIR-2003, supra note 4, at 89.

442 deborah l. swenson

What is notable is the contrast between this method of furthering international economic integration with earlier approaches for trade liberalization, which were largely based on the efforts of multilateral negotiations. While there have been calls for multilateral negotiations aimed at the concerns of international investors, multilateral efforts to liberalize the international investment environment have not borne fruit.33 One view that has inhibited progress toward the creation of a true multilateral investment agreement has been the view that signatories to such an agreement would provide foreign firms a host of economic benefits, while extracting no new concessions and imposing no new responsibilities on the beneficiary firms.34 As a result, while the European Union was successful in adding discussion about such an investment agreement to the Doha Round of WTO talks,35 many countries view such agreements with suspicion. Consequently, there is no multilateral protection of investment that is similar in scope to the WTO protections provided to trade in goods and services and the prospects for a multilateral solution are believed to be slim.36 In this policy void, BITs provide a piecemeal set of investment protections that augment investment provisions provided through some regional agreements.37 While economic models confidently predict that BITs should increase the level of foreign investment, the wide range of results discussed above suggests that the motives and the effects of BIT signing deserve further attention. To think about this question, and the paradoxical inability of academic studies to uncover consistent evidence that BITs are associated with investment increases,

33. See, e.g., The Global Policy Forum, Multilateral Agreement on Investment, http:// www.globalpolicy.org/socecon/bwi-wto/indexmai.htm (last visited Oct. 16, 2005). This web page is devoted to concerns about multilateral investment approaches and notes that OECD discussions of a multilateral investment agreement, which commenced in 1995, have failed to move toward any new agreement. 34. E.g., Das, supra note 31. 35. World Trade Organization, Ministerial Declaration of 14 November 2001, WT/ MIN(01)/DEC/1, 41 I.L.M. 746 (2002), available at http://www.wto.org/english/thewto_e/ minist_e/min01_e/mindecl_e.htm [hereinafter Doha Declaration]. The Doha Declaration includes statement 20, which reads: “ Recognizing the case for a multilateral framework to secure transparent, stable and predictable conditions for long-term cross-border investment, particularly foreign direct investment, that will contribute to the expansion of trade, and the need for enhanced technical assistance and capacity-building in this area as referred to in paragraph 21, we agree that negotiations will take place after the Fifth Session of the Ministerial Conference on the basis of a decision to be taken, by explicit consensus, at that session on modalities of negotiations.” 36. See, e.g., WIR-2003, supra note 4, at 91 (commenting on the fact the efforts to negotiate multilateral investment agreements have always failed, even when the agreements were non-binding on countries). 37. A more limited set of WTO measures included in the Trade Related Investment Measures (TRIMS) and the General Agreement on Trade in Services (GATS), help protect some forms of investments.

why do developing countries sign bit s ? 443

I study two questions. First, I examine the correlation between previous foreign investment and the signing of BITs to explore whether there is any evidence that the signing of BITs is investor-driven. Such a correlation would imply that BIT signing is affected not only by the national governments of the signatory nations, but also by the interests of the investing firms who wish to gain further protections for their assets. This link may provide insight into the findings that BITs are only sometimes related to increases in foreign investment, since BITs signed to meet the interests of existing investors may not lead to the receipt of future investments. Next, I study how BITs affected the flow of investments between countries when a wide range of controls for the economic environment, such as home and host GDP, wage rates, and risk measures are considered. Unfortunately, while these variables are certainly primary determinants of investment, it is impossible to control for all country characteristics that cause firms to invest in one country versus another. Rather than adding further control variables, I add previous foreign investment as a control in my regression analysis. Such a control helps to provide a direct indicator of country attractiveness that is otherwise unobservable through direct measures.38 A. Which Countries Sign BITs? Although progress toward a multilateral investment agreement remains stalled, an increasing portion of foreign investment benefits from investment protections that are provided by BITs. According to Hallward-Driemeier, by 2000 onehalf of the foreign investment developing countries received from investors in Organization for Economic Cooperation and Development (OECE) countries was covered by the promises and protections conferred by BITs.39 Since the benefits of BIT signing are as yet difficult to identify, one may ask whether BIT signers receive an adequate exchange for their promises to foreign investors. To address this question, it is important to begin by analyzing which countries chose to sign BIT agreements. Elkins, Guzman and Simmons, and Neumayer have contributed to this area by using Cox analyses of the data that examine the timing of treaty signing.40 In both papers, the authors ask what determines the amount of time that elapses

38. While a researcher can directly control for education, for example, the researcher can not directly control for the effectiveness of the educational system. However, a country that has an effective education system that produces skilled workers is likely to attract investment now and in the future. 39. Hallward-Driemeier, supra note 16. 40. Zachary Elkins et al., Competing for Capital: The Diffusion of Bilateral Investment Treaties, 1960–2000 (UC Berkeley Public Law Research Paper No. 578961, 2005) [hereinafter Elkins]; Eric Neumayer, Own Interest and Foreign Need: Are Bilateral Investment Treaty Programs Similar to Aid Allocation? (Jan. 2005) (unpublished manuscript, on file with the author) [hereinafter Neumayer].

444 deborah l. swenson

between the signing of BITs. While Neumayer focuses on the economic characteristics of the investor and recipient countries,41 Elkins, Guzman and Simmons examine whether competition with other countries influenced the motives of countries.42 To further this debate, I will address an alternative possibility: countries signed BITs due to the interest of foreign investors rather than due to pressures exerted by local constituencies or the altruism of the investing nations. I will also change the frame of the debate by examining the overall frequency of treaty signing, rather than the amount of time that elapsed between treaty signings. The motive for this switch is to examine the factors that cause some countries to pursue a larger number of treaties than others. B. Who signs BITs? Empirical results Since countries differ so dramatically in their propensities to sign BITs, this section first asks whether country characteristics can predict the signing of BITs. Next, the analysis is augmented to examine whether the interests of existing foreign investors had any influence on the signing of BITs. To answer these questions, I examine whether the number of BITs signed by developing countries was related to economic conditions of the countries, or to the investment positions taken by foreign investors. The dependent variable is the number of new BITs signed. By 2000, countries that had signed BITs had an average number of 21.2 treaty signings each.43 However, there was a great deal of heterogeneity across countries in their treaty signing proclivities. In particular, since the median number of treaties signed was 14.5, we can infer that some countries signed a great number of treaties, whereas others were much less active.44 Because the number of treaties signed is always non-negative, I use a negative binomial estimation framework.45 In addition, since treaty signing isn’t a frequent event, the dependent variable is measured at two different periods of time, 1990–1994 and 1995–1999. To gain insight into the importance of country characteristics, Table 1 displays results from a simple analysis that relates the number of treaties signed to country per capita income, country risk, and the region where a country is located. Country risk is included, since one would expect that countries might sign BITs to provide assurances to foreign investors when external assessment deems the country as having a level of risk that might otherwise deter foreign investors from entering the country.46 Said differently, a country that is characterized by

41. Neumayer, supra note 40. 42. Elkins, supra note 40. 43. Author’s calculation based on roster of treaty signings included in BITs, supra note 5. 44. Author’s calculation based on roster of treaty signings included in BITs, supra note 5. 45. Standard regression models assume that the dependent variable can take on both positive and negative values, which are not seen in this context. 46. See Hallward-Driemeier, supra note 16.

why do developing countries sign bit s ? 445

table 1. economic determinants of bit signing Dependent Variable

Per Capita Income Risk

Region Effects South America Africa CEE Asia Constant Log Likelihood Number of Countries

Number of BITs Signed 1990−1994

1995−1999

.729∗∗∗ (.246) .013 (.022)

419∗∗∗ (.151) .022∗ (.014)

.423 (.386) −.169 (.449) 1.11 (.390) .937 (.447) −.890 (1.39) −258 90

−.227 (.257) −.415 (.272) .345 (.263) .450 (.286) .615 (.826) −301 90

Notes: Negative binomial regressions. Standard errors in ( ); ∗∗∗, ∗∗, and ∗ represent significance at the 10, 5 and 1 % levels, respectively.

a high level of risk or corruption might not be able to implement domestic reforms that succeed in attracting international investors, since investors might worry that the reform would not be permanent, or if permanent, that the reform might not be enforced. Per capita income is included, since treaty signing is found to be positively related to country incomes.47 Consequently, I expect to find a positive relationship between both per capita income and country risk and the signing of BITs. Finally, regional dummy variables are included in the analysis to capture regional differences in institutions or resources, which might predispose local governments toward or against the signing of BITs. The regression coefficients on these regional variables will take on either positive or negative signs, indicating whether the region

47. See Neumayer, supra note 40.

446 deborah l. swenson

was more or less active in BIT signing than were countries in other regions of the world. For example, some countries might have been less likely to sign BITs since the region had already provided credible investor protections through a regional trade agreement, or an alternative mechanism such as Friendship, Commerce and Navigation treaties that included investment provisions. Regional dummy variables may also capture the essence of competitive BIT signing that the results of Elkins, Guzman, and Simmons suggest.48 Specifically, countries may have decided to sign a BIT when they were concerned about retaining foreign investments that might otherwise relocate to neighbors that had recently signed their own BITs with key investing countries. Since neighboring countries within a region are likely to resemble each other to a greater degree than geographically distant countries, regional variables help to capture this aspect of competition.49 The results in Table 1 confirm the basic predictions. During the 1990s, highincome countries, and those countries that were viewed as more risky exhibited a propensity for entering into a larger number of BITs. The results also show that there were marked regional differences in country propensities for signing BITs. Throughout the 1990s the highest probability of signing was observed in the developing countries of Asia. While this may indicate that BIT signing was driven by cross-country competition in Asia, there may have been characteristics that were present in Asian economies that caused the Asian countries to sign more treaties than other countries did. If multinational firms can convince countries that the firms are footloose and prepared to relocate their operations to countries that provide a superior investment climate, pressure from multinational investors may influence whether countries sign BITs.50 Assuming that the stock of current investors propels the impetus towards investor-driven BIT signing, we can test whether the number of BITs signed is positively related to the stock of investments countries had already received. In particular, if country decisions were based on the fear of relocation by multinational firms, countries that had amassed greater foreign investment stocks would be more inclined to sign a BIT than countries that were economically similar but had yet to attract significant levels of foreign investment. To test whether BIT signing was related to previous foreign investment, the regressions displayed in Table 2 add a country’s foreign investment stock at the beginning of the period as an explanatory variable. The foreign investment stock

48. Elkins, supra note 40. 49. See Deborah L. Swenson, The Effect of U.S. State Tax and Investment Promotion Policy on the Distribution of Inward Direct Investment, in Geography and Ownership as Bases for Economic Accounting 285 (J. David Richardson, Robert Lipsey & Robert E. Baldwin eds., 1998) (showing that one identifies bigger effects of taxes on foreign investment in the U.S., if one controls taxes of competing neighbor states). 50. See Elkins, supra note 40.

why do developing countries sign bit s ? 447

table 2. economic determinants of bit signing, including fdi stocks Dependent Variable

FDI Stock in beginning Per Capita Income Risk

Region Effects South America Africa CEE Asia Constant Log Likelihood Number of Countries

Number of BITs Signed 1990−1994

1995−1999

.387∗∗∗ (.110) .400 (.294) .010 (.025)

.165∗∗∗ (.056) .380∗∗ (.150) .029∗∗ (.012)

.660 (.539) .026 (.574) 1.97∗∗∗ (.756) .482 (.597) −2.98∗∗ (1.59) −178 69

−.787∗∗∗ (.261) −.853∗∗∗ (.270) .021 (.263) −.364 (.310) −.232 (.800) −266 82

Notes: Negative binomial regressions. Standard errors in ( ). ∗∗∗, ∗∗, and ∗ represent significance at the 10, 5 and 1% levels, respectively. FDI stock for 1990–94 regression is stock in 1990. The FDI stock in 1995 is used for the regression describing 1995–99 BITs signed.

accumulated by 1990 was added to the regression describing country treaty decisions for 1990 to 1994, while the foreign investment stock amassed by 1995 was added to the regression analysis for 1995 to 1999. The results in Table 2 show that countries that attracted more foreign investment in previous years were much more likely to sign BITs. This result suggests that concern that foreign firms might relocate if the current host government failed to sign its own BIT motivated BIT signing, at least in part. Such an outcome is plausible, since foreign investors may become more effective at lobbying host governments after they make their investments due both to their economic threat to leave, and to insights about the political processes in the host country that are gained as the firm operates in the host country. Notably, the economic significance of

448 deborah l. swenson

the other variables, per capita income and country risk, are unchanged when the regression is augmented with foreign investment stocks.51 Since BIT signing appears to be correlated with the presence of foreign investment, one may also ask whether countries signed BITs as a means of gaining contemporaneous FDI flows. While BITs may have had no observable effect on future foreign investment flows, it may be the case that countries signed BITs in order to secure immediate foreign investments that were realized at the time of BIT signing. To consider this issue, a final variable supplements the analysis: FDI flows that occurred during the period of BIT signing. Table 3 shows the new results. Once again, the general economic variables introduced earlier continue to perform as before. BIT signing, country per capita incomes, and country risk are still positively related. However, the new results show that the inclusion of contemporaneous FDI flows provides additional explanatory value to the regressions. Countries that were in the midst of FDI surges were more likely to sign BIT agreements. In summary, the data for the 1990s show that the signing of BITs was positively correlated with previous investment levels. Such a correlation suggests that countries may have agreed to sign these treaties since foreign investors located in their borders were lobbying for the investor protections they could gain from BITs. In addition, BITs depend on country characteristics, as the number of treaty signings was positively correlated with per capita income and country risk. C. Do BITs Increase Foreign Investment? Conceptual Issues Basic economics suggests that, all else being equal, countries that sign BITs will be rewarded with increased levels of foreign investment. This belief is based on the economic premise that firms seek to maximize their economic returns, and on the assumption that BITs will help firms to protect their profits and to reduce their uncertainty about the application of laws to their investments.52 The fact that the review of foreign investment data comes to contradictory findings may say something about the assumptions that underpin this simple view. To address this problem, this section reviews timing issues that may affect the measurement of BIT effects. The empirical section will then use the discussion of timing issues to guide the implementation of the data analysis. The evidence from Table 3 shows that foreign investment rose contemporaneously with BIT signing. However, since signing a BIT implies that a country has agreed to a set of ongoing obligations, it is reasonable to ask whether such countries were rewarded with increases in foreign investment in later years.

51. The regressions reported in Tables 2 and 3 have fewer observations than the regressions in Table 1. This is because foreign investment is included as a determinant of BIT signing in Tables 2 and 3. As a result, developing countries that failed to receive foreign investment in earlier years were not included in the regression. 52. See Neumayer & Spess, supra note 15.

why do developing countries sign bit s ? 449

table 3. economic determinants of bit signing, including fdi stocks and flows Dependent Variable

FDI Stock in beginning FDI Flow in period Per Capita Income Risk

Region Effects South America Africa CEE Asia Constant Log Likelihood Number of Countries

Number of BITs Signed 1990−1994

1995−1999

429∗∗∗ (.098) 702∗∗∗ (.213) .506∗ (.272) .022 (.022)

.248∗∗∗ (.069) .353∗∗ (.148) .302∗∗ (.157) .026∗∗ (.012)

.528 (.488) .057 (.519) .355 (.778) .172 (.539) −4.34 (1.50) −173 69

.844∗∗∗ (.255) −.764∗∗∗ (.266) −.105 (.265) −.492 (.305) −.902 (.827) −260 81

Notes: Negative binomial regressions. Standard errors in ( ). ∗∗∗, ∗∗, and ∗ represent significance at the 10, 5 and 1% levels, respectively. FDI stock for 1990–94 regression is stock in 1990. The FDI stock in 1995 is used for the regression describing 1995–99 BITs signed. FDI flow is the contemporaneous change in FDI.

As was discussed in the introduction, work on this question has come to mixed conclusions. While Salacuse and Sullivan, and Neumayer and Spess find evidence that BITs appeared to facilitate subsequent foreign investment flows,53 work based on a smaller set of host countries conducted by Hallward-Driemeier and

53. Salacuse & Sullivan, supra note 29; Neumayer & Spess, supra note 15.

450 deborah l. swenson figure 2. identifying economic effects of bit s FDI BIT EFFECT

BIT SIGNED

TIME

Tobin and Rose-Ackerman comes to the opposite conclusion.54 To revisit this question I use a large set of developing countries and propose two ways of reframing the question in hopes of better identifying the effects of BITs on foreign investment flows. Figure 2 represents the simple view of BIT effects. Here, a country signs a BIT, and the path of FDI flows rises immediately upon the country’s signing. The investment response shown in Figure 2 is based on the idea that investor uncertainty about a country’s treatment of foreign investments is resolved on the date of the signing. And the rise in investment that is observed at the time of signing is measured as the effect of the BIT provisions on foreign investment. Figure 2 conveys the basic idea that lies behind the econometric analysis: The effect of BITs is to be measured by examining how foreign investment flows change when a country signs a BIT. However, the simple framework fails to capture at least two phenomena that are likely to influence the economic value that is attributed to BIT signing. The first caveat relates to issues of timing. Although Figure 2 assumes that all investors change their expectations about a host country’s treatment of foreign investment on the date when a treaty is signed, it is easy to imagine that the time pattern of investment responses will unfold much less neatly than those portrayed in Figure 2. In some cases, if BIT signing is expected, it is likely that investors may invest in the host country before the BIT signing takes place, since the investors confidently anticipate that their current investments will receive further protections shortly after the signing occurs. In this scenario, the level of foreign investment jumps to a higher level before the BIT is signed, as is shown by the dotted line, Anticipation, in Figure 3. 54. Hallward-Driemeier, supra note 16; Tobin & Rose-Ackerman, supra note 28. While working with a larger set of countries implies that the data set is more comprehensive, the smaller set of countries included in the work by Hallward-Driemeier or Tobin and RoseAckerman attracted the bulk of all foreign investments. As a result, the difference in the two sets of results may mean that the effects of BITs were greatest for the countries that received the least amount of investment. While this means that BITs may have helped these countries to achieve a noticeable percentage increase in the foreign investment they received, the economic value of these investments may have been miniscule, since they represented a large percentage increase applied to a tiny base.

why do developing countries sign bit s? 451

figure 3. leads and lags in investment reactions Delay FDI Anticipation

BIT SIGNED

TIME

In contrast, investors may doubt the resolve of other host countries to sign a BIT. The investors may be concerned that the ultimate treaty will fail to include all the provisions that are of great importance for their particular investment. In this scenario, such investors will wait to verify that a treaty with all vital provisions has been signed before they make their investment decision. However, once the treaty has been signed, investment may be delayed further, as the investor takes time to plan the details of the investment and to begin its implementation. In this second scenario, the BIT does cause investors to elevate their investment. However, the actual response is delayed relative to signing, as shown by the higher dotted line, Delay, in Figure 3. Since timing issues are likely to affect all investors, it is unlikely that foreign investment will jump on the date of treaty signing, as is illustrated in Figure 2. Further, it is important to remember that timing issues may not be the same for all firms, even if the firms are evaluating the same host country. To deal with the timing and account for the possibility that investments may be characterized by anticipation or delay relative to the date on which a BIT is signed, I use data that are grouped into two time periods, 1990–1994 and 1995–1999, rather than examining changes that occurred at yearly intervals. Another data issue that influences the measured importance of BITs relates to the identity of BIT partners. Consider, for example, a developing country that first signs a BIT with the U.S., and later signs a BIT with Belgium. While the U.S. and Belgium are both rich countries whose firms often engage in foreign investment, U.S. foreign investment is much greater than Belgian foreign investment. As a result, a host country is likely to see its overall receipt of foreign investment rise considerably after it signs a treaty with the U.S. The same country will experience further foreign investment gains when it signs a treaty with Belgium. However, the magnitude of those gains is likely to be much smaller, since Belgium is a much smaller player in the market for foreign investment.55 55. Data on outward investment in services in WIR-2004, supra note 20, at 306–307, show that Belgium’s investments for 1995–1999 were less than 3% as large as foreign investments by U.S. investors during the same interval.

452 deborah l. swenson figure 4. the effect of country identity of bit effects FDI

BIT SIGNED: with U.S.

with Belgium

TIME

To convey these differences, Figure 4 illustrates the differential effects that come from signing a BIT with the U.S. versus signing a similar treaty with Belgium. Both treaties elevate the level of foreign investment, though the increase following the U.S. treaty signing is larger. Tobin and Rose-Ackerman also considered the importance of country identity, as they sought to determine whether the effects of BIT signing with low-income countries were quantitatively different from the effects of BIT signings with high-income countries.56 They found that countries that signed more BITs attracted a greater share of world FDI.57 However, while they found that a higher number of BIT signings with high-income countries sometimes correlated with receipt of higher foreign investor shares, the result was somewhat sensitive to the choice of regression specification.58 D. Economic evidence from the 1990s on the value of signing BITs Tables 4 and 5 explore how BIT signing affected the investment received by developing countries. Each of the regressions includes regional variables to capture emerging trends in foreign investment. For example, to the extent that foreign investment in the 1990s helped to facilitate outsourcing of electronics industries in Asia, one might expect Asia to receive a higher than average amount of foreign investment. Alternatively, to the extent that South American economies were emerging from the problems of their 1980s debt crises, one might predict that they would have benefited from increased foreign investment inflows. In any case, the inclusion of regional variables helps to control for trends in investment that were driven by otherwise unobservable factors that were common at the regional level. These include, among other factors, commonality in

56. Tobin & Rose-Ackerman, supra note 28. 57. Id. 58. Id.

why do developing countries sign bit s? 453

table 4. the effect of bit signing on fdi flows: all bit s Dependent Variable

New BITs FDI Flow, prev period Risk

Region South America Asia CEE Africa Constant Adjusted R2 Observations

Foreign Direct Investment Flow 1990−1994

1995−1999

.002 (.091) −.040 (.149) .011 (.011)

.143∗∗ (.069) .186∗∗ (.076) .010 .173

−076 (.257) .155 (.298) 2.37∗∗∗ (.831) −.243 (.259) .445 (.403) .063 80

.192 (.177) .029 (.195) .147 (.278) .061 (.188) −.111 (.298) .129 84

Notes: Negative binomial regressions. Standard errors in ( ). ∗∗∗, ∗∗, and ∗ represent significance at the 10, 5 and 1% levels, respectively.

resources and distance from customer markets. The coefficients on regional variables may also indicate that there were similarities in the policy environment, which arose due to commonality of culture and tradition. Of course, there is much additional heterogeneity among countries within a region. Unfortunately, identifying all country differences that influence the level of foreign investment is difficult, since it is virtually impossible to collect a comprehensive data set that includes information on every country characteristic that is of interest to foreign investors. However, as Lipsey shows, there is a high level of persistence in foreign investment decisions.59 In particular, some 59. Robert E. Lipsey, The Location and Characteristics of U.S. Affiliates in Asia (Nat’l Bureau of Econ. Res., Working Paper No. 6876, 1999).

454 deborah l. swenson table 5. the effect of bit signing on fdi flows: bit identity Dependent Variable

New U.S. BIT New non-U.S. BITs FDI Flow, prev period Risk

Region South America Asia CEE Africa Constant Adjusted R 2 Observations

Foreign Direct Investment Flow 1990−1994

1995−1999

−.057 (.440) .009 (.093) −.043 (.151) .011 (.011)

.943∗∗∗ (.309) .125∗ (.066) 194∗∗∗ (.073) .008 (.007)

−.072 (.263) .152 (.299) 2.365∗∗∗ (.836) −.239 (.262) .433 (.407) .050 80

.124 (.170) .084 (.187) .187 (.264) .089 (.179) −.067 (.284) .211 84

Notes: Negative binomial regressions. Standard errors in ( ). ∗∗∗, ∗∗, and ∗ represent significance at the 10, 5 and 1% levels, respectively.

countries attract high levels of foreign investment from one year to the next, whereas other countries repeatedly fail in their pursuit of foreign investment.60 To control for ongoing differences in country attractiveness, I include previous years’ foreign investment as an explanatory variable in the regressions, as an indirect means of capturing the unobservable factors that affected foreign investment flows. This means that the regression is designed to show whether a country’s decision to sign more BITs caused the country to receive a greater level

60. Id.

why do developing countries sign bit s? 455

of foreign investment in the current period, as compared with predictions based on their previous receipt of foreign investment. The answers provided by the results presented in Table 4 are mixed. In particular, the results suggest that new BIT signing in the early 1990s was not correlated with increased levels of investment, whereas BIT signing in the late 1990s was. It is possible that differences in the business cycle were responsible for the differences in the effects observed in the two periods. In particular, foreign investment was depressed in the early 1990s,61 and as a result, countries may have experienced declining investment in the early 1990s even if they agreed to sign new BITs. In contrast, world FDI flows rose considerably in the late 1990s,62 so multinational firms, who were increasing their foreign investment presence at this time, may have been more responsive to cross-country differences in the BIT environment provided by host country governments. The analysis presented in Table 5 explores whether there were notable differences in the value of signing a BIT with the U.S. To do so, the BIT variable was split into two unique variables. The first component takes on a value of zero or one, depending on whether the host country signed a BIT with the U.S. The second component is the number of BITs signed with other countries. As in Table 4, BITs were not found to foster investment in the early 1990s, whereas all BIT expansion was correlated with increased investment in the late 1990s. However, BIT signing with the U.S. was correlated with a much larger stimulus to investment flows. In fact, the relative magnitude of the estimated coefficients suggests that countries who signed a BIT with the U.S. received a boost to foreign investment that was seven and a half times larger than the boost experienced when the average BIT was signed with other countries. Nonetheless, the positive coefficient on the number of non-U.S. BITs is likely to mean that countries that actively sought to provide a more stable environment using BITs were also actively promoting investment through other channels. E. Interpretation and Further Issues There are other factors that may influence the results in this study as well as factors that deserve attention in future studies. One of the first is the effect of alternative investment promotion measures. In this study, as in others, the only investment tool that is included in the regressions is the number of BITs. However, countries certainly have other avenues for protecting the interests of foreign investors.63

61. See WIR-2003, supra note 4 (annex tables on investment document the general decline in investment in the early 1990s). 62. See WIR-2003, supra note 4 (annex tables show the rapid increase in foreign investment in the late 1990s). 63. See WIR-2003, supra note 4, at 89–93 (describing how foreign investment could be protected under provisions of regional trade agreements, or ways that multilateral agreements could accomplish the task).

456 deborah l. swenson

As a result, a more comprehensive treatment of the data might be achieved if the regressions were augmented with variables that indicated whether investors could receive protection from elements of regional trade agreements or from the provisions of Maritime and Friendship treaties.64 While work by Yeyati, Stein and Daude suggests that regional agreements have influenced foreign investment, there is no study that looks at the full set of investment measures at the same time.65 Another reason why BITs may appear to have been less effective in stimulating investment than they actually were is that large cross-country analyses necessarily involve generalizations that are unable to capture the full degree of heterogeneity in choices and outcomes. To begin, although some BITs provide more comprehensive investor benefits than others, the regression analysis is limited to noting whether a BIT was in place or not. Unfortunately, the creation of a more tailored variable that ranked BITs by value, such as a one to ten scale, would be difficult. First, one would have to assemble a large data set that indicated which details were present or absent from each of the BITs signed. More difficult yet, one would have to decide how to weight the different elements of BITs. Given the diversity in procedures and enforcement, it is difficult to imagine that one could easily assign weights to varying sets of dispute settlement measures, thus enabling easy comparisons of the treatment of foreign nationals in other countries. Ultimately, an index measure would require arbitrary judgment calls about the weights. What is worth remembering, when interpreting this study or other work that examines how investment responds to BITs, is that the use of BIT counts relies on an imprecise measure of the investment environment. As such, the analysis is less likely to observe results, even if BITs are effectively fostering new investment. A third element of the data that may affect the results is that all foreign investment is grouped together. However, it is possible that some investments will be more responsive to BITs than others. For example, investors deciding where to build a manufacturing facility may be able to choose from a large number of countries, and the availability of BIT protections may influence their ultimate choice. In contrast, natural resource investors are constrained to locate in countries that offer the appropriate resource. While investors might hope for BIT protection, they will choose to invest in the resource-rich countries, regardless of whether an investment treaty has been signed. As a result, grouping all invest-

64. See Subramanian & Wei, The WTO Promotes Trade, Strongly But Unevenly (Nat’l Bureau of Econ. Res., Working Paper No. 10024, 2003) (showing that uncovering the trade-promoting effects of the WTO requires careful attention to other trade promotingagreements, including regional trade agreements, or aternatively, application of the generalized system of preferences). 65. Eduardo L. Yeyati et al., Regional Integration and the Location of FDI (Inter-Am. Dev. Bank, Working Paper No. 492, 2002), available at http://www.iadb.org/res/publications/ pubfiles/pubwp-492.pdf.

why do developing countries sign bit s ? 457

ments together provides an estimate of the average response to BITs. However, this approach misses the subtle differences across different investment sectors.66 It would be interesting for future researchers to explore the differential effects of BITs on manufacturing, service and resource investments. It would also be interesting to learn whether the effects of BITs differ by investment type, having more or less effect on acquisition FDI than they have on greenfield investments.

conclusion A remarkable expansion in the number of signings of BITs marked the 1990s.67 On the face of it, this expansion in the number of treaty obligations presents a puzzle, since a number of studies have come to question whether the receipt of new foreign investment flows rewarded signatory countries.68 I find two factors that are likely to influence the observed benefits of country decisions to enter into BITs. First, although treaties are viewed as forward-looking tools that are signed to gain future investments, I show that treaty signing also had a backward-looking element. In particular, countries that had already received larger stocks of foreign investment were more likely to sign BITs than were countries that had been less successful in attracting foreign investment. This result suggests that the interest of exiting foreign investors drove the signing of BITs, at least in part. Nonetheless, the signing of BITs may have been in the interest of individual countries if the signing of BITs allowed countries to retain investments that otherwise might have relocated to another country. However, rather than fostering new investments, BITs may have enabled these countries to hold onto previous investments. The second conclusion of this chapter is that controls for timing, intrinsic country attractiveness and investor identity are all important. When these issues are addressed, data from the late 1990s suggests that BIT signing did help developing countries attract a larger volume of foreign investment.

66. Compare Deborah L. Swenson, Investment Distinctions: The Effect of Taxes on Foreign Direct Investment in the U.S., in The Welfare State, Public Investment, and Growth (Hirofumi Shibata and Toshihiro Ihori eds., 1998), and Deborah L. Swenson, Transaction Type and the Effect of Taxes on the Distribution of Foreign Direct Investment in the U.S., in International Taxation and Multinational Activity (James R. Hines, Jr. ed., 2001) (showing how different classes of investment are differentially affected by host location taxes), with Keith Head et al., Attracting Foreign Manufacturing Investment Promotion and Agglomeration, 29 Regional Sci. & Urb. Econ. 197 (1999) (showing how careful treatment of the full set of tax provisions provides cleaner estimates of the effect of taxes on investment decisions). 67. See WIR-2003 supra note 4 (noting the remarkable growth of BITs after 1989). 68. See, e.g., Hallward-Driemeier, supra note 16; Tobin & Rose-Ackerman, supra note 28 (neither find consistent evidence suggesting that BITs foster foreign investment in developing country hosts).

This page intentionally left blank

part three exploring the impact of double taxation treaties on foreign direct investment flows

This page intentionally left blank

17. do bilateral tax treaties promote foreign direct investment?∗ bruce a. blonigen and ronald b. davies introduction Economists have long been concerned about the effect of taxation on foreign direct investment (FDI). A plethora of studies have examined whether and to what extent FDI responds to tax incentives. Although the specific results vary, the general consensus is that firms do indeed respond to a variety of tax policies.1 This issue is of primary concern to economists because inappropriate tax policies can result in inefficient allocation of investment across countries. As governments use their tax policies to affect the rates of return on capital, provide public goods, or simply capture part of the profits that would otherwise be repatriated to other countries, tax policies can divert investment from its most productive locations. One potential method of relieving this inefficiency is a bilateral tax treaty on FDI. These treaties adjust the tax environment for investment between treaty partners by specifying the applicable tax base, the withholding taxes that can be applied, and other measures affecting the taxation of FDI. Worldwide, over 2,000 of these treaties are in force and they govern the taxation of the large majority of FDI (Radaelli 1997). Since treaties indicate cooperative taxation by treaty partners, many economists assume that treaties increase investment. However, it is by no means certain that treaties do so. In particular, since treaties can reduce tax avoidance and other tax-saving strategies by firms, they might actually have a dampening effect on FDI. Furthermore, legal researchers have questioned whether FDI promotion is even a primary goal of treaty formation. In light of these conflicting arguments, the effect of tax treaties on FDI is an open question. This chapter presents empirical results on treaty formation by OECD members. Our findings suggest that, at least for recent treaties, treaty formation most likely does not increase FDI activity between members and may in fact decrease it. Thus, our results stand at odds with the FDI promotion rationale for treaty formation. ∗ This chapter was reprinted with permission from Blackwell Publishers. The chapter was originally published in Handbook of International Trade, Volume II: Economic and Legal Analysis of Laws and Institutions, J. Hartigan, ed., (Blackwell Publishers, 2005). The authors would like to thank James Hartigan for comments and Sarah Lawson for excellent research assistance. 1. See Wilson (1999) and Gresik (2001) for surveys of the both the theory and empirical economic literature regarding the effects of taxation on foreign investment.

462 bruce a. blonigen and ronald b. davies

Before delving into our data analysis, it is instructive to consider the conflicting opinions about tax treaties. The FDI promotion view of treaties springs from the idea of a tax distortion to investment. To illustrate this concept, consider the following simple model of FDI. An investor in the parent country has an amount of capital K that she can invest either in her parent country, in a host country, or in both. The amount of her exported capital is Z, thus Z represents the level of FDI. The rate of return in the parent country is a decreasing function of capital invested in the parent country r(K–Z), while rate of return in the host country is a decreasing function of FDI r∗(Z). Efficiency requires that capital be allocated between the two countries such that the rates of return are equal (or that all the capital is invested in the high-return country). However, the investor is not concerned with the total return on capital, but rather her share of it. In other words, the investor bases her decisions on the after-tax rate of return in each country, not the gross rate of return. As a result, if t and t∗ are the marginal effective tax rates on investment in the parent country and in the host country, respectively, the investor will compare (1–t)r(K–Z) with (1–t∗)r∗(Z).2 Unless the marginal effective tax rates are equal, the equilibrium distribution of capital will be inefficient because efficiency calls for capital export neutrality in which only real, that is non-tax, variations govern capital flows. Therefore, as governments use their tax policies, it can quite easily lead to differential tax rates across locations and inefficient levels of FDI. Note that differential tax rates do not require different statutory taxes since the effective tax depends on many factors, including the definition of the tax base, accelerated depreciation rules, research and development tax credits, double taxation relief, and the like. It is generally believed that the effective tax rate against host investment exceeds that on investment in the parent country, implying that FDI is inefficiently low. Hines (1988) and Wilson (1993) support this concept by illustrating how typical parent-country tax policies such as accelerated depreciation for domestic investments effectively result in tax rates that discriminate against overseas investment. Tax treaties can help to alleviate this problem by coordinating tax policies between treaty partners. This idea is mirrored in the introduction of the OECD’s model tax treaty, which states that a primary goal of treaty formation is “removing the obstacles that double taxation presents,” thus reducing its “harmful effects on the exchange of goods and services and movements of capital, technology, and persons” (OECD, 1997, p. I-1).3 If treaties do indeed reduce

2. Since we are examining the marginal capital allocation, the appropriate tax rate is the marginal rate. Identification of the marginal tax rate is notoriously difficult. Graham (1996) provides discussion of the various proxies researchers have used for the marginal tax on corporate income. 3. For an excellent discussion of the workings of the OECD model tax treaty, see Baker (1994).

do bilateral tax treaties promote foreign direct investment? 463

these tax barriers to FDI, one would expect that FDI activity would rise after a treaty is enforced. Although individual treaties include a wide range of specific investment incentives, overall treaties reduce the barriers to FDI in two ways.4 First, by harmonizing the tax definitions and the tax jurisdictions of treaty partners, a treaty can reduce the double taxation of investment. For example, income is typically taxed in a host country when it is generated through a permanent establishment. However, without a treaty each country can form its own definition of a permanent establishment. If this definition differs between countries, it can lead to double taxation of overseas profits (Hamada 1966). Janeba (1996) discusses how these definition differences can result in inefficient capital flows. This idea is confirmed empirically by Hines (1988), who finds that the 1986 Tax Reform Act, which revised U.S. tax definitions, led to an increase in U.S. outbound investment. Since treaties standardize tax definitions and jurisdictions (often by matching them to those provided by the OECD’s model tax treaty (OECD 1997)), they have a similar potential to increase FDI. Second, tax treaties affect the actual statutory taxation of multinationals. They do so through the rules affecting double taxation relief and the withholding taxes levied on repatriations by FDI. Following the OECD model treaty guidelines, most tax treaties specify that both countries must either exempt foreign-earned profits from domestic taxation or offer foreign tax credits when calculating the domestic tax bill.5 Although most countries already offer their investors credits or exemptions, certain treaties do alter the relief method applied by one or both treaty partners.6 In addition to the provisions for double taxation relief, treaties usually reduce maximum allowable withholding taxes on three types of remitted income: dividend payments, interest payments, and royalty payments.7 Some treaties lower these withholding rates to as low as zero. Most treaties specify that

4. For specifics of the treaties, see the treaties themselves as reprinted by Diamond and Diamond (1998). 5. Under a tax credit, the domestic tax bill is calculated by applying the standard domestic tax rate to the pre-foreign tax level of overseas profits. A credit against this amount is then applied up to the amount of the foreign taxes paid. If this credit exceeds the domestic tax liability, the firm is in an “excess credit” position and pays no additional taxes on these overseas profits. If the parent country tax bill is greater than the amount of the credit, the firm is in an “excess limit” position and pays the remaining amount to the parent country’s government. 6. For instance, Belgium applies a reduced domestic tax rate to the foreign-earned profits of its residents. Under the U.S./Belgium treaty, however, income earned by Belgian firms in the U.S. is exempt from Belgian taxation. 7. Note that a lower foreign tax rate does not guarantee a reduction in the firm’s total tax bill. When a parent country offers foreign tax credits, only firms in an excess credit position will benefit by a reduction in host taxes (Altshuler and Newlon 1991). Thus, a decrease in the overseas tax rate may not improve capital flows.

464 bruce a. blonigen and ronald b. davies

the same maximum rates apply to both treaty partners.8 If these reductions in the withholding tax reduce the tax burden on overseas investment, equivalent to reducing t∗ above, this should increase FDI.9 Note that even though withholding tax rates fall under a treaty, this does not imply that tax receipts from inbound investment must decline. Since withholding taxes can be tailored to the specific investment from a treaty partner, it may be possible to set tax rates which encourage tax-sensitive inbound investment and actually raise total tax receipts.10 In addition, treaties are accompanied by improved information exchange between partner governments.11 Because of this, tax evasion may fall under a treaty, leading to increased tax revenue.12 Combining these arguments, it is easy to understand the common expectation that tax treaties serve to increase the amount of FDI activity between treaty partners. Nevertheless, there exist several economic and legal arguments which suggest that treaties may have no effect on FDI. For instance, Dagan (2000) asserts that the use of treaties to promote foreign direct investment is “a myth” (p. 939). She claims that since a parent country could unilaterally adjust its tax policy to eliminate distortions caused by differing parent and host country tax policies, promotion of efficiency plays little role in treaty formation. Instead, Dagan suggests that treaties are intended to reduce administration costs, reduce tax evasion, and to extract tax concessions from treaty partners. Radaelli (1997) also dismisses the double taxation objective in favor of the view that tax treaties are geared towards reducing tax evasion. Furthermore, there is the concern that tax treaties arise due to lobbying efforts by profit-seeking investors. If this is the case, then treaties may be geared towards maximizing investor profits rather

8. One exception is the U.S. treaty with Pakistan in which U.S. firms receive no tax break from Pakistan while Pakistani firms do receive reduced tax rates from the U.S. 9. Several researchers, including Altshuler and Newlon (1991), Hines (1992), Altshuler, Newlon, and Randolph (1995), and Mutti and Grubert (1991) have found that firms do respond to changes in withholding tax rates by changing both the timing and method of repatriation. However, this evidence suggests that treaties can affect the profitability of overseas investment, and does not necessarily imply that treaties will induce new FDI. 10. As derived in Bond and Samuelson (1989) among others, the tax revenue maximizing tax rate on inbound FDI is 1/(1+ ε), where ε is the elasticity of inbound FDI supply. If ε varies across countries, the revenue maximizing tax rate will differ across countries. Since a treaty allows a country to lower its tax rate, this can actually raise tax revenues if the current tax rate is greater than the optimal tax on FDI from that particular country. 11. In addition to the bilateral treaties, the OECD has established the multilateral Convention on Mutual Assistance in Tax Matters which provides for information exchange even between members without bilateral treaties (OECD 1989). 12. Devereux, Griffith, and Klemm (2001) find that, even though corporate tax rates for the U.K. appear to have declined since the mid-1960s, tax revenues from corporate income have not. They suggest that part of this result may stem from a broadening of the incomes classified as corporate.

do bilateral tax treaties promote foreign direct investment? 465

than promoting efficient investment.13 In addition to uncertainties about government objectives in treaty formation, it is by no means clear that firms’ investment activities will necessarily respond to reductions in withholding tax rates. Hartman (1985) and Sinn (1993) argue that withholding taxes are irrelevant for expanding multinational firms since it is cheaper to expand an overseas affiliate through retained earnings than through repatriated and re-exported funds. This is because retained earnings avoid the withholding taxes applied to repatriated funds. As a result, they suggest that only mature, non-expanding foreign affiliates will repatriate earnings, implying no effect of withholding taxes on FDI. Finally, it is even possible that tax treaties may actually increase the tax barriers for certain kinds of investment. As noted above, through information exchange treaties can reduce a firm’s ability to engage in transfer pricing. This is the practice by which, through manipulation of the price of goods traded between their various subsidiaries, firms can shift profits to low tax locations and minimize their global tax revenues. As argued by Casson (1979), transfer pricing provides firms with an incentive to invest in those low tax locations in order to shield profits from taxes.14 Since treaties streamline and promote the exchange of tax information by governments, this reduces firms’ ability to avoid taxes through misrepresentation of costs. As a result, treaties may reduce the incentive to engage in investment for tax minimization reasons, leading to decreased FDI activity. In addition, recent tax treaties have sought to eliminate treaty shopping. Treaty shopping is a practice in which investments are funneled through a treaty country by a third nation for the purpose of avoiding or reducing taxes. According to Radaelli (1997), concerns over treaty shopping have been a primary focus of many new treaties and have prompted the U.S. to renegotiate many of its older treaties. In addition, certain so-called “tax haven” countries (particularly Aruba, Malta, and the Netherlands-Antilles) have seen several of their treaties cancelled due to perceived insufficient efforts to prevent treaty shopping. If a treaty is revised to close this possibility, then this could easily reduce the investment activity between treaty partners as third nation investors choose to simply send their capital directly to the ultimate host. While there are many variations in the regulations that address treaty shopping, the most common rules restrict treaty benefits if more than 50% of a corporation’s stock is held by a third, nontreaty country’s residents (Doernberg 1997). With these conflicting arguments in mind, it is by no means certain that bilateral treaties will increase the amount of FDI between partner countries. This issue is of importance because of the sizable costs of treaty formation. For most countries, tax treaties are like any other international agreement in that they must 13. According to Radaelli (1997), treaty formation by the U.S. is free of such business lobbying efforts. 14. Caves (1993) provides an explanation of transfer pricing. Graham and Krugman (1995) provide case studies of firms prosecuted by the U.S. for engaging in transfer pricing.

466 bruce a. blonigen and ronald b. davies

be drawn up and then ratified by the appropriate governing body. Ratification is no mean feat since the treaty cannot conflict with other national policies. For the U.S., this is not a severe problem since tax treaties are federal instruments and thus supercede state or local laws.15 This is not always true elsewhere. Anders (1997), for example, gives an in-depth discussion of the difficulties the U.S.German treaty faces since it may violate the anti-discrimination rules of the European Union. Even ignoring such possible roadblocks, the simple conflicts of bilateral negotiation require much time and effort. Conflicts over the terms of the treaty can lead to failure during the development stage or even during ratification. For example, the U.S.-Cyprus treaty required three attempts before it was finally ratified in 1988. Thus, in light of these large costs, it is important that we gauge the potential gains from treaty formation. We seek to provide one possible measure of the gains from treaty formation by estimating the impact of treaty formation on the FDI of OECD members.16 Using recent models developed by Carr, Markusen and Maskus (2001) and Markusen and Maskus (2001), we test whether treaty formation by OECD members is associated with changes in FDI activity. Using data from 1982 to 1992, we examine the behavior of FDI stocks and FDI flows. Initial results indicate that treaties seem to increase FDI. However, this result is suspect because the sample includes many older treaties which were enacted well before our data series begins. Therefore, we also split our sample into “old” and “new” treaties. Here, we find that new treaty formation is not significantly correlated with FDI activity. In fact, when we restrict ourselves to just those countries that enacted a treaty during our sampling period, we find that FDI stocks are significantly decreased after treaty formation. These results are consistent with those of Dagan (2000) and Radaelli (1997), who claim that recent treaties are not geared towards the promotion of FDI but rather towards reductions in tax evasion. While these results do not mean that treaties cannot be used to increase FDI they do suggest that the FDI promotion argument seems suspect. The chapter proceeds as follows. In Section A, we present our empirical methodology. We discuss our data in Section B. Section C presents our results and the last section presents our conclusions.

15. Generally, the only limitation U.S. tax treaties place on state taxation is that a foreign corporation must be treated the same as a firm incorporated in another state (White 1991). 16. Hines and Willard (1992) empirically examined the number of treaties a country signs as well as the tax concessions dictated by a particular treaty. However, they do not include the amount of FDI activity as an explanatory variable nor do they consider the effect of taxation on FDI. The UN (1998) has studied the effects of bilateral treaties for the promotion and protection of FDI. These treaties are generally geared towards increasing investment in developing nations by ensuring a favorable political and economic climate and do not address tax concerns.

do bilateral tax treaties promote foreign direct investment? 467

A. Empirical Framework For Statistical Analysis In order to examine how treaties affect FDI, we require a framework that describes the determinants of FDI. Over the past two decades, James Markusen and colleagues have developed formal general equilibrium theory models of multinational enterprise (MNE) activity.17 These theory models lead to predictions of equilibrium FDI activity across bilateral pairs of countries in terms of a few observable factors. Recently, Carr, Markusen and Maskus (2001) and Markusen and Maskus (2001) tested the “knowledge-capital” version of the Markusen model of MNE activity using data on affiliate sales of U.S. firms in other countries and foreign affiliate sales in the U.S. over the period 1986–1994. They find substantial empirical fit of the knowledge-capital model to the data. The Carr, Markusen, Maskus (CMM) empirical framework is specified by equation 1: FDIij = f (SUMGDPij, GDPDIFSQij, SKDIFFij, SKDIFFij∗GDPDIFFij, (SKDIFFij)2∗T_OPENj, Zij, TREATYij)

(1)

The dependent variable, FDIij is a measure of FDI activity from a parent country (i) to a host country (j). CMM use affiliate sales data as their “FDI” measure, though Blonigen and Davies (2000; 2001) find that the CMM empirical framework also fits U.S. FDI stock (and to some extent, FDI flow) data reasonably well. The first five independent variables on the right hand side of equation 1 are the variables specific to the CMM framework. The first two terms are relatively straightforward, with SUMGDP defined as the sum of the two countries’ real gross domestic products (GDPs), and GDPDIFSQ defined as the squared difference between the two countries’ real GDP. There is an expected positive correlation between SUMGDP and FDI activity and an expected negative correlation between GDPDIFSQ and FDI activity. The intuitive expectation is that with some positive level of trade frictions, larger and more similar sized markets better support the higher fixed costs associated with setting up production across countries (versus exporting) and lead to greater FDI activity. The third, fourth, fifth terms on the right-hand side of equation 1 are more complicated terms related to differences in the two countries’ relative endowments of skilled labor to unskilled labor. Skilled labor is important in the theory of MNE activity because firms that have firm-specific assets (which are developed using skilled labor) will have the greatest incentives to expand their operations across borders. In some theories of MNE activity, MNE firms are additionally attracted to place production activities in less-skilled countries because production activity is less skill-intensive than the headquarter activities conducted in the parent country. These are generally models which emphasize vertical reasons for MNE activity; i.e., to locate its production activities in countries with lower wages for low-skill-intensive activities. These models would suggest that FDI activity 17. For recent examples of this work see Markusen, Venables, Eby-Konan, and Zhang (1996) or Markusen and Venables (1997).

468 bruce a. blonigen and ronald b. davies

should be stronger between countries with greater differences in skilled labor abundance. Other models that emphasize MNE motives to locate production in large markets to avoid tariffs and transport costs (horizontal MNE models) often predict that FDI activity should increase with greater similarities in skilled labor abundance. In equation 1 the SKDIFF variable is the parent country’s skilled labor abundance minus the host country’s skilled labor abundance. Given the vertical MNE motives that exist in the CMM knowledge-capital model, CMM predict a positive correlation between SKDIFF with FDI activity from the parent to the host country. The fourth term is an interaction term between SKDIFF and GDPDIFF, the parent country’s GDP minus the host country’s GDP. CMM suggest that the knowledge-capital model predicts a negative correlation between this variable and FDI from the parent to the host country, since the theory predicts larger effects of SKDIFF on FDI when the parent country is much smaller than the host country. Finally, CMM include a fifth term that interacts the square of SKDIFF multiplied by trade openness in the host country. The predicted sign from their model on this term is positive.18 In addition to the variables just described, additional control variables (Zij) have typically been included in the CMM empirical framework. First, distance (DISTANCEij) is included as a proxy for transport and other trade costs that may affect a firm’s decision about whether to become an MNE. Second, trade openness for both the parent and host countries (T_OPENi and T_OPENj) affects the MNE’s ability to trade intermediates and final goods, which then affects the location of MNEs. Greater openness in the host country should lower FDI activity because it lowers trade frictions and makes exporting a relatively more attractive method of serving the host market than FDI. Greater openness in the parent country should increase FDI, since openness makes it easier to ship goods back to the parent country from foreign affiliates. FDI openness of the host country (F_OPENj) proxies for the costs of setting up an MNE, with greater openness expected to increase FDI activity. The last independent variable on the right hand side of equation 1 is our focus variable: a measure of bilateral investment tax treaty activity. There are substantial measurement issues that determine how we define this variable. In particular, we can observe when countries make bilateral investment tax treaties with each other, but these treaties certainly differ from each other along many dimensions which are very difficult to quantify. In addition, treaties which appear the same on paper can have vastly different consequences for different pairs of countries depending on the unilaterally-adopted tax practices of the countries before entering the treaty. There is little that can be quantified across treaties with the 18. In the CMM (2001) study, the authors interact the skill difference term with host trade costs, the opposite of host trade openness. Thus, we expect the opposite sign on this coefficient than does CMM (2001). See CMM (2001) for further details on the knowledgecapital model.

do bilateral tax treaties promote foreign direct investment? 469

exception of agreed-upon tax withholding rates. Unfortunately, these withholding rates are typically maximum allowable rates and there is evidence that countries sometimes set rates below these maximum allowable rates after the treaty, making them uninformative for our purposes of analysis (Blonigen and Davies 2000; 2001). Because of these difficulties, we primarily measure tax treaty activity in this chapter as a binary variable taking the value of “1” if two countries have signed a bilateral tax treaty that was eventually entered into force and “0” if they have not. As a result, we will only be able to estimate an average total impact of tax treaties across our sample of countries. While we only report results on the effect of bilateral tax treaties using the empirical CMM framework described above, we note that we get qualitatively similar results for the impact of tax treaties when using alternative empirical frameworks. This includes variations of the CMM framework found in Markusen and Maskus (1999) and a “gravity-type” framework which posits FDI activity as a function only of the size of the countries (proxied by real GDP) and distance. In addition, Blonigen, Davies and Head (2002) show that the CMM framework misspecifies the SKDIFF and GDPDIFF terms, which should be expressed in absolute values. While this correction strongly affects the implications of the relationship between skill differences and FDI activity, it has virtually zero quantitative effect on our estimates of treaty effects. Although results from these alternative empirical specifications have been omitted here to save space, they are available upon request from the authors. B. Data Among the major hurdles in the analysis of the effect of bilateral tax treaties on FDI are data constraints. As noted above, this begins with measuring the treaty activity. However, there are also significant measurement issues with respect to the data on bilateral FDI activity as well. For this study we use OECD data on bilateral FDI stocks and flows, as reported by OECD-member countries.19 Such data were not even compiled into a publicly-available form until 1993, with the first annual OECD International Direct Investment Statistics Yearbook.20 Since the data are collected from national sources in each country, there is substantial variation in coverage by country source and by year, and there is variation in measurement of FDI activity itself. Almost all countries report inflows and outflows

19. As mentioned in the section above, Carr, Markusen and Maskus (2001) and Markusen and Maskus (2001) use U.S. data on MNEs’ foreign affiliate sales, which is a theoretically preferable measure of MNE activity. However, to our knowledge, there are very few countries that keep track of affiliate sales, and there is no comprehensive crosscountry database of foreign affiliate sales activity, even for OECD countries. 20. These data are available in print form in these annual yearbooks or in electronic form on the OECD Statistical Compendium CD-ROM, available for purchase from the OECD.

470 bruce a. blonigen and ronald b. davies

for at least some select bilateral pairings, with about half also reporting measures of inward and outward stocks of FDI. The earliest data available begin in 1982. Table 1 provides further details on data coverage across OECD countries and years in our sample. There are definite comparability issues regarding FDI measures across countries. For example, a number of OECD countries do not include reinvested earnings by firms in their measures of FDI. Countries can also differ in what percentage of foreign-owned shares of a firm is necessary for it to be classified as FDI rather than as a portfolio investment.21 However, with only a couple exceptions, we note that FDI definitions are fairly consistent for the same country over time. This is important because in our statistical analysis we will be able to use techniques that estimate relationships between variables using only the sample data variation within individual countries over time, avoiding cross-country data measurement consistency problems that arise for other statistical techniques. Besides our FDI and tax treaty measures, our statistical analysis employs the additional CMM empirical framework control variables. Our data on real GDP, which is used to construct the SUMGDP and GDPDIFSQ variables, come from the well-known Penn World Tables, which are described by Summers and Heston (1991) and are available online at http://datacentre.chass.utoronto. ca:5680/pwt/. To construct our measure of differences in skilled labor endowments, we use World Bank data on the average country-level education attainment as a proxy for such endowments. Nehru and Dhareshwar (1993) provide further details on these data. DISTANCE was measured as the distance between capital cities as reported by the Bali Online Corporation. This distance calculator can be found at Trade openness measures of the two countries were obtained from the Penn World Tables and are defined as a country’s total trade flows (exports plus imports) divided by its GDP. The F_OPENj was constructed in a similar manner as a country’s total FDI flows divided by its GDP using the United Nations’ World Investment Directory. Details on this variable construction can be found in Blonigen and Davies (2001). The Penn World Tables run only through 1992, whereas our FDI data begin in 1982 at the earliest. These data availability issues limit our sample and analysis to the period from 1982 through 1992. Table 2 provides descriptive statistics of our dependent and independent variables. We have 3276 observations on FDI outflows by OECD countries, and 2235 observations on FDI outbound stock. Average annual bilateral FDI outflows are almost $284 million with substantial variation across observations 21. IMF and OECD guidelines, which many of the countries follow or eventually adopted, specify investment as FDI when acquired shares are 10% or higher of the target firm’s outstanding stock. Graham and Krugman (1993) find that the foreign parent of a MNE in the U.S. on average owns 77.5% of the affiliate’s equity, suggesting that this problem may not be overwhelming.

do bilateral tax treaties promote foreign direct investment? 471

table 1. fdi data coverage by oecd country Country

Type of FDI Measure

Australia

Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inflows Outflows Inbound stock Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inflows Inflows Inflows Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inbound stock Outbound stock

Austria

BelgiumLuxembourg Canada

Denmark Finland

France

Germany Greece Iceland Ireland Italy

Japan

Number of Partner Countries Reporteda 18 17 17 12 11 13 9 11 44 45 3 3 29 30 20 22 10 17 1 23 23 43 34 19 18 14 0 0 20 20 22 23 9 36 9 37

Years Coveredb 1982– 1982– 1982– 1982– 1982– 1982– 1982, 1986– 1982, 1986– 1982– 1982– 1982– 1982– 1982– 1982– 1982– 1982– 1982– 1982– 1989– 1982– 1982– 1987– 1987– 1982– 1982– 1987– 1987– 1983– 1982– 1982– 1985– 1985– 1982– 1982– 1982– 1982– Continued

472 bruce a. blonigen and ronald b. davies table 1. fdi data coverage by oecd country (cont’d...) Country

Type of FDI Measure

Netherlands

Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inflows Outflows Inflows Outflows Inflows Outflows Outbound stock Inflows Outflows Inflows Outflows Inbound stock Outbound stock Inflows Outflows Inbound stock Outbound stock

New Zealand Norway

Portugal Spain Sweden Switzerland

Turkey United Kingdom

United States

a

Number of Partner Countries Reporteda 9 9 14 14 4 4 14 19 18 24 18 10 20 18 15 36 0 0 0 18 6 18 36 19 35 39 41 28 41

Years Coveredb 1982– 1982– 1984– 1984– 1984– 1984– 1986– 1987– 1987– 1988– 1982– 1982– 1982– 1982– 1982– 1982– 1983– 1986– 1986– 1982– 1989– 1982– 1984, 1987– 1982– 1984, 1987– 1982– 1982– 1982– 1982–

Figures for 1990. Not all reported countries are necessarily reported each year during range indicated. “0” indicates that data are available only by region, not specific countries. b

do bilateral tax treaties promote foreign direct investment? 473

table 2. descriptive statistics of variables Variables

Number of Mean Observations

Standard Deviation

Minimum

Maximum

Dependent Variables FDI outbound stock FDI outflows

2235 3276

3378.13 283.82

8002.59 1056.61

0.00 –2550.56

92733.80 42267.61

Independent Variables SUMGDPij GDPDIFSQij SKDIFFij SKDIFFij∗GDPDIFFij (SKDIFFij)2∗T_OPENj DISTANCEij T_OPENi T_OPENj F_OPENj TREATYij OLD TREATYij NEW TREATYij

3276 3276 3276 3276 3276 3276 3276 3276 3276 3276 3276 3276

1259.53 1993241 1.21 2124.00 475.35 5695.07 56.38 68.22 12.42 0.77 0.74 0.03

1260.96 4710473 2.54 5374.54 821.85 5129.36 31.01 55.26 14.20 0.42 0.44 0.17

61.79 0.11 –6.65 –7494.69 0.00 174.00 17.62 8.96 0.15 0.00 0.00 0.00

6449.00 2.06e+07 8.10 31011.47 8850.60 19007.00 156.45 386.23 93.69 1.00 1.00 1.00

Notes: FDI outbound stock, FDI outflows, SUMGDP, and GDPDIFSQ are measured in millions of real U.S. dollars. Distance is measured in miles between capital cities. Skill difference is measured in mean years of female and male educational attainment. See text for variable definitions.

(standard deviation of just greater than than $1 billion).22 Average FDI outbound stock is $3.378 billion with a standard deviation of $8 billion.23 About 77% of the sample observations are of bilateral country pairs with bilateral tax treaties in place. The majority of these (74%) are what we term “old” treaties which were in place before our sample begins, with three percent of the sample observations connected to “new” treaties that were enacted during our sample. We discuss the importance of this distinction for our statistical analysis in more detail below.

22. These figures are in 1995 U.S. dollars. 23. We remind the reader that these are figures for those bilateral pairs reported by each OECD country, which are typically skewed toward reporting only the countries with which the reporting country has the largest FDI activity.

474 bruce a. blonigen and ronald b. davies

C. Statistical analysis and results Our statistical analysis will proceed in stages, beginning with a relatively naive estimation of equation 1, and then proceeding with more sophisticated estimation procedures to correct for potential statistical problems. We do this to show that it is not easy to estimate the true effect of tax treaties on FDI, primarily because many treaties by OECD countries were in place before our sample begins. Most of these “old” treaties are between OECD countries, with many beginning in the decade after World War II. If we get a positive correlation between our tax treaty variable and our dependent variable, FDI activity, it is not clear whether other unobservable characteristics of the tax treaty country pairings may be leading to both increased FDI activity and a tax treaty. This occurs because the tax treaty variable, a simple binary variable, will pick up any residual effects on FDI that are not measured by the other control regressors. This problem, known as simultaneity, makes identification of the treaty effect difficult to measure. For example, there are likely a number of underlying reasons beyond their bilateral tax treaty that explain why the U.S. and the U.K. have large FDI activity with each other. These other reasons, however, may not be observable. Because of this, statistical analysis will assign the influence of the unobserved factors to the observed existence of a treaty. As such, although statistical analysis would indicate a strong positive relationship between the existence of a treaty and FDI activity, one hesitates to say that the treaty causes the activity between the two countries. However, our sample also includes a number of bilateral tax treaties completed by OECD countries after our sample data begins. These “new” treaties afford a much better opportunity to measure the impact of a tax treaty, as we have data on FDI activity both before and after the treaty is signed. Presuming that there are no other changes occurring at the time of the treaty that would affect FDI (besides those captured by our control regressors), we can estimate the effect of these new treaties on FDI more precisely by comparing the pre- and post-treaty information. Table 3 lists new treaties that were completed by OECD countries after the first year of our sample, 1983, through the last year of our sample, 1992.24 As we show below, the estimated effects of old treaties versus new treaties on FDI activity are quite different. Given this discussion, we give much more weight to the credibility of our evidence for the effects of new treaties. Column 1 of Table 4 provides statistical results when we use ordinary least squares (OLS) regression techniques to estimate equation 1 on our full sample of countries and years when FDI activity is measured as the parent country’s FDI stock in the host country. Column 2 of Table 4 provides statistical results when we conduct the same procedure, but define our FDI activity measure as the 24. We note that because of missing data, some of the bilateral pairs completing treaties in Table 3 are not covered by our sample. More specifically, nineteen of the new bilateral treaties in Table 3 are covered in our sample in at least one of the two possible directions.

do bilateral tax treaties promote foreign direct investment? 475

table 3. new treaties by oecd countries from 1983–1992 Country

Bilateral Tax Treaties (Year of Treaty in Parentheses)

Australia

Italy (1983), Korea (1983), Norway (1983), Ireland (1984), Finland (1985), Austria (1987), China (1990), Thailand (1990), Hungary (1992), Poland (1992) Thailand (1986), Australia (1987) Korea (1985), Turkey (1988)

Austria BelgiumLuxembourga Canada

Denmark Finland France Germany Greece Iceland Ireland Italy Japan Netherlands New Zealand Norway Portugal Spain Sweden Switzerland Turkey United Kingdom United States

a

Bangladesh (1983), Brazil (1985), Cyprus (1986), India (1986), China (1987), Kenya (1987), Poland (1989), Slovak Republic (1987) Cyprus (1984), Indonesia (1986), China (1987) Australia (1985), New Zealand (1985), Turkey (1985), Thailand (1986), China (1987), Yugoslavia (1987), Indonesia (1988) Bangladesh (1988), Trinidad and Tobago (1988), Nigeria (1991) Philippines (1984), Turkey (1986), Indonesia (1988) Hungary (1984), Switzerland (1984), Czechoslovakia (1987), Norway (1989) None. Australia (1984), New Zealand (1987) China (1987), India (1987) Indonesia (1983) Pakistan (1983), Romania (1983), China (1987), India (1989), Brazil (1991) Norway (1983), Finland (1985), Ireland (1987), Indonesia (1988) Australia (1983), New Zealand (1983), Yugoslavia (1984), China (1987), Pakistan (1987), Philippines (1988), Greece (1989) None Yugoslavia (1983) Trinidad and Tobago (1985), China (1987), Cyprus (1989), Indonesia (1989), Turkey (1989) Greece (1984), Egypt (1988), Indonesia (1989), China (1991) Finland (1985), Germany (1986), Belgium-Luxembourg (1988) Sweden (1989) India (1983), Thailand (1983), China (1985), India (1987) Barbados (1984), China (1985), Tunisia (1986), Cyprus (1988), Indonesia (1989), India (1990), Spain (1991)

Includes only treaties made by Belgium, not Luxembourg.

476 bruce a. blonigen and ronald b. davies

parent country’s FDI flow into the host country. With the exception of SKDIFF and the trade openness measures, the control variables have their predicted signs and are generally statistically significant in the FDI stock and FDI flow regressions. The SKDIFF variable is estimated with an incorrect negative sign in both regressions, but is not statistically significant at standard levels of statistical confidence. The general fit of the empirical framework to the data is better for the FDI stock regression, where the control variables explain about 34% of the variation in the dependent variable, as compared to the FDI flow regression, where only about 10% of the variation in the dependent variable is explained by the independent variables. Both statistical regressions provide strong evidence for a positive effect of tax treaties on FDI activity. Everything else being equal, the presence of a tax treaty means almost an extra $2.5 billion of parent FDI stock in the host country compared to a situation where there is no bilateral tax treaty. This is quite significant given an average of $3.4 billion in FDI stock in the sample. The effect of treaties on FDI flows is estimated to be somewhat larger relative to its mean. Everything else being equal, the presence of a tax treaty means an extra $234 million of annual parent FDI flows into the host country, compared to an average annual flow of $284 million. Even setting issues of magnitude aside, these estimates suggest that we can be over 99% confident that these coefficient estimates are not zero. As mentioned above, our sample includes both old treaties that occurred before our sample period began and new treaties that occurred during our sample’s time period. Because of this, it is not clear whether the strongly positive treaty effects in columns 1 and 2 are the result of the treaties themselves or some other unobserved factor. To deal with this, columns 3 and 4 of Table 4 provide results when we estimate based on the same empirical framework as in columns 1 and 2, but allow for separate effects for old and new treaties; i.e., separate binary variables indicating whether an old or new treaty was present between the bilateral country pair in a given year or not. The estimated effects of old and new treaties are quite different for both of our specifications (FDI stock and FDI flows). Old treaties continue to show a positive effect on FDI activity, and that effect continues to be highly statistically significant and of slightly larger magnitude than our results in columns 1 and 2. Our estimates suggest that the presence of an old treaty increases outbound FDI stock by $2.8 billion (compared to $2.5 billion in the column 1 estimates) and increases outbound FDI flows by $258 million (compared to $242 million in the column 2 estimates). In contrast, new treaties yield coefficient estimates that are negative in sign. However, given high standard errors relative to the small coefficient estimates, we cannot statistically reject the hypothesis that the new treaties have no impact on FDI activity. As mentioned above, it is difficult to assign causation to the old treaty effects because they were in place before our sample period began—the positive correlation

do bilateral tax treaties promote foreign direct investment? 477

table 4. estimated treaty effects on oecd outbound fdi stock and flows using ols Empirical Models and Dependent Variables Ordinary Least Squares with Ordinary Least Squares Treaty Variable with Separate Old and New Treaty Variables Regressors

FDI Stock(1)

FDI Flows(2)

FDI Stock(3)

FDI Flows(4)

233.81∗∗ (0.00) –

–

–

NEW TREATY

2446.9∗∗ (0.00) –

OLD TREATY

–

–

–263.06 (0.74) 2782.8∗∗ (0.00)

–20.586 (0.86) 257.81∗∗ (0.00)

4.368∗∗ (0.00) –0.0001 (0.16) 2.948 (0.97) –0.315∗∗ (0.00) –0.697∗∗ (0.00) –0.396∗∗ (0.00) –7.567 (0.42) –26.184∗∗ (0.00)

0.426∗∗ (0.00) –0.00004∗∗ (0.00) –5.747 (0.51) –0.020∗∗ (0.00) –0.049 (0.10) –0.031∗∗ (0.00) –0.430 (0.54) –1.347∗ (0.02)

4.426∗∗ (0.00) –0.0001 (0.09) 65.085 (0.41) –0.313∗∗ (0.00) –0.623∗∗ (0.01) –0.392∗∗ (0.00) –4.767 (0.61) –27.754∗∗ (0.00)

0.431∗∗ (0.00) –0.00004∗∗ (0.00) –1.017 (0.91) –0.020∗∗ (0.00) –0.039 (0.18) –0.030∗∗ (0.00) –0.335 (0.63) –1.461∗∗ (0.01)

Treaty Variables TREATY

CMM Controls SUMGDPij GDPDIFSQij SKDIFFij SKDIFFij ∗ GDPDIFFij (SKDIFFij)2∗T_OPENj DISTANCEij T_OPENi T_OPENj

Continued

478 bruce a. blonigen and ronald b. davies table 4. estimated treaty effects on oecd outbound fdi stock and flows using ols (cont’d...) Empirical Models and Dependent Variables Ordinary Least Squares with Ordinary Least Squares Treaty Variable with Separate Old and New Treaty Variables R-Squared F Test Sample Size

0.34 112.14∗∗ 2235

0.11 38.31∗∗ 3276

0.34 113.21∗∗ 2235

0.11 35.45∗∗ 3276

Notes: P-values are in parentheses, with ∗∗ and ∗ denoting statistical confidence levels at the 99 and 95% levels, respectively. P-values indicate statistical probability (in decimal form) that the true parameter value is zero (i.e., has no effect on the dependent variable). R-squared is the ratio of the variation in the dependent variable explained by the regressors. F test is a statistical test of the probablility of rejecting the hypothesis that the coefficients of the regressors are jointly zero.

may just suggest that countries that would naturally have high FDI activity due to other (unobserved) factors also make sure they have a bilateral tax treaty in place.25 This is particularly worrisome since our (observed) control variables do not even account for half of the variation in the dependent variables. A common way to control for unobserved characteristics that affect the FDI activity between a bilateral pair of countries is to estimate what is known as a fixed effects specification. This means that, in addition our control variables, we also include a binary variable for each bilateral country pairing. These binary fixed-effect country-pair variables will estimate the aggregate effect of time-invariant characteristics (both observed and unobserved) that raise or lower the FDI activity for that bilateral pairing versus the average. In other words, the variable acts as an intercept term that is specific to the bilateral country pair. With fixed effects, our estimates then come only from the time series variation in our variables within each bilateral country pair, not the variation across bilateral country pairs. Note that now the effects of our time-invariant regressors, 25. There is a similar, though less serious, concern that the new treaties occur because FDI activity is increasing or expected to increase between the countries due to unobserved factors. However, this would be expected to bias us toward finding a positive correlation between new treaties and FDI activity. In contrast, our results below find evidence for a negative correlation.

do bilateral tax treaties promote foreign direct investment? 479

distance and old treaties, will be subsumed into these fixed effects. In other words, we cannot identify the effects of old treaties separately from other timeinvariant characteristics, such as historical conditions, that affect the overall FDI activity between a bilateral country pair. Because of this, we focus solely on the new treaties enacted at some point during the sample of years. Estimation using only the variation within bilateral country pairs provides an additional benefit since it can help to reduce the problems caused by different definitions of FDI activity across the countries in our sample, providing more credible estimates. Columns 1 and 2 of Table 5 present our fixed effects empirical estimation results. The fit of the equations increases dramatically: 94% of the variation in the dependent variable is explained in the FDI stock regression and 38% of the variation is explained in the FDI flow regression. Not surprisingly, statistical tests easily confirm that the fixed effect variables are jointly statistically significant for explaining FDI activity. The main CMM framework control variables still perform well. In fact, the skilled-labor difference variable (SKDIFF), now has the correct sign and is statistically significant. However, the trade openness measures generally have the wrong sign in these regressions. The surprising result in columns 1 and 2 of Table 5 is that the estimated effect of the new treaties on FDI activity is now strongly negative. The estimated impact has increased substantially and the effects are much more statistically significant: A new treaty leads to a $2.6 billion decrease in FDI stock and $351 million decrease in annual FDI flows. In column 3, we introduce a lagged dependent variable to control for dynamic adjustment in our FDI stock equation as there is probably persistence in the FDI stock over time that can lead to statistical problems with our estimates. The coefficient on the lagged dependent variable is statistically significant, but does not seriously affect our other regressor coefficient estimates, including the new treaty variable.26 In summary, we estimate very different correlations between old treaties and FDI activity versus new treaties and FDI activity. Given the simultaneity concerns described above, it is difficult to assign much weight to our old treaty evidence for positive effects on FDI, even though it is the best we can do given data constraints. Because consistent recording of FDI activity began much more recently than the enactment of these old treaties, it is virtually impossible to identify the effect of the treaties on FDI activity from other country-pair characteristics that might affect both the inherent FDI activity and the incentives to have treaties. In contrast, we are much more confident in our new treaty estimates from a statistical standpoint and these results yield a very surprising conclusion: new treaties are not promoting FDI activity and the evidence suggests that they may even be decreasing FDI activity.

26. A lagged dependent variable is not statistically significant in the FDI flow specifications.

480 bruce a. blonigen and ronald b. davies table 5. estimated treaty effects on oecd outbound fdi stock and flows using fixed effects Empirical Models and Dependent Variables Fixed Effects with New Treaty Variable Only Regressors

FDI Stock (1)

FDI Flows (2)

FDI Stock with Lagged Dependent Variable (3)

–2597.6∗∗ (0.00)

–350.83 (0.06)

–2212.3∗∗ (0.00)

–

–

R-Squared

9.527∗∗ (0.00) –0.0004∗∗ (0.00) 2677.6∗∗ (0.00) –0.819∗∗ (0.00) 1.029∗ (0.01) –1.924 (0.90) –47.163∗∗ (0.00) 124.66∗∗ (0.00) 0.94

1.379∗∗ (0.00) –0.0001∗∗ (0.00) 67.116 (0.63) –0.008 (0.89) 0.010 (0.94) –3.164 (0.50) –0.552 (0.85) 7.596 (0.19) 0.38

0.308∗∗ (0.00) 7.053∗∗ (0.00) –0.0003∗∗ (0.00) 1331.6∗∗ (0.00) –0.540∗∗ (0.00) 0.610 (0.09) –2.145 (0.89) –28.957∗∗ (0.00) 79.207∗∗ (0.00) 0.96

F Test

95.46∗∗

12.94∗∗

192.87∗∗

Sample Size

2235

3276

2041

Treaty Variables NEW TREATY

Control Regressors LAGGED FDI STOCK SUMGDPij GDPDIFSQij SKDIFFij SKDIFFij ∗ GDPDIFFij (SKDIFFij)2∗T_OPENj T_OPENi T_OPENj F_OPENj

Notes: P-values are in parentheses, with ∗∗ and ∗ denoting statistical confidence levels at the 99 and 95% levels, respectively. P-values indicate statistical probability (in decimal form) that the true parameter value is zero (i.e., has no effect on the dependent variable). R-squared is the ratio of the variation in the dependent variable explained by the regressors. F test is a statistical test of the probablility of rejecting the hypothesis that the coefficients of the regressors are jointly zero.

do bilateral tax treaties promote foreign direct investment? 481

While this new treaty result is surprising, it is confirmed by Blonigen and Davies (2001) using a completely different database on U.S.-only FDI activity. The U.S. data, collected by the U.S. Bureau of Economic Activity (BEA), is arguably the best-measured and most comprehensive data on FDI activity of any country in the world. The BEA has data on U.S. bilateral FDI activity (both inbound and outbound) stretching back as far as 1966 for FDI flows and stock and the early 1980s for affiliate sales.27 Using this U.S. database and U.S. bilateral tax treaty activity, Blonigen and Davies (2001) also found evidence for substantial negative effects of new treaties on FDI activity. In an earlier version of the paper, Blonigen and Davies (2000), had similarly addressed the issue of old versus new treaties by estimating the U.S.-only sample without observations connected with Canada, Japan and European countries. We found that the remaining sample still displayed positive effects of tax treaties on FDI activity, but with a very long lag. However, this result was driven by the fact that the reduced sample still contained FDI activity with a few remaining old treaty partners, specifically Australia and New Zealand. Once all of the old treaty partners are eliminated from the sample, as was effectively done by Blonigen and Davies (2001), results are consistent with this chapter’s results for OECD countries: There is no credible evidence in the data that tax treaties have significant positive effects on FDI activity.

conclusion The majority of economic and legal texts stress the intuitive notion that bilateral tax treaties should promote FDI activity. As discussed in the introduction, there are a number of reasons why this may not be true in theory. This chapter adds to this debate by providing some of the first evidence on the effect of bilateral tax treaties on FDI activity. Using OECD data we find that new treaty activity (during the 1983–1992 period) suggests strong negative impacts on FDI. While we find a positive correlation in the case of much older treaties, we cannot weight this evidence very heavily as we cannot observe FDI activity before these treaties were in place. These results are consistent with previous work by Blonigen and Davies (2001) using only U.S. data. Thus, in conjunction with this earlier work, our results cast doubt upon the FDI promotion rationale for treaty formation, which stands in contrast to the conventional wisdom among many economists and lawyers. One possible reason for the non-promotion effect of treaties on FDI activity is that treaties reduce firms’ abilities to evade taxes through transfer pricing or treaty shopping. Our data on aggregate FDI activity are not well-suited to address whether these issues connected with firm-level behavior are behind the 27. Of course, the disadvantage of the U.S. data versus the OECD data is that its observations are all tied to one country, making it difficult to know whether the results from such a sample generalize to the rest of the world.

482 bruce a. blonigen and ronald b. davies

overall result. An additional possibility for non-promotion of FDI activity by new treaties is that treaties may increase investment uncertainty, at least in the short run. Since a new treaty has yet to be tested in the courts of the partner countries, it may actually increase the perceived risk of investment between treaty partners until the legal interpretation of the treaty has been resolved. Thus, in the short run, the treaty may lead to a reduction in FDI activity. Over the long run, however, this uncertainty will be resolved, clearing the way for the treaty to promote investment. However, when we alter our new treaty dummy variable to only take the value of “1” a year (or even two years) after the treaty was signed, we get similar negative and statistically significant effects of new treaties on FDI activity. This would argue that the uncertainty issue is not behind the effects we find unless it takes many years to resolve such uncertainty. On the other hand, our work with U.S. data in Blonigen and Davies (2001) found evidence that the negative drop in FDI activity seemed to occur mainly at the time of the treaty. We leave these and other important issues for future work and hope that these results serve as a guidepost for continued exploration of the relationship between tax treaties and foreign investment.

references Altshuler, Rosanne and T. Scott Newlon (1991). “The effects of U.S. tax policy on the income repatriation patterns of U.S. multinational corporations,” National Bureau of Economic Research Discussion Paper Series, No. 571. Altshuler, Rosanne, T. Scott Newlon and William Randolph (1995). “Do repatriation taxes matter? Evidence from the tax returns of U.S. multinationals,” in Martin Feldstein, James Hines, Jr. and R. Glenn Hubbard, eds., Effects of Taxation on Multinational Corporations, (Chicago: University of Chicago Press), pp. 253–272. Anders, Dietmar (1997). “The limitation on benefit clause of the U.S.-German tax treaty and its compatibility with European law,” Northwestern Journal of International Law and Business, 18. Baker, Philip, (1994). Double Taxation Conventions and International Tax Law: A Manual on the OECD Model Tax Convention on Income and on Capital of 1992, 2nd ed. (London: Sweet & Maxwell). Blonigen, Bruce A. and Ronald B. Davies (2000). “The effects of bilateral tax treaties on U.S. FDI activity,” National Bureau of Economic Research Working Paper Series, No. 7929. Blonigen, Bruce A. and Ronald B. Davies (2001). “The effects of bilateral tax treaties on U.S. FDI activity,” Manuscript. Blonigen, Bruce A., Ronald B. Davies and Keith Head (2002). “Estimating the knowledge-capital model of the multinational enterprise: comment,” Manuscript. Bond, Eric and Larry Samuelson (1989). “Strategic behaviour and the rules for international taxation of capital,” Economic Journal, 99, pp. 1099–1111. Carr, David, James R. Markusen and Keith E. Maskus (2001). “Estimating the knowledge-capital model of the multinational enterprise,” American Economic Review, 91, pp. 693–708.

do bilateral tax treaties promote foreign direct investment? 483

Casson, Mark (1979). Alternatives to the Multinational Enterprise (London: Macmillan Press). Caves, Richard E. (1993). Multinational Enterprise and Economic Analysis, 2nd Edition (New York: Cambridge Surveys of Economic Literature, Cambridge Press). Dagan, Tsilly (2000). “The tax treaties myth,” New York University Journal of International Law and Politics, pp. 939–996. Devereux, Michael, Rachel Griffith and Alexander Klemm (2001). “Have taxes on mobile capital declined?” Manuscript. Diamond, Walter and Dorothy Diamond (1998). International Tax Treaties of All Nations (New York: Ocean Publications). Doernberg, Richard (1997). International Taxation in a Nutshell, 3rd Edition (Minnesota: West Publishing) Graham, Edward and Paul Krugman (1995). Foreign Direct Investment in the United States, 3rd ed. (Washington, D.C.: Institute for International Economics). Graham, John R. (1996). “Proxies for the corporate marginal tax rate,” Journal of Financial Economics, 42, pp. 187–221. Gresik, Thomas A. (2001). “The taxing tax of taxing transnationals,” Journal of Economic Literature, 39, pp. 800–838. Grubert, Harry and John Mutti (1991). “Taxes, tariffs and transfer pricing in multinational corporate decision making,” Review of Economics and Statistics, 73, pp. 285–293. Hamada, Koichi (1966). “Strategic aspects of taxation on foreign investment income,” Quarterly Journal of Economics, 80, pp. 361–375. Hartman, David (1985). “Tax policy and foreign direct investment,” Journal of Public Economics, 26, pp. 107–121. Hines, James R. Jr. (1988). “Taxation and U.S. multinational investment,” in Lawrence Summers, ed., Tax Policy and the Economy, 2 (Boston: MIT Press), pp. 33–61. Hines, James R. Jr. (1992). “Credit and deferral as international investment incentives,” National Bureau of Economic Research Working Paper Series, No. 4191. Hines, James R. and Kristen L. Willard (1992). “Trick or treaty? Bargains and surprises in international tax agreements,” Manuscript. Janeba, Eckhard (1996). “Foreign direct investment under oligopoly: profit shifting or profit capturing?” Journal of Public Economics, 60, pp. 423–445. Markusen, James R. and Anthony J. Venables (1997). “The role of multinational firms in the wage-gap debate,” Review of International Economics, 5, pp. 43–51. Markusen, James R. and Keith E. Maskus (1999). “Discriminating among alternative theories of the multinational enterprise,” National Bureau of Economic Research Working Paper Series, No. 7164. Markusen, James R. and Keith E. Maskus (2001). “Multinational firms: reconciling theory and evidence,” in Magnus Blomstrom and Linda S. Goldberg, eds., Topics in Empirical International Economics: A Festschrift in Honor of Robert E. Lipsey (Chicago: University of Chicago Press for National Bureau of Economic Research), pp. 71–95. Markusen, James R, Anthony J. Venables, Denise Eby-Konan, and Kevin Honglin Zhang (1996). “A unified treatment of horizontal direct investment, vertical direct investment, and the pattern of trade in goods and services,” National Bureau of Economic Research Working Paper Series, No. 5696. Nehru, Vikram and Ashok Dhareshwar (1993). “A new database on physical capital stock: sources, methodology and results,” Rivista de Analisis Economico, 8, pp. 37–59.

484 bruce a. blonigen and ronald b. davies Organization for Economic Cooperation and Development (1989). Explanatory report on the convention on mutual administrative assistance in tax matters, (Paris: OECD Committee on Fiscal Affairs). Organization for Economic Cooperation and Development (1997). Model tax convention on income and on capital, (Paris: OECD Committee on Fiscal Affairs). Radaelli, Claudio M. (1997). The Politics of Corporate Taxation in the European Union, (London: Routledge Research in European Public Policy). Sinn, Hans-Werner (1993). “Taxation and the birth of foreign subsidiaries,” in Herberg Heberg and go Van Long, eds., Trade, Welfare, and Economic Policies, Essays in Honor of Murray C. Kemp (Ann Arbor: University of Michigan Press) pp. 325–352. Summers, Robert and Alan Heston (1991). “The Penn-World table (mark 5): an expanded set of international comparisons, 1950–1988,” Quarterly Journal of Economics, 106, pp. 32–368. White, Fredrick (1991). “The United States perspective,” Tax Treaties and Local Taxes, International Fiscal Association, 16, pp. 15–24. Wilson, G. Peter (1993). “The role of taxes in location and sourcing decisions,” in A. Giovannini, R. G. Hubbard and Joel Slemrod, eds., Studies in International Taxation (Chicago: University of Chicago Press). Wilson, John D. (1999). “Theories of tax competition,” National Tax Journal, 52, pp. 269–304. United Nations (1998). Bilateral Investment Treaties in the Mid-1990s (New York: United Nations Conference on Trade and Development).

18. the effects of bilateral tax treaties on u.s. fdi activity∗ bruce a. blonigen and ronald b. davies introduction Empirical studies on the effects of taxation on foreign direct investment (FDI) are as numerous and varied as the tax policies they study. Until this point, however, the impacts of bilateral tax treaties governing the taxation of FDI activity have been unexplored.1 Worldwide, there exists a network of over 2,000 treaties which affect the taxation of the large majority of FDI (Radaelli 1997). According to the Bureau of Economic Analysis (BEA; 1998), by 1998 these treaties covered approximately $774 billion of U.S. investment abroad (outbound FDI) and $586 billion of foreign investment within the United States (inbound FDI). These amounts represented 78% of total U.S. outbound FDI and 96% of the total U.S. inbound FDI. The introduction to the Organization for Economic Cooperation and Development’s (OECD) model tax treaty states that a primary goal of a tax treaty is “removing the obstacles that double taxation presents,” thus reducing its “harmful effects on the exchange of goods and services and movements of capital, technology, and persons” (OECD, 1997, I-1). Janeba (1995), Davies (2003), and Chisik and Davies (forthcoming) model these goals and demonstrate that by lowering withholding

∗ This chapter was reprinted with permission from International Tax and Public Finance. The chapter was originally published as “The effects of bilateral tax treaties on U.S. FDI activity,” International Tax and Public Finance 11(5), 601 (2004). This chapter has benefited greatly from comments received from two anonymous referees, Reuven Avi-Yonah, Andrew Bernard, Ann Harrison, Stephen Haynes, Keith Head, Robert Lipsey, James Markusen, William Randolph, and seminar participants at the 2000 Empirical Investigations in International Trade conference, the National Bureau of Economic Research, the Fall 2000 National Tax Association meetings, the Summer 2000 Western Economic Association meetings and the Winter 2001 International Economic and Finance Society meetings. Blonigen acknowledges financial support from a Richard A. Bray Fellowship. Any errors or omissions are the responsibility of the authors. 1. Hines and Willard (1992) empirically examined the number of treaties a country signs as well as the tax concessions dictated by a particular treaty. However, they do not include the amount of FDI activity as an explanatory variable, nor do they consider the effect of taxation on FDI. A handful of recent studies have examined whether bilateral investment treaties (BITs) increase FDI activity. BITs more clearly address promotion and protection of FDI than bilateral tax treaties and do not address tax concerns. Empirical evidence from United Nations (1998) and Hallward-Dreimeier (2003) provides only limited support for positive effects of such treaties.

486 bruce a. blonigen and ronald b. davies

taxes on reparations to treaty partners and harmonizing tax laws treaties, tax treaties can increase FDI. Empirical evidence from Altshuler and Newlon (1991), Hines (1992), Altshuler, Newlon, and Randolph (1995), and Mutti and Grubert (1996) indicates that changes in withholding taxes affect both the timing of repatriation decisions and the mode of repatriation by firms. While this suggests that these treaties can affect the profitability of existing foreign investors, it does not necessarily say anything about the effects of these changes on overall FDI activity. Furthermore, there are several legal and economic arguments contrary to the potential FDI-promotion effect of tax treaties. First, it is not clear that reducing firms’ withholding tax burdens necessarily impacts the amount of investment. Hartman (1985) and Sinn (1993) point out that, since withholding taxes are only paid upon repatriation, it is less expensive to expand an overseas affiliate through retained earnings than through repatriated and re-exported funds. Because of this, they contend that only mature, non-expanding foreign affiliates will repatriate earnings. As Sinn notes, this implies that treaty-specified withholding tax reductions may have no effect on the size of FDI activity. In addition, there is the possibility that treaties stem from the rent-seeking lobbying efforts of current foreign investors rather than a desire on the part of governments to reduce distortions and enhance efficiency.2 Second, there are reasons to think that treaties may reduce the incentive to engage in FDI. Dagan notes that double taxation can be alleviated unilaterally just as well as through a bilateral agreement, and suggests that the actual intent of treaties is to reduce administration costs, reduce tax evasion, and extract tax concessions from host countries. In fact, she goes so far as to condemn the double taxation objective as “a myth” (Dagan 2000, 939). Radaelli (1997) and Gravelle (1998) also assert that U.S. tax treaties are primarily geared toward reducing tax evasion rather than promoting FDI. As discussed by the OECD (1994), treaties combat transfer pricing and treaty shopping (a practice in which investments are funneled through a treaty country by a third nation for the purpose of avoiding or reducing taxes) by promoting the exchange of information between tax authorities.3 Thus, if investment is taking place on the margin to avoid taxes, treaty formation may reduce FDI. These varied arguments indicate that the effect of bilateral tax treaties on the amount of FDI between the United States and partner countries is an open question. The answer is of particular importance in light of the efforts expended by the U.S. Treasury to negotiate the treaties, as well as the necessary ratification by the U.S. Senate. This chapter makes a first attempt at answering this question by estimating the impact of bilateral tax treaties on both U.S. inbound and U.S. outbound FDI over the period from 1980 to 1999. The sample we construct 2. Radaelli (1997) argues that, at least for the United States, treaty formation is wellinsulated from the interests of business groups. 3. For an explanation of transfer pricing, see Caves (1993). Graham and Krugman (1995) provide examples of firms recently prosecuted for transfer pricing by the United States.

the effects of bilateral tax treaties on u.s. fdi activity 487

represents easily the most comprehensive set of consistent data on FDI used in the literature to date, and allows us to examine the impact of new bilateral tax treaties for a wide variety of U.S. partner countries that occurred during the 1980s and 1990s. Our initial empirical framework for testing the determinants of FDI activity is one recently developed by Carr, Markusen, and Maskus (2001), as modified by Blonigen, Davies, and Head (2003). As opposed to the traditional gravity model, this framework is more strongly based on established theory of multinational enterprises (MNEs). However, we also consider modifications to the Carr, Markusen, and Maskus (CMM) framework that allow us to better fit the data. Robust to a wide variety of alternative empirical specifications, we find little evidence that bilateral tax treaties increase FDI activity, contrary to OECD-stated goals for such treaties. While there is some heterogeneity in individual country experiences, we generally find that the average new treaty effect is not statistically different from zero with a very small point estimate. This is true for both FDI into the United States and U.S. outbound FDI. Thus, our results find no support for the idea that tax treaties increase FDI, and may instead lend credence to the concerns expressed by Radaelli (1997) and Dagan (2000). The chapter proceeds as follows. In Section A, we discuss U.S. tax treaties including their history, how they are formed, and their key functions. In Section B, we present our empirical methodology and data. Section C presents our results and the last section concludes. A. Tax Treaties In this section, we briefly discuss the history and functions of bilateral tax treaties. Given our chapter’s focus, this section concentrates on these issues from a U.S. perspective; however, most of the discussion applies to other countries as well. The process of creating an international tax treaty with the United States is the same as for other international treaties. The Assistant Secretary for Tax Policy and the International Tax Counsel, acting on behalf of the Department of the Treasury, undertake the actual treaty negotiation. After the treaty is signed by the President or his or her delegate, it proceeds to the Senate. There, the Senate Foreign Relations Committee holds hearings before the treaty proceeds to the full Senate for ratification.4 Not all signed treaties receive Senate approval. For example, the Bangladeshi treaty was signed in 1981 but failed in the Senate. Some treaties fail multiple times before they are finally ratified.5 Note that because

4. Since tax treaties are federal instruments, they supercede state or local laws. Generally, the only limitation U.S. tax treaties place on state taxation is that a foreign corporation must be treated the same as a firm incorporated in another state (White 1991). 5. The original draft of the U.S.–Cyprus treaty was signed in 1981 but did not pass the Senate. An amended version also failed in 1985. In 1987, a final version of the treaty was signed but was not ratified until 1988.

488 bruce a. blonigen and ronald b. davies

of this delay, there can be a lag between when a treaty is signed and when it enters into force. Additionally, the enforcement date can differ from the effectiveness date, which is the year in which the treaty provisions initially applied. The effective date can come before or after the enforcement date and can even precede the signing date. Once entered into force, treaties usually remain in effect, although most contain provisions for termination after a six-month notification by either nation (Doernberg 1997). Of those treaties allowed to lapse, the majority (Aruba, Malta, and the Netherlands-Antilles) were the result of insufficient attempts by partner nations to prevent treaty shopping.6 As with many types of international agreements, the number of tax treaties the United States has in force is on the rise. Table 1, which lists U.S. treaty partners and the effective dates, shows that the United States has gone through three major waves of treaty negotiations. The first followed World War II and saw the United States complete treaties with its major economic partners: the western European countries, Canada, Japan, Australia, and New Zealand. The second wave of the late 1970s and early 1980s focused on a wider variety of countries, including China, Egypt, Korea, India, Indonesia, the Philippines, and Spain. More recently, beginning in the early 1990s, the fall of the former Soviet Union has led to additional U.S. treaty activity with respect to many of the former Soviet Bloc countries. Overall, the pace of new treaty implementation has increased. In fact, half of the treaties currently in force were ratified only in the past twenty years, with nearly one-third of the current treaties implemented since 1990. Tax treaties perform four primary functions. The first is to standardize tax definitions and solidify the tax jurisdictions of treaty partners. It is common practice that only income generated by foreigners through a permanent establishment in the host country is subject to host taxation. If the countries differ in their definitions of permanent establishments, this can lead to double taxation and inefficient capital flows (Hamada 1966). The impact of differing tax definitions on FDI is explored theoretically by Janeba (1996). Hines (1988) finds that the 1986 Tax Reform Act, which revised U.S. definitions, had a significant effect on U.S. MNEs. Since a common goal for tax treaties is to reduce double taxation and the inefficiencies it causes, standardization of tax definitions and jurisdictions is a powerful tool. Treaty tax definitions generally match those provided by the OECD’s model tax treaty (OECD 1997).

6. The treaty with South Africa was terminated as part of the Comprehensive AntiApartheid Act of 1987. It was reinstated in 1997. The U.S.–U.S.S.R. treaty, which was officially terminated in 1992, has been extended to cover former Soviet block countries until individual treaties can be negotiated. The only terminated treaty which was not cancelled by the U.S. was that with Honduras. This treaty eliminated withholding taxes on FDI by both countries. Due to the almost entirely one-way nature of FDI flows, Honduras felt that all gains from the treaty were accruing to the U.S. and terminated it in 1967 (Diamond and Diamond 1998).

the effects of bilateral tax treaties on u.s. fdi activity 489

table 1. list of u.s. tax treaties (years effective) Australia (1953–) Belgium (1953–) Cyprus (1985–) Egypt (1981–) France (1936–) Honduras (1957–1966) India (1990–) Israel (1994–) Japan (1955–) Latvia (1999–) Malta (1982–1995) Netherlands (1948–) Pakistan (1959–) Portugal (1995–) Slovakia (1993–) Spain (1990–) Thailand (1997–) Turkey (1997–) U.S.S.R. (1976–)b

Austria (1957–) Canada (1937–) Czech Republic (1993–) Estonia (1999–) Germany (1954–)a Hungary (1979–) Indonesia (1990–) Italy (1956–) Kazakhstan (1996–) Lithuania (1999–) Mexico (1993–) New Zealand (1951–) Philippines (1982–) Romania (1976–) Slovenia (2001–) Sweden (1939–) Trinidad and Tobago (1967–) Ukraine (2000–) Venezuela (1999–)

Barbados (1986–) China (1986–) Denmark (1948–) Finland (1952–) Greece (1953–) Iceland (1975–) Ireland (1951–) Jamaica (1981–) Korea (1979–) Luxembourg (1964–) Morocco (1981–) Norway (1951–) Poland (1976–) Russia (1993–) South Africa (1952–1987,1997–) Switzerland (1951–) Tunisia (1990–) United Kingdom (1946–) –

Notes: a Extended to cover reunified Germany in 1990. b Treaty provisions extended to former members until individual treaties can be negotiated. Source: Worldwide Tax Treaties Database at tax.com.

A second role for tax treaties is to reduce transfer pricing and other forms of tax avoidance. They do so in three key ways. First, they influence the method used by tax authorities to calculate the internal price used by firms.7 Second, they establish rules for resolving conflicts between tax authorities or between governments and firms. Third, they promote the exchange of tax information. Information exchange can be viewed narrowly or broadly. The narrow view limits exchanges to those necessary to implement the provisions of the treaty. The broad view uses exchanges to help in overall tax policy implementation. The OECD’s model treaty takes the broad view in its commentary, where it states that information exchange is intended to be used to the widest possible extent (OECD 1997). The United States has taken this to heart and is one of the most vigorous users

7. The three most common provisions are the comparable uncontrolled pricing, cost plus pricing, or resale pricing methods.

490 bruce a. blonigen and ronald b. davies

of information exchange (Easson 1999).8 Furthermore, the United States has made information exchange a requirement for continuing existing treaties, and has cancelled some treaties for their failure to meet this requirement (Eden 1998). Information exchange is also linked to the third goal of tax treaties, which is to prevent treaty shopping. According to Ault and Bradford (1990) and Radaelli (1997), this is a primary focus of recent treaties and has prompted the United States to renegotiate many of its older treaties as well. Although there are many variations in the regulations regarding treaty shopping, the most common rules restrict treaty benefits if more than 50% of a corporation’s stock is held by a third, non-treaty country’s residents (Doernberg 1997). Finally, tax treaties affect the actual taxation of MNEs. They do so through the rules affecting double taxation relief and the withholding taxes levied on repatriations by FDI. Following the OECD model treaty guidelines, tax treaties with the United States specify that both countries must either offer foreign tax credits when calculating the domestic tax bill or exempt foreign-earned profits from domestic taxation.9 This does not usually affect U.S. tax policy as it already provides foreign tax credits to its investors whether their income is earned in a treaty or a non-treaty country. It does, however, affect the taxation of some treaty partners.10 In addition to the provisions for double taxation relief, the treaties also often reduce maximum allowable withholding taxes on three types of remitted income: dividend payments, interest payments, and royalty payments. There are no tax treaties that raise tax rates. From the above discussion, the primary costs and benefits of entering into a tax treaty can be summarized as follows. Since a treaty can lower the overseas taxes, more income is repatriated to the home country. At the same time, lower overseas tax rates can promote a more efficient global allocation of investment.11

8. The United States has also entered into a small number of tax information exchange agreements with countries that are not tax treaty partners. Additionally, the United States ratified the multilateral OECD Convention on Mutual Assistance in Tax Matters, which provides for information exchange (OECD 1989). However, many OECD members have not ratified this agreement so its usefulness is limited. 9. Under a tax credit, the domestic tax bill is calculated by applying the standard domestic tax rate to the pre-foreign tax level of overseas profits. A credit against this amount is then applied up to the amount of the foreign taxes paid. If this credit exceeds the domestic tax liability, the firm is in an “excess credit” position and pays no additional taxes on these overseas profits. If the home tax bill is greater than the amount of the credit, the firm is in an “excess limit” position and pays the remaining amount to the home government. 10. For instance, Belgium applies a reduced domestic tax rate to the foreign-earned profits of its residents. Under the U.S.–Belgium treaty, however, income earned by Belgian firms in the United States is exempt from Belgian taxation. 11. Note that a lower foreign tax rate does not guarantee a reduction in the firm’s total tax bill. As noted by Altshuler and Newlon (1991), only firms in excess credit will benefit

the effects of bilateral tax treaties on u.s. fdi activity 491

This gain is accompanied by potentially reduced tax receipts since treaties also lower the tax rates levied on inbound investment. This cost can be mitigated in two ways. First, improved international cooperation may reduce tax evasion, and can also reduce the administrative costs associated with enforcement and tax collection. Second, since withholding taxes can be tailored to the specific investment from a treaty partner, it may be possible to set tax rates that encourage tax-sensitive inbound investment and actually raise total tax receipts.12 Finally, a tax treaty, regardless of its specific terms, can promote investment by reducing uncertainty about the overseas tax environment. Since a treaty can signal the willingness of governments to work out differences, this may boost investor confidence, which leads to increased willingness to export capital.13 These treaty provisions, however, can cut both ways and may instead reduce the incentive to engage in FDI. If firms maximize after-tax profits, some of them may choose to operate in multiple countries in an effort to reduce their taxes. For example, it is well known that MNEs can minimize their tax burdens through transfer pricing. By manipulating the price of goods traded between their various subsidiaries, firms can shift profits to low-tax locations. As put forth by Casson (1979), this provides an incentive to invest in those low-tax locations in order to shield profits from taxes. A treaty that reduces the ability to transfer price also reduces the incentive to invest for tax-minimization reasons. Additionally, because treaties streamline the tax environment and encourage the exchange of tax information, a treaty may make it easier for governments to reduce other types of tax evasion, such as misrepresentation. Therefore, to the extent that some firms engage in FDI simply to minimize taxes, tax treaties may reduce FDI activity, not increase it. Thus, the effect of tax treaties on FDI activity is an open question, and we next turn to the data to estimate this effect for U.S. FDI activity. B. Empirical methodology and data 1. Empirical methodology To estimate the effect of U.S. bilateral tax treaties, one requires an empirical model of MNEs’ FDI activity that reasonably captures the effect of other factors. While variations of a gravity model have perhaps been

from this change. Thus, a decrease in the overseas tax rate may not improve capital flows. Also, as discussed by Ramaswami (1968) and Bond and Samuelson (1989), this effect may be to the detriment of the home country due to foreign factor-market effects. 12. As derived in Bond and Samuelson (1989) among others, the tax revenue–maximizing tax rate on inbound FDI is 1/(1+ε), where ε is the elasticity of inbound FDI supply. If ε varies across countries, the revenue-maximizing tax rate will differ across countries. Since a treaty allows a country to lower its tax rate, this can actually raise tax revenues if the current tax rate is greater than the optimal tax on FDI from that particular country. 13. See Jones (1996) or Sasseville (1996) for additional discussion of the effects of treaties on tax certainty.

492 bruce a. blonigen and ronald b. davies

the most popular empirical framework for examining FDI activity across countries, they lack connection to a recognized formal theory of MNE activity. In contrast, Carr, Markusen, and Maskus (2001) and Markusen and Maskus (2001, 2002) establish and test an empirical specification of FDI activity that is arguably more grounded in the formal theories of MNE activity. The CMM specification is based on the Markusen’s (2002) knowledge-capital MNE model, which allows for both the horizontal and vertical motivations for FDI. This specification for the FDI from country i to country j is: FDIij = f (SUMGDPij, GDPDISQ ij, SKDIFFij, SKDGDPDij,

T _ COSTi, T _ COSTj, F _ COSTj, HTSKDij, DISTij) The first two terms control for partner countries’ sizes where SUMGDP is defined as the sum of the two countries’ real Gross Domestic Products (GDPs), and GDPDIFSQ is defined as the squared difference between the two countries’ real GDP. Since horizontal MNEs (those that invest to gain access to the host market) are most common between large countries of similar size, there is an expected positive correlation between SUMGDP and FDI activity and an expected negative correlation between GDPDIFSQ and FDI activity. The intuition is that with some positive level of trade frictions, larger and more similarly-sized markets better support the higher fixed costs associated with setting up production across countries (versus exporting), and lead to greater MNE activity. The next two terms capture relative factor endowment effects. SKDIFF is the skill difference between the home and host country and is intended to capture relative factor abundance motivations for FDI. According to the CMM interpretation of the knowledge-capital model, this variable should have a positive coefficient. Greater skill differences should proxy for greater wage differences, which encourages vertical FDI by firms looking to outsource some of their activities to low-wage areas. As demonstrated by Blonigen, Davies, and Head (2003), however, interpretation of estimated coefficients on difference variables is sensitive to whether the variable takes negative or positive values. With bilateral U.S. FDI data, observations of inbound FDI find that the skill difference between the home and host country (here, the United States) almost always has a negative value. In these cases, the coefficient on SKDIFF will tell us the effect on FDI as skill differences decline, that is, as SKDIFF becomes less negative. Therefore, we allow the SKDIFF variable to have different coefficients depending on whether our SKDIFF variable is positive or negative in value. To more easily interpret coefficient estimates, we multiply our SKDIFF term by –1 for the U.S. inbound sample, for which SKDIFF is always negative. Thus, for both the inbound and outbound samples, a positive coefficient on SKDIFF will indicate that MNE activity increases as skill differences increase, as hypothesized by CMM. However, Blonigen, Davies, and Head (2003) estimated a statistically significant negative coefficient on SKDIFF in their analysis, suggesting that skill similarities increase FDI. One explanation for this puzzling result is that skill similarities are proxying for

the effects of bilateral tax treaties on u.s. fdi activity 493

other (omitted) factors that increase FDI between developed countries beyond those captured by the GDP variables. Given that we use similar data to Blonigen, Davies, and Head, our empirical (not theoretical) expectation is a negative coefficient on the SKDIFF variable. The fourth term, SKDGDPD, interacts the skill difference with the difference in GDPs, which CMM hypothesizes will have a negative coefficient. Finally, the last five terms capture trade and investment frictions. T_COST is the trade cost of either the home or host country. Lower trade costs in the host country should lower FDI activity because they make exporting more attractive relative to FDI. Lower trade costs in the home country should increase FDI, since they make it easier to ship goods back to the home country from foreign affiliates. F_COST is the cost of investing in the host. As this rises, FDI should fall. HTSKD interacts host trade costs with the squared skill difference. Since host trade costs should matter less when FDI is vertical (i.e., skill differences are large), this carries an expected negative coefficient. Finally, DIST is the distance between countries. Since higher distances make both trade and control of overseas investment more difficult, the net effect is ambiguous. Table A1 in the Data Appendix provides descriptive statistics of our variables for various specifications we use in our analysis. As discussed earlier, tax treaties may alter investment incentives in a variety of ways. Unfortunately, most of these are difficult to quantify, and it is unclear how important the few quantifiable aspects are for the overall tax treaty effect. Thus, to examine bilateral tax treaty effects, we begin with inclusion of a simple binary variable in the CMM empirical framework, which takes the value of “1” when an effective U.S. treaty is in place for a given partner country and year. From this initial examination, however, we employ a variety of alternative specifications of this bilateral tax treaty variable to explore the heterogeneity of tax treaty effects across countries and over time. One additional concern is endogeneity of our dependent variable and the tax treaty variable. The United States may be negotiating treaties with only the countries for which there are large and/or increasing amounts of FDI activity. Such an endogeneity problem would bias our coefficient on tax treaties upwards and possibly cause a spurious positive correlation. In our sample, there are two types of U.S. treaty partners: those with which the United States has had a bilateral tax treaty before our sample begins in 1980, which we call old treaty partners, and those with which treaties are negotiated during our sample spanning the 1980s and 1990s, which we call new treaty partners. Endogeneity problems are likely largest with the old treaty partners, which are primarily composed of Canada, Japan, Australia, New Zealand, and the western European countries. However, since these treaties were in place well before our sample begins, we are already skeptical about our ability to separately identify tax treaty effects on FDI activity from other uncontrolled factors. Thus, our focus will be on the effects for new treaty countries whose treaties with the United States came into place after the

494 bruce a. blonigen and ronald b. davies

beginning of our sample. For these partners there is little evidence of an endogeneity problem, as we next show. Table 2 lists each new treaty country for which we have data, the rank of the new treaty country in terms of total FDI activity with the United States relative to other countries at the time of the treaty, and the number of countries ranked ahead of the new treaty country that did not yet have a bilateral tax treaty with the United States14 Table 2 clearly shows that the United States was not pursuing a strategy of signing bilateral tax treaties with countries that had the largest FDI activity with the United States during the 1980–1999 sample period. To the extent that one may believe endogeneity exists to bias our coefficient toward showing a positive relationship between new tax treaties and FDI, the insignificant effects we report below are even more surprising.15 On a final note, we run empirical specifications on separate samples of U.S. outbound and U.S. inbound FDI activity. This allows us to examine whether bilateral tax treaties asymmetrically impact outbound FDI of U.S. firms versus FDI into the United States by foreign firms. Splitting into inbound and outbound samples also gets around Blonigen, Davies, and Head’s (2003) critique of CMM. 2. Data Empirical FDI studies have always been hampered by data difficulties. Missing or incomplete data for one or more variables often leads researchers to examine cross-sectional data, with little or no time series dimension.16 Data issues are also why researchers generally focus on U.S. data, which is more complete and detailed than the FDI statistics for virtually any other country. Our approach is to collect data on inbound and outbound U.S. FDI activity for as many years as possible across as many countries as available. As a result, 14. Table 2 also reports that for a number of countries, primarily former Soviet Bloc countries, data on FDI activity with the United States before the treaty was signed is unavailable, making it difficult to estimate the treaty effect just as in the case of old treaty partner effects. These treaty countries are the Czech Republic, Estonia, Kazakhstan, Latvia, Lithuania, Russia, and Slovakia. We designate these countries as new treaty partners in our sample, but note that they have no impact on our new treaty coefficient once we employ country fixed effects. 15. One referee of this chapter made the point that Latin American countries have historically refused to sign tax treaties with the United States because they did not want to lose tax revenue on a significant stock of existing FDI. In fact, Chisik and Davies (2004) find that this concern exerts influence over both the existence of a treaty and the withholding taxes setout therein. In any case, this idea would also argue against the existence of any systematic trend for the United States to sign up large FDI partners to treaties, especially with developing economies for which such taxes are an important source of revenue. 16. Grubert and Mutti (1991) examine a cross-section of 33 countries for 1982, while Brainard (1997) samples U.S.-country-industry combinations across 63 tradable industries and 27 countries for 1989. Markusen and Maskus (2001, 2002) examine an unbalanced panel data set of 36 U.S.-country pairs from 1986–1994. In contrast, this study examines an unbalanced panel of up to 88 countries for up to 20 years.

the effects of bilateral tax treaties on u.s. fdi activity 495

table 2. relative rank of new u.s. bilateral tax treaty countries in u.s. outward fdi position in the year the treaty is effective Year

Country

1981 1982

Egypt Jamaica Philippines

30th 44th 29th

14 22 15

1985

Cyprus

87th

51

1986

China

58th

30

1990

1993

India Indonesia Spain Tunisia Mexico

49th 22nd 15th 76th 11th

23 8 4 42 2

1994

Israel

40th

16

1995

Portugal

43rd

18

1997

Thailand Turkey Venezuela

32nd 56th 27th

11 20 8

1999

Rank of country in terms of U.S. outward FDI stock

Number of countries ahead in rankings with no treaty

Notes: This table excludes those countries (both with and without treaties) for which FDI data were not available. Data may not be available for a number of reasons, but in many cases they were publicly suppressed since they had such small volume that they would reveal proprietary information on individual investing firms. The table also excludes new treaty activity with former Soviet bloc countries, since there are virtually no inbound or outbound U.S. FDI data for these countries before the treaty comes into place. Country rankings are relative to all countries listed by the U.S. Bureau of Economic Activity table of outbound FDI activity for the particular year. Source: U.S. Bureau of Economic Activity and Worldwide Tax Treaties Database at tax.com.

we have an unbalanced panel of 88 countries spanning up to 20 years from 1980 through 1999. As is typical in virtually all FDI studies, the data are for activity in non-financial sectors only. Previous studies have typically measured FDI activity through affiliate sales or FDI stock in the host country. We use FDI stock as our dependent variable, as this provides a greater number of observations than affiliate sales data. We note, however, that our estimates are qualitatively identical when using affiliate sales as the dependent variable.17 Data for FDI activity variables were collected from 17. Affiliate sales data are only available back to 1983 for outbound affiliate sales and 1984 for inbound affiliate sales. As one would guess, there are strong, statistically significant

496 bruce a. blonigen and ronald b. davies

official BEA data available at the BEA’s Internet site and converted into millions of real 1996 U.S. dollars using the U.S. chain-type price index for gross domestic investment as reported in the Economic Report of the President. A Data Appendix provides more details on our dependent variable, as well as the other variables we employ in our study. With respect to our independent variables, data on real GDP come from the Penn-World Tables and are scaled in trillions of real U.S. dollars. Data on skill differences come from Barro and Lee International Data on Educational Attainment, which reports the average years of schooling for those aged twentyfive or more. These data are measured every five years, so we used linear interpolation for in-between years. This measure differs from that of CMM, who use occupation information as a proxy for skill. Our measure yields similar predictions regarding skill differences and FDI, but covers a longer time period and more countries. Distance data come from the Bali Online Corporation (1999) and are measured in miles between capital cities. Following previous studies, trade cost measures come from the Penn-World Tables and are defined as 100 minus the ratio of the sum of imports and exports to GDP. In an ideal setting, we would decompose trade costs into institutional, political, and geographic components. However, such variables are difficult to come by and would limit our sample by excluding many of our developing countries. In any case, since an increase in any of these barriers would reduce the share of a country’s economic activity devoted to international trade, our chosen measure acts as a reasonable proxy for general trade barriers and allows us to focus on our variable of interest. To construct our measure of investment barriers, we use the composite score compiled by Business Environment Risk Intelligence, S.A. (BERI). This composite includes measures of political risk, financial risk, and other economic indicators, and ranges between zero and 100, with higher numbers signifying more openness. To construct a cost measure, we use 100 minus the BERI’s composite score. The BERI measure allows us to consider more countries over a longer time period than CMM. There is a strong relation between the two, with a correlation of 0.81 for the observations in the CMM database. Finally, information on U.S. bilateral tax treaties was collected from the Worldwide Tax Treaties database at tax.com (2002). As discussed above, there are three dates that we could use for each treaty: the date signed, the date in force, and the date effective. The date in force is no earlier that the date signed, however the effective date can pre-date the signing of the treaty or even follow the enforcement date. While we report the results using only the effective date, we get similar results regardless of which measure we use. These additional results are available upon request. correlations between affiliate sales and FDI stock. In our sample, the pairwise correlation is 0.92 for inbound FDI activity and 0.90 for outbound FDI. The results from the affiliate sales data are available on request.

the effects of bilateral tax treaties on u.s. fdi activity 497

C. Results 1. Initial results We begin our exploration of the effect of bilateral tax treaties by using initial specifications that follow the CMM specification, adding only treaty variables. Following Blonigen, Davies, and Head (2003), we examine inbound and outbound samples separately, and note that when we use U.S. inbound FDI data, subscript i refers to the foreign parent country and subscript j refers to the United States (the host country), while when we use U.S. outbound FDI data, subscript i refers to the United States (the home country) and subscript j refers to the foreign host country. As discussed above, we use FDI stock data for our dependent variable. Following our discussion of these initial CMM specification issues in this subsection, statistical problems with our estimates will lead us to explore alternative specifications in following subsections. Table 3 presents estimates for determinants of both inbound and outbound U.S. FDI stock using the CMM specification with various tax treaty variables. We begin with a CMM specification that includes only a simple binary variable for whether the observation involves an effective treaty country or not, with column 1 displaying these results for inbound FDI stock and column 3 displaying these results for outbound FDI stock. For both the inbound and outbound samples, the data appear to fit the overall model reasonably well, with R2s around 0.35 and an F-test statistic that rejects the null of zero slopes at the 1% significance level. The primary variables of the CMM model (SUMGDP, GDPDIFSQ, SKDIFF) are generally statistically significant and have estimated signs that are consistent with a world that is dominated by horizontal MNEs. In particular, the negative coefficient on the SKDIFF variable that suggests that FDI activity is greater for country pairs with lower skill differences, not high skill differences as predicted by a vertical model of MNE activity. The trade cost terms and interaction terms (SKDGDPD and HTSKD) are generally of incorrect sign, while the FDI cost variable and distance both display statistically significant negative impacts on FDI activity. Our variable of interest, the indicator variable for the presence of an effective tax treaty between the United States and a given country in a given year, is estimated with a negative coefficient for both the inbound and outbound samples, with the inbound estimate statistically significant at the 1% significance level. In columns 2 and 4 we relax the assumption that bilateral treaty effects are similar across old treaty countries and new treaty countries. The estimates find that old treaties are significantly and positively related to both inbound and outbound FDI, whereas new treaties are significantly and negatively related to FDI in both directions.18 Such results suggest that, at least for recent treaties, the role 18. An earlier draft of this chapter (Blonigen and Davies 2000) found no significant effect for the simple treaty dummy variable. In that version, we then moved toward a “treaty age” variable to measure vintage effects. There, we found that older treaties were significantly and positively correlated with FDI activity. Initially, we attributed this to

498 bruce a. blonigen and ronald b. davies

of treaties in enforcing anti-tax avoidance measures may well outweigh any FDI-promotion effects. The magnitudes of the treaty variable coefficients in Table 3 are quite large. For example, a new treaty is estimated to lower inbound FDI stock by $9.9 billion annually even though the average inbound stock for these countries is just $329 million. Likewise, outbound FDI stock is estimated to be $7.8 billion lower as compared to an average of $3.05 billion of outbound FDI stock in new treaty partner countries. The estimated new treaty effects from these specifications are therefore implausibly large, and hence we next explore alternative specifications to obtain estimates with more reasonable magnitudes. 2. Alternative specifications One concern with our estimates in Table 3 is that they are driven by unobserved characteristics not controlled for by the CMM independent variables. In fact, the residuals from these specifications without treaty variables included show that rich partner countries have large positive residuals, while non-rich partner countries have large negative residuals.19 Since old treaty partners are all rich countries and most new treaty partners are not, this systematic bias is likely captured by our treaty variables, leading to spurious inferences. A first step to mitigate this problem is the inclusion of country fixed effects. Unfortunately, this means we can no longer estimate old treaty partner effects, since these treaties became effective before our sample begins. Thus, the old treaty effects are subsumed into the fixed-effects coefficients and cannot be separately identified. Columns 1 and 5 of Table 4 present estimates that include country-specific fixed effects to filter out time-invariant country-specific heterogeneity. As one might guess, these fixed effects are quite important and their inclusion leads to R2s around 0.80 with F-tests strongly supporting their inclusion. Because these fixed effects are so important for fitting the data, we necessarily focus on estimating treaty effects for only the new treaty partners in the rest of the chapter, and concede that data constraints do not allow us to credibly estimate treaty effects for old treaty partners.20

significant lag effects from the treaties. Given the current results, however, the evidence suggests that this effect is driven by the particular experience of the old treaty countries. Unfortunately, given the limitations of the data set, we cannot say more on the effects of older treaties. 19. We define rich partner countries as the European Union countries, as well as Austria, Australia, Canada, Finland, Hong Kong, Japan, New Zealand, Norway, Sweden, and Switzerland. 20. Credible estimation of the old treaty effects would first require consistent data on FDI activity back to at least World War II. In addition, since these old-treaty countries were clearly the United States’ main economic partners, endogeneity is likely a greater concern. Finally, one would have to control for the many economic factors, such as post-war reconstruction, that were occurring at the time.

the effects of bilateral tax treaties on u.s. fdi activity 499

table 3. estimates of treaty effects on u.s. inbound and outbound fdi stock: initial results Expected Sign

Inbound FDI Stock

Outbound FDI Stock

Pooled

Pooled

Pooled

Pooled

–

–1220∗ (–1.68) –

–

TREATYij

?

OLD TREATYij

?

–2174∗∗ (–2.51) –

NEW TREATYij

?

–

CMM Controls SUMGDPij

+

GDPDIFSQij

–

SKDIFFij

–

SKDGDPDij

–

F_COSTj

–

T_COSTi

–

HTSKDij

–

T_COSTj

+

DISTij

?

Constant

–

Observations R2

– –

7.4∗∗∗ (4.74) –0.001∗∗∗ (–5.90) –12774∗∗∗ (–8.43) 0.6∗∗∗ (4.14) –610.6∗∗∗ (–2.82) –2115∗∗ (–2.44) 7.8∗∗∗ (9.41) –51.0∗∗∗ (–6.18) –1.4∗∗∗ (–10.51) 230840∗∗∗ (2.65) 1470 0.35

7172.4∗∗∗ (4.96) –9879.0∗∗∗ (–8.79)

6.6∗∗∗ (4.48) –0.001∗∗∗ (–5.47) –6873∗∗∗ (–4.52) 0.3∗ (1.82) –500.1∗∗ (–2.40) –2359.7∗∗∗ (–2.84) 4.5∗∗∗ (5.52) –45.3∗∗∗ (–5.86) –1.1∗∗∗ (–9.23) 232328∗∗∗ (2.79) 1470 0.41

–

–0.04 (0.02) –8.9e–4∗∗∗ (–4.15) –2231∗ (–1.67) 0.02 (0.12) –326.7∗∗∗ (–4.90) –72.2∗∗∗ (–3.11) 1.3∗∗∗ (2.90) –3534∗∗∗ (–3.20) –2.4∗∗∗ (–7.66) 372367∗∗∗ (3.40) 871 0.30

6518∗∗∗ (4.87) –7798∗∗∗ (–6.30)

0.5 (0.30) –9.0e–4∗∗∗ (–4.26) –1023 (–0.75) –0.1 (–0.26) –65.4 (–0.91) –65.3∗∗∗ (–3.03) 0.4 (0.80) –3694∗∗∗ (–3.41) –2.1∗∗∗ (–7.59) 362877∗∗∗ (3.43) 871 0.34

Notes: Robust t-statistics are in parentheses, with ∗∗∗, ∗∗, and ∗ denoting statistical significance (two-tailed test) at the 1%, 5%, and 10% levels, respectively.

Inbound FDI stock

NEW TREATYij

CMM Controls SUMGDPij GDPDIFSQij SKDIFFij SKDGDPDij F_COSTj T_COSTi

Outbound FDI stock

Levels, fixed effects

Logs, fixed effects

Logs, fixed effects

Logs, fixed effects

Levels, fixed effects

Logs, fixed effects

Logs, fixed effects

Logs, fixed effects

–8570∗∗∗ (–6.22)

0.17 (0.63)

0.39 (1.44)

0.30 (1.12)

–5206∗∗∗ (–4.68)

0.06 (0.53)

0.02 (0.13)

0.07 (0.67)

10.1∗∗∗ (4.06)

8.0∗∗∗ (4.83) –3.5∗∗∗ (–4.80) –4.9∗∗∗ (–4.64) – –0.1 (–0.07)

87.7∗∗∗ (3.79)

3.5∗∗ (2.06)

15.4∗∗∗ (2.75) –4.3∗∗∗ (–4.39) – 63.5∗∗∗ (3.44)

–1951∗∗∗ (–4.15)

4.8∗ (1.73)

4.8 (1.27)

–0.16∗∗∗ (–2.74)

7.3∗∗∗ (2.88) –3.4∗∗∗ (–3.40) 0.8 (1.16) – –1.1 (–1.45) –0.3∗∗∗ (–3.46)

8.5 (0.63)

–8.3e–4∗∗∗ (–4.31) –3991 (–1.31) –0.1 (–0.69) –344.7∗ (–1.84)

8.5∗∗∗ (6.02) –2.7∗∗∗ (–4.13) –1.3∗∗∗ (–5.94) – –0.3 (–0.50)

6.7∗∗∗ (3.15) –3.8e–4∗∗∗ –3.2∗∗∗ (–3.07) (–3.63) 7192∗∗∗ 0.2∗ (5.47) (1.86) –0.6∗∗∗ (–4.92)– 177.3∗∗ (2.56) –0.6∗ (–1.68) –100.1∗∗∗ –0.2∗∗∗ (–4.67) (–3.64)

–3.4 (–1.33) –1.1∗∗∗ (–4.36) – 1.6∗∗∗ (3.98) –0.1 (–0.03)

500 bruce a. blonigen and ronald b. davies

table 4. estimates of treaty effects on u.s. inbound and outbound fdi stock: alternative specifications

HTSKDij T_COSTj FXij TAXj

Observations R-squared

–

–

–

–

–

–

–2.3 (–0.54) 0.05 (0.93) –

2.0∗∗∗ (4.79) –2357∗∗∗ (–4.14) – –

–0.1∗∗∗ (–2.92) – –

–0.1∗∗ (–2.47) – –

–6.8∗∗∗ (–2.92) – –

–5.6 (–1.34) – –

Yes

Yes

No

No

Yes

–0.03 (–1.05) 0.01 (0.74) –0.14∗∗ (–2.21) Yes

No

No

133051∗∗∗ (2.69) 1470 0.81

–41.9∗∗ (–2.14) 1470 0.91

–23.1 (–1.12) 1470 0.91

–1296.2∗∗∗ (–3.47) 1423 .91

167174∗∗∗ (2.89) 871 0.80

37.7∗∗ (2.22) 871 0.83

30.7∗ (1.68) 871 0.83

–4.4 (–0.03) 753 0.94

Notes: Robust t-statistics are in parentheses, with ∗∗∗, ∗∗, and ∗ denoting statistical significance (two-tailed test) at the 1%, 5%,and 10% levels, respectively.

the effects of bilateral tax treaties on u.s. fdi activity 501

Rich-Country Interactions Constant

9.9∗∗∗ (4.62) 44.5∗∗∗ (3.87) – –

502 bruce a. blonigen and ronald b. davies

After controlling for country-specific fixed effects, we still find that the coefficients on the new treaty dummy variable in columns 1 and 5 of Table 4 are negative and statistically significant for both U.S. inbound and outbound FDI stock. Though the magnitude of the new treaty coefficient falls somewhat, these numbers remain implausibly large. While the inclusion of fixed effects means residuals for any group of countries (such as rich ones) are zero on average, differing trends between groups may still remain. Specifically, over the time dimension of our sample the rich countries’ average residuals become increasingly positive, while the poor countries’ average residuals grow increasingly negative. For both, each additional year in time changes the magnitude of the residual approximately $400 million in absolute value. Thus, the new treaty variable, which tends to kick in later in the sample, potentially reflects this negative trend in the residual of poor countries. We tried a number of alternative specifications to address these concerns, including year dummies, trend terms, separate rich country interactions with all our independent variables, and first differencing of the data with inclusion of country-partner-specific fixed effects.21 Nevertheless, these approaches continued to yield implausible coefficient estimates for the new treaty variable. A different approach is to log the data. The linear model proposed by CMM is not a structural equation derived from theory, so there is nothing inherently inconsistent with specifying a log-linear model. This alternative is particularly useful because the data are highly skewed. For example, the standard deviation of inbound FDI stock in our sample is almost four times larger than the mean. In particular, rich country inbound FDI is almost twenty times larger than inbound FDI for other countries. Logging the data reduces this skewness considerably. Furthermore, doing so leads to new treaty estimates that are plausible in size, free of many of the above-noted statistical problems, and quite robust. One issue in logging the data is dealing with negative values of the dependent variable and trade cost measures for some observations. We truncate these observations to 0.1 before taking logs, although we get qualitatively similar results if we simply drop these observations. Additionally, note that once logged, the interaction terms in the CMM model, SKDGDPD and HTSKD, become collinear with the main control variables and must therefore be dropped.22 Despite these changes, logging the data improves the fit of our regressions. Columns 2 and 6 of Table 4 provide estimates when we log our data, but otherwise use the same fixed-effect specification used in columns 1 and 5 with the above-noted exceptions.

21. Inclusion of country-partner-specific fixed effects with first-differenced data will control for country-partner-specific trends in FDI activity. These additional results are available on request. 22. The treaty variable is still in levels and takes either the value of zero or one.

the effects of bilateral tax treaties on u.s. fdi activity 503

The fit of the log specification for both the inbound and outbound samples is high with R2s of 0.91 and 0.83, respectively. The control regressors are generally of the expected sign and are statistically significant. However, there is substantial change in the treaty coefficient. In both the inbound and outbound samples the treaty coefficient is now slightly positive, although we cannot reject the null hypothesis that the effect is zero. Examining residuals from this regression, we still find that there is a statistically significant increasing trend in the residuals for the rich countries and a significant decreasing trend for non-rich countries. However, the magnitude of these trends is quite small. Nevertheless, in columns 3 and 7 of Table 4, we include rich-country interactions with our regressor matrix for the logs specification, which eliminates any rich or non-rich trends in our residuals. While the rich-country interactions are jointly significant at the 1% significance level, suggesting that their inclusion is important, the coefficient estimates of our original regressors remain qualitatively similar. We no longer estimate a significantly negative effect of new treaties, but continue to find that there is no significant positive effect. We make one final modification to achieve our preferred specification. The CMM framework stems from a long-run general equilibrium model. In our panel data, however, FDI patterns may be substantially influenced by short-run factors. In columns 4 and 8 of Table 4 we include the log of the exchange rate (FXij) and year dummies in our specifications. We also include the log of a corporate tax rate variable (TAXj, which is described in the Data Appendix) in the outbound regressions. This reflects the varying tax rates faced by U.S. firms in host countries.23 While the tax rate variable is negative and significant in the outbound regression, as expected, the inclusion of these short-run factors does little to the new treaty coefficient. While our modifications to the CMM framework eliminate systematic trends in our residuals for rich and poor countries, a Shapiro and Wilk (1965) test for normality of the residuals continues to reject the null hypothesis of normality. Likewise, a misspecification test by Ramsey (1969) examines whether higher orders of the fitted dependent variable have additional explanatory power in the regression equation; we find that this is indeed true in all of our specifications, which rejects the null hypothesis of no omitted variables. In response, we have tried numerous alternative specifications, including first differences in logs and inclusion of squared terms of all the variables. None of these specifications allows us to pass the Shapiro and Wilk or Ramsey tests. However, none of these alternative specifications finds significant effects of new treaties either. This clearly illustrates the need for continued efforts to develop better, econometrically robust empirical models of FDI.

23. A similar variable for inbound U.S. FDI would capture only changes over time in the U.S. corporate tax rate. This is already captured in the year dummies.

504 bruce a. blonigen and ronald b. davies

Another issue we address is the binary nature of our treaty variable. As mentioned above, we choose this representation of bilateral tax treaties because such treaties often stipulate a number of changes between countries that are difficult to quantify. However, the most quantifiable difference in various treaties is the stipulated maximum withholding tax rates governing taxation of repatriated foreign income. As an alternative we tried specifications in which new treaty effects were measured as the withholding (pre- and post-treaty) rates for (1) dividend income for the parent firm, (2) dividend income for non-parent entities, (3) interest income, and (4) royalty income. To be clear, these variables were an interaction between whether the country partner was a new treaty country during our sample and the withholding rates. For all other countries in our sample, withholding rates would likely not change and their effects would be subsumed into the country fixed effects. A negative coefficient is expected as it would indicate that FDI activity goes up when withholding rates come down as a result of a treaty. Our estimates for both inbound and outbound FDI found no significant pattern in these withholding treaty variables, which is consistent with our results when using a binary treaty variable. These additional results are available on request. 3. Individual new partner country effects All our regressions to this point have assumed identical treaty effects across new treaty partner countries, which may mask important heterogeneity across these new partner countries. Table 5 explores this by estimating individual treaty-year dummies for each partner country using our logs specification with rich-country interactions from Table 4. Columns 1 and 3 of Table 5 provide our estimated effects by country for inbound and outbound FDI stock, respectively, after translating our coefficient estimates into millions of real U.S. dollars.24 To gauge the relative magnitude of the estimated treaty effects, columns 2 and 4 provide the amount of FDI stock in the first effective year of the treaty in millions of real U.S. dollars. For inbound FDI, about two-thirds of the treaty effects have a positive coefficient. Of these, five—Indonesia, Israel, Mexico, Spain, and Turkey—are significant. On the other hand, the estimated treaty effects for Jamaica, the Philippines, and Portugal are negative and significant, whereas the remaining four estimated treaty effects are statistically insignificant. Thus, for inbound FDI, we find significantly positive effects of treaties less than half of the time. For outbound FDI, the individual country treaty effects are even more mixed. Here, eight of the thirteen treaty estimates are negative, with three of those statistically significant. Only one treaty effect (Spain) shows any significant positive effect. Thus, our individual treaty estimates continue to find no overwhelming evidence for 24. The coefficient on our new treaty variable indicates how much the log of FDI stock changes with a new treaty in place. How much this log change means for the underlying change in the level of FDI stock is dependent on the starting value one chooses. We calculate the marginal effect by starting with the log of FDI stock in the year before the treaty becomes effective.

the effects of bilateral tax treaties on u.s. fdi activity 505

table 5. individual treaty effects on fdi stock based on coefficient estimates from fixed-effects logs specification with rich-country interactions Inbound FDI stock New treaty country

Estimated treaty effect

Stock in year of treaty

Outbound FDI stock Estimated treaty effect

Stock in year of treaty

China (1986)

6.98253

11.6

1311.797

211.5

Egypt (1981)

—

2.5

–135.944

1376.2

India (1990)

–10.2265

31.7

–208.233∗∗∗

392.8

Indonesia (1990)

584.5292∗∗∗

26.4

–976.09∗

3386.4

Israel (1994)

8173.057∗∗∗

1986.5

–245.726

1499.2

Jamaica (1981)

–0.04847∗

0.0

Mexico (1993)

1624.693∗∗∗

1276.4

Philippines (1982)

–57.4505∗∗∗

Portugal (1995) Spain (1990) Thailand (1997) Turkey (1997) Venezuela (1999)

—

546.1

3324.761

15617.7

98.0

–502.543∗∗∗

1586.1

–21.1166∗∗

23.0

–256.944

1411.0

624.9369∗∗∗

836.3

1829.36∗∗

8308.3

178.7193

224.2

–1529.14

4335.0

839.8532∗∗∗

59.0

–67.9615

1033.7

—

–62.7

1123.919

7422.2

Notes: All numbers are in millions of real U.S. dollars. ∗∗∗, ∗∗, and ∗ denoting statistical significance (two-tailed test) at the 1%, 5%, and 10% levels, respectively.

treaties increasing FDI. An important qualification of our individual country effects is that broader regime changes for new treaty partners may coincide with the date of a newly effective treaty. For example, the bilateral tax treaty with Mexico roughly coincides with the North American Free Trade Agreement. This is something we cannot address with our estimates.

conclusion This chapter presents the first estimates of the effects of bilateral treaties governing the taxation of FDI on bilateral FDI activity. Though our sample covers the most comprehensive set of countries and years to date, the data constraints allow

506 bruce a. blonigen and ronald b. davies

us to credibly estimate only the effects of U.S. bilateral tax treaties that were implemented in the 1980s and 1990s, which we call “new” treaties. Our initial estimates based on the an empirical model of MNE activity provided by Carr, Markusen, and Maskus (2001) suggest very large and statistically negative effects of new treaties on FDI, which would be consistent with the hypothesis that such treaties allow tax authorities to substantially reduce tax evasion practices that were a significant motivation for FDI. However, we find a number of undesirable statistical properties of these initial CMM estimates. We eliminate these statistical problems by modifying the CMM framework to include countrypartner fixed effects and a log-linear specification of the variables. Our estimates from this specification provide robust estimates of new treaty effects that are reasonable in magnitude. In general, we find that the average new treaty effect is not statistically different from zero with a very small point estimate. This is true for both FDI into the United States and U.S. outbound FDI. We find substantial heterogeneity in new treaty effects when we identify them on a countryby-country basis. Specifically, a number of treaties appear to have had significant positive effects on FDI activity (especially inbound FDI), while a handful of treaties are estimated to have had significant negative effects. It must be noted, however, that a number of the cases in which we find significant positive effects are arguably situations in which much larger regime changes coincided with the bilateral tax treaty. Thus, we find no systematic evidence that bilateral tax treaties affect FDI activity, despite statements by the OECD and other sources that such agreements are meant to increase the efficiency of world capital flows. Instead, our results suggest that either the provisions of a treaty have no effect, or the positive and negative aspects of treaty formation largely cancel one another.

the effects of bilateral tax treaties on u.s. fdi activity 507

data appendix bureau of economic analysis (bea) data on fdi activity The FDI activity data reported by the BEA is subject to two types of censorship. First, if revealing the bilateral FDI activity would reveal the information of a single firm, only this information is reported due to confidentiality requirements. These observations were dropped. Second, if the FDI activity lies between –$500,000 and $500,000, only this information is given. These observations were set to zero. This censorship then left us with an unbalanced panel which covered different time periods for different activity measures. Outbound FDI stocks were available as far back as 1966 for some partner countries, while inbound FDI stock data were available only back to 1980. All measures are for non-financial institutions and can be found at http://www.bea.doc.gov/bea/di1.htm.

tax treaties Information on tax data all came from the Worldwide Tax Treaties database at tax. com (2002). This database reports the calendar year a treaty was signed, the calendar year it became effective, and the calendar year the treaty entered into force.

other data Our GDP (both total and per capita), trade openness, and exchange rate measures are those from version 6.1 of the Penn-World Tables, which are available online at http://pwt.econ.upenn.edu. For a detail discussion of these measures, see Summers and Heston (1991). Total GDP is measured in millions of real base 1996 dollars. Our education variable is the mean years of education for adults over age twenty-five. This data comes from the Barro-Lee dataset which is available at http://www.worldbank.org/research/growth/ddbarle2.htm. Details on these data are given by Barro and Lee (1996). The exchange rates are measured as the home country currency price of one unit of the host country currency. Distance is measured as the distance between capital cities as reported by the Bali Online Corporation. This distance calculator can be found at http://www.indo.com. For our measure of investment costs, we use the composite score compiled by Business Environment Risk Intelligence, S.A. (BERI). This composite includes measures of political risk, financial risk, and other economic indicators, and ranges between 0 and 100, with higher numbers signifying more openness. To compare these estimates to previously used measures of investment barriers, we define Investment Barriers as 100 minus the BERI’s composite score.

508 bruce a. blonigen and ronald b. davies

Corporate tax rate data are average rates by country, as reported in Appendix Table 1a of Altshuler, Grubert, and Newlon (1998) for the period 1980–1992, and on the IRS website for years after 1992: http://www.irs.gov/taxstats/ article/0,,id=96282,00.html. The IRS website’s data are only available for evennumbered years, so following their lead we take the average of the two adjoining even-numbered years to estimate these corporate tax rate data for odd-numbered years. For 1999, the value in 1998 was used, provided it was available. The rates were calculated by dividing the income tax paid by the current earnings and profits before income taxes using only firms with positive current earnings and profits, similar to method of Altshuler, Grubert, and Newlon.

descriptive statistics of data Summary statistics for the data set are reported in Table A1 table a1. descriptive statistics Inbound

Outbound

Variable Obs.

Mean

Std. Dev.

Obs.

Mean

Std. Dev.

Levels Specification Stockij SUMGDPij GDPDIFSQij SKDIFFij SKDGDPDij F_COSTj T_COSTi HTSKDij T_COSTj DISTij NEW TREATYij

1470 1470 1470 1470 1470 1470 1470 1470 1470 1470 1470

5897.798 6958.45 4.22e+07 5.829699 36746.31 29.20743 80.7157 3329.2 34.94435 5033.416 .1659864

20662.22 1458.937 1.76e+07 2.63687 18079.52 2.578377 4.947221 2526.155 44.61486 2419.184 .372195

871 871 871 871 871 871 871 871 871 871 871

11100.92 7146.119 4.00e+07 4.973219 30162.27 47.16219 42.38257 1492.986 80.6973 5123.3 .2089552

22311.85 1523.354 1.70e+07 2.401347 15258.98 12.4759 43.25036 2208.56 4.909313 2256.359 .4067959

Logs Specification Stockij SUMGDPij GDPDIFSQij SKDIFFij F_COSTj T_COSTi T_COSTj NEW TREATYij

1470 1470 1470 1470 1470 1470 1470 1470

3.201628 8.826261 17.47075 1.616645 3.370186 4.389011 2.900666 .1659864

4.436201 .206615 .4276306 .6138984 .0940595 .0623969 2.124915 .372195

871 871 871 871 871 871 871 871

7.915625 8.852257 17.41155 1.458387 3.811664 3.292179 4.388814 .2089552

2.111452 .2094638 .4407623 .5884139 .3053918 1.848956 .0619008 .4067959

the effects of bilateral tax treaties on u.s. fdi activity 509

references Altshuler, Rosanne, Harry Grubert and T. Scott Newlon (1998). “Has U.S. investment become more sensitive to tax rates?,” NBER Working Paper No. 6383. Altshuler, Rosanne and T. Scott Newlon (1991). “The effects of U.S. tax policy on the income repatriation patterns of U.S. multinational corporations,” National Bureau of Economic Research Discussion Paper No. 571. Altshuler, Rosanne, T. Scott Newlon and William Randolph (1995). “Do repatriation taxes matter? Evidence from the tax returns of U.S. multinationals,” in Martin Feldstein, James Hines, Jr. and R. Glenn Hubbard, eds., Effects of Taxation on Multinational Corporations (Chicago: University of Chicago Press), pp. 253–272. Ault, Hugh J. and David F. Bradford (1990). “Taxing international income: an analysis of the U.S. system and its economic premises,” in Assaf Razin and Joel Slemrod, eds., Taxation in the Global Economy (Chicago: University of Chicago Press), pp. 11–46. Bali Online Corporation (1999). http://www.indo.com. Barro, Robert and J.W. Lee (1996). “International measures of schooling years and schooling quality,” American Economic Review: Papers and Proceedings, 86, pp. 218–223. Blonigen, Bruce A. and Ronald B. Davies (2000). “The effects of bilateral tax treaties on U.S. FDI activity,” National Bureau of Economic Research Working Paper No. 7929. Blonigen, Bruce A., Ronald B. Davies and Keith Head (2003). “Estimating the knowledge-capital model of the multinational enterprise: comment,” American Economic Review, 93, pp. 980–994. Bond, Eric and Larry Samuelson (1989). “Strategic behaviour and the rules for international taxation of capital,” Economic Journal, 99, pp. 1099–1111. Brainard, S. Lael (1997). “An empirical assessment of the proximity-concentration trade-off between multinational sales and trade,” American Economic Review, 87, pp. 520–544. Bureau of Economic Analysis [BEA] (1998). Detailed Annual Balance of Payments and Position Estimates (Washington, D.C.: Bureau of Economic Analysis International Data). Carr, David, James R. Markusen and Keith E. Maskus (2001). “Estimating the knowledge-capital model of the multinational enterprise,” American Economic Review, 91, pp. 693–708. Casson, Mark (1979). Alternatives to the Multinational Enterprise (London: Macmillan Press). Caves, Richard E. (1993). Multinational Enterprise and Economic Analysis, 2nd edition (New York: Cambridge Press). Chisik, Richard and Ronald B. Davies (2004). “Asymmetric FDI and tax treaty bargaining: theory and evidence,” Journal of Public Economics, 88, pp. 1119–1148. —— (forthcoming). “Gradualism in tax treaties with irreversible foreign direct investment,” International Economic Review. Davies, Ronald B. (2003). “The OECD model tax treaty: tax competition and two-way capital flows,” International Economic Review, 44, pp. 725–753. Dagan, Tsilly (2000). “The tax treaties myth,” New York University Journal of International Law and Politics, 32, pp. 939–996. Diamond, Walter and Dorothy Diamond (1998). International Tax Treaties of All Nations (Dobb’s Ferry: Ocean Publications).

510 bruce a. blonigen and ronald b. davies Doernberg, Richard (1997). International Taxation in a Nutshell, 3rd edition (St. Paul: West Publishing). Easson, Alex (1999). Taxation of Foreign Direct Investment: An Introduction (The Hague: Kluwer Law International). Eden, Lorraine (1998). Taxing Multinationals: Transfer Pricing and Corporate Income Taxation in North America (Toronto: University of Toronto Press). Graham, Edward and Paul Krugman (1995). Foreign Direct Investment in the United States, 3rd edition (Washington D.C.: Institute for International Economics). Gravelle, Pierre (1988). “Tax treaties: concepts, objectives and types,” International Bureau of Fiscal Documentation, Bulletin 522, pp. 522–526. Grubert, Harry and John Mutti (1991). “Taxes, tariffs and transfer pricing in multinational corporate decision making,” Review of Economics and Statistics, 73, pp. 285–293. Hallward-Dreimeier, Mary (2003). “Do bilateral investment treaties attract foreign direct investment? A bit . . . and they could bite,” World Bank Working Paper No. 3121. Hamada, Koichi (1966). “Strategic aspects of taxation on foreign investment income,” Quarterly Journal of Economics, 80, pp. 361–75. Hartman, David (1985). “Tax policy and foreign direct investment,” Journal of Public Economics, 26, pp. 107–121. Hines, James R., Jr. (1988). “Taxation and U.S. multinational investment,” in Lawrence Summers, ed., Tax Policy and the Economy, 2 (Boston: MIT Press), pp. 33–61. —— (1992). “Credit and deferral as international investment incentives,” National Bureau of Economic Research Working Paper No. 4191. Hines, James R., Jr. and Kristen L. Willard (1992). “Trick or treaty? Bargains and surprises in international tax agreements,” Mimeo. Janeba, Eckhard (1995). “Corporate income tax competition, double taxation treaties, and foreign direct investment,” Journal of Public Economics, 56, pp. 311–325. —— (1996). “Foreign direct investment under oligopoly: profit shifting or profit capturing?,” Journal of Public Economics, 60, pp. 423–445. Jones, Bob (1996). “International tax developments and double taxation agreements in Australia,” in Richard Vann, ed., Tax Treaties: Linkages Between OECD Member Countries and Dynamic Non-Member-Economies (Paris: OECD Committee on Fiscal Affairs), pp. 19–26. Markusen, James R. (2002). Multinational Firms and the Theory of International Trade (Boston: MIT Press). Markusen, James R. and Keith E. Maskus (2001). “Multinational firms: reconciling theory and evidence,” in Magnus Blomstrom and Linda S. Goldberg, eds., Topics in Empirical International Economics: A Festschrift in Honor of Robert E. Lipsey (Chicago, IL: University of Chicago Press for National Bureau of Economic Research), pp. 71–97. —— (2002). “Discriminating among alternative theories of the multinational enterprise,” Review of International Economics, 10, pp. 694–707. Mutti, John and Harry Grubert (1996). “The significance of international tax rules for sourcing income: the relationship between income taxes and trade taxes,” National Bureau of Economic Research Working Paper No. 5526. Organisation for Economic Co-operation and Development [OECD] (1989). Explanatory Report on the Convention on Mutual Administrative Assistance in Tax Matters (Paris: OECD Committee on Fiscal Affairs).

the effects of bilateral tax treaties on u.s. fdi activity 511

—— (1994). Tax Information Exchange Between OECD Member Countries: A Survey of Current Practices (Paris: OECD Committee on Fiscal Affairs). —— (1997). Model Tax Convention on Income and on Capital (Paris: OECD Committee on Fiscal Affairs). Radaelli, Claudio M. (1997). The Politics of Corporate Taxation in the European Union (London: Routledge Research in European Public Policy). Sasseville, Jacques (1996). “Current issues in international tax policy,” in Richard Vann, ed., Tax Treaties: Linkages Between OECD Member Countries and Dynamic Non-Member-Economies (Paris: OECD Committee on Fiscal Affairs), pp. 9–13. Shapiro, Samuel S. and Martin B. Wilk (1965). “An analysis of variance test for normality (complete samples),” Biometrika, 52, pp. 591–611. Sinn, Hans-Werner (1993). “Taxation and the birth of foreign subsidiaries,” in Horst Heber and Ngo V. Long, eds., Trade, Welfare, and Economic Policies: Essays in Honor of Murray C. Kemp (Ann Arbor: University of Michigan Press), pp. 325–352. Ramsey, James B. (1969). “Tests for specification errors in classical linear least squares regression analysis,” Journal of the Royal Statistical Society, Series B 31, pp. 350–371. Ramaswami, V. K. (1968). “International factor movement and the national advantage,” Economica, 35, pp. 309–310. Summers, Robert and Alan Heston (1991). “The Penn-world table (Mark 5): an expanded set of international comparisons, 1950–1988,” Quarterly Journal of Economics, 106, pp. 327–368. White, Fredrick (1991). “The United States perspective,” Tax Treaties and Local Taxes, International Fiscal Association, 16, pp. 15–24. United Nations (1998). Bilateral Investment Treaties in the Mid-1990s (New York: United Nations Conference on Trade and Development).

This page intentionally left blank

19. the impact of endogenous tax treaties on foreign direct investment: theory and empirical evidence∗ peter egger, mario larch, michael pfaffermayr, and hannes winner introduction There is a large body of theoretical and empirical literature viewing taxation as an important determinant of a country’s locational attractiveness for investors (see Hines 1999; Gresik 2001; de Mooij and Ederveen 2003, for comprehensive surveys). One of the most visible obstacles to cross-border investment is the double taxation of foreign-earned income. Double taxation arises if the same tax base (e.g., income or wealth) of a specific taxpayer (i.e., a person or enterprise) is taxed in two or more jurisdictions. Bilateral tax treaties, representing mutual agreements on the definition of tax liabilities, the assignment of taxing rights or the determination of withholding tax rates, are the most important way to circumvent double taxation of enterprises. Worldwide, the importance of tax treaties has increased tremendously in the last decades. The number of treaties in force has risen from 100 in the 1960s to more than 2,500 today (see Easson 2000; Arnold, Sasseville, and Zolt 2002). Most of them are based on model conventions established by international organizations such as the Organization for Economic Co-operation and Development (OECD) or the United Nations, where the former accounts mainly for the interests of industrialized countries and the latter for those of less developed economies. Apart from double taxation, tax treaties are usually motivated by additional objectives, depending on the specific economic and legal situation of the treaty partners. The preamble of the OECD Model Tax Convention on income and capital (henceforth OECD Model) emphasizes two major objectives: the alleviation of double taxation and the restriction of tax avoidance. The treaty partners are left

∗

This chapter was reprinted with permission from Canadian Journal of Economics. The chapter was originally published as “The impact of endogenous tax treaties on foreign direct investment: theory and empirical evidence”, Canadian Journal of Economics 39(3), 2006, 901–931. The authors are grateful to Ron Davies and the seminar participants at the 2004 IIPF Congress in Milan for helpful comments and suggestions. Financial support from the Austrian Fonds zur Förderung der wissenschaftlichen Forschung (FWF, grant no. P17028-G08) is gratefully acknowledged.

514 peter egger, mario larch, michael pfaffermayr, and hannes winner

free to agree on one or both objectives.1 The U.S.–Canada treaty, for instance, focuses on “the avoidance of double taxation and the prevention of fiscal evasion with respect to taxes on income and on capital . . .” The coexistence of different aims highlights a fundamental conflict in tax treaty formation. As far as double taxation is concerned, tax treaties may on the one hand promote international production and investment activities. For instance, the assignment of taxing rights via tax exemption (Art. 23A OECD Model) or tax credits (Art. 23B OECD Model), or agreement on maximum withholding tax rates (e.g., for dividends paid from foreign subsidiaries), lower the overall tax burden of multinational enterprises. On the other hand, restricting tax evasion and tax avoidance, primarily through transfer pricing rules (Art. 9 OECD Model), anti-treaty shopping rules, or the exchange of information (Art. 26 OECD Model), may induce negative effects on international investment. The conventional view of tax lawyers and administratives is that tax treaties are desirable to attract foreign investment. In contrast, Dagan (2000) argues that tax treaties are primarily intended to redistribute tax revenue (from poor to rich countries), or to lower administrative costs (e.g., via the definition of tax terms between contracting countries). Dagan concludes that double taxation should be more effectively eliminated by unilateral arrangements, which are less expensive and more flexible than tax treaties. In fact, some countries (e.g., Chile in the early 1990s) have not signed tax treaties but have unilaterally offered multinational enterprises a similar environment with respect to tax rules. However, there are several aspects of tax treaties that cannot be replicated by a country unilaterally. Most importantly, tax treaties reassure international investors ex-ante on how they are taxed with their overseas profits and this, in turn, may attract FDI. Although international issues of taxation have been of growing concern in economics, tax treaties have attracted less attention so far. The majority of available studies examines theoretically the role of different tax assignment rules for international investment flows. There are three methods of international tax assignment: Under the credit method, foreign-earned profits are taxed in both the parent and the host country, but the taxes paid abroad are deductible from the domestic tax liability.2 Under the exemption method, in contrast, domestic and foreign profits are taxed separately within their jurisdiction.3 A third method,

1. The aim of preventing tax avoidance was introduced recently in the 2003 revision of the OECD Model. In the 1963 and 1977 versions of the OECD Model, tax evasion was limited to fraud or criminal tax evasion (see Arnold 2004). 2. If the foreign tax burden is higher than the domestic one, the tax rebate is limited to the tax level of the home country. The limitation may be applied either for a single country (per-country-limitation) or for all countries (overall-limitation). 3. Consequently, this reduces the domestic tax base by the amount of foreign profits, which, in turn, lowers the overall tax burden in the presence of a progressive income tax system. In practice, this ‘splitting advantage’ is reduced via an exemption with progression.

the impact of endogenous tax treaties on foreign direct investment 515

though not proposed in the OECD Model, is the deduction method, which allows a deduction of foreign taxes from the domestic tax base. Bond and Samuelson (1989), relying on a model with a capital-exporting and a capital-importing country and discriminatory taxation (i.e., different tax rates for domestic and foreignearned income), demonstrate that tax rates will increase in equilibrium if the treaty partners agree on a credit system instead of the deduction method. This, in turn, induces a decline in investment flows to zero at the extreme (see Hamada 1966) for an earlier contribution). Davies and Gresik (2003) show that this result no longer holds under the assumption that subsidiaries are able to finance their operations in the host country. Considering non-discriminatory taxation, Janeba (1995) finds that the home country sets its tax rate equal to zero, irrespective of the assignment of taxing powers. Davies (2003a), focusing on two-way capital flows, shows that there are always positive capital flows in equilibrium, regardless of which tax relief method is employed. In contrast to Janeba (1995), he finds that the home and host country set positive tax rates in equilibrium. In sum, the theoretical hypotheses with respect to the impact of tax treaties on international investment flows are ambiguous, so it remains an empirical question whether tax treaties promote or hinder foreign direct investment (FDI). Empirical studies on the relationship between tax treaties and FDI are still scarce, and the available ones adopt methods that rely on the assumption that tax treaties are exogenous. Depending on whether a specification of FDI in logs or levels is chosen, previous studies find negative or no effects on FDI (Davies 2003b; Blonigen and Davies 2004a, 2004b; see Davies [2004] for an excellent survey). In this regard, Blonigen and Davies (2004a) indicate that empirical models of FDI should be formulated in logs rather than levels.4 However, based on such a specification recent research does not identify a significant impact of exogenous tax treaties on FDI (see Blonigen and Davies 2004a). For the average country-pair, the question then arises, whether implementing a treaty can be justified from an economic point of view. This chapter sheds further light on the obviously relevant question of how tax treaties affect outward FDI, emphasizing two important issues that are not raised in previous research. First, it is reasonable to assume that new treaties are implemented only, if both treaty partners gain in welfare. Hence, tax treaties are endogenous from a general equilibrium perspective. Accordingly, it is a major concern under which conditions positive, treaty-induced welfare effects arise. Put differently, the determinants of tax treaties are at issue. Second, given that a treaty is implemented (i.e., given its positive welfare effects), the associated impact on FDI is of interest. Focussing on the double taxation aspect of tax treaties, we set up a two-country general equilibrium model of trade and FDI, which is 4. Mutti and Grubert (2004) significantly reject level specifications versus log specifications based on a RESET test motivated by Davidson and MacKinnon (1993) in the context of taxation and FDI.

516 peter egger, mario larch, michael pfaffermayr, and hannes winner

closely related to the knowledge-capital model of the multinational enterprise (Carr, Markusen, and Maskus 2001; Markusen 2002; Markusen and Maskus 2002). The model is parameterized and the hypotheses on the impact of tax treaties on welfare and outward FDI (rather than foreign affiliate sales) are derived from numerical simulations. Abstracting from the potential administrative costreducing effect of tax treaties and tax avoidance issues,5 we find that treaties often have positive welfare effects, but the associated effect on FDI can even be negative without tax evading FDI. The simulation results for welfare motivate an empirical model to explain the probability of implementing new tax treaties. Further, given positive welfare effects, the results indicate under which circumstances one should expect negative FDI effects from implementing a treaty. In econometric terms, we estimate the average treatment on the treated, that is, the effect of a tax treaty on FDI between countries that have actually implemented a treaty. We construct a bilateral panel data set for tax treaties over the period from 1985 to 2000. Applying different matching estimators and focusing on differences-in-differences, we estimate a negative effect of newly implemented tax treaties of about –34% on changes in bilateral OECD outward FDI. The remainder of the chapter is organized as follows. Section A outlines the general equilibrium model. The theoretical hypotheses are summarized in Section B. Section C presents the database, the econometric methods, and the empirical results. In Section D, we discuss our results in the light of previous research, and provide an extension regarding the time pattern of accumulation of the treaty-induced effect on FDI. The last section concludes with a summary of the main findings. A. A Numerically Solvable General Equilibrium Model of FDI and International Taxation 1. Households There are two economies, referred to as countries 1 and 2 and indexed as {i, j} = {1, 2}. Both countries produce two goods, Z and X. Z is a homogeneous good produced at constant returns to scale by a competitive industry. X-goods are differentiated in the usual Dixit and Stiglitz (1977) fashion. We consider the following firm types: national firms (NEs) sell on the local market and export to the other country, where the number of national firms of country i is denoted by ni; horizontal multinational enterprises (MNEs) are running production plants in both countries, where hi denotes the number of horizontal multinationals headquartered in i; vertical MNEs are able to unbundle the headquarters and the production plant, where vi is the number of vertical multinationals with headquarters in i and production plants only in j. In contrast to horizontal MNEs,

5. The related FDI-inhibiting effects of tax treaties are well understood (see Davies 2003b).

the impact of endogenous tax treaties on foreign direct investment 517

vertical ones engage in goods trade. are the exports of country i–based NEs to country j. Similar definitions apply for the other quantities, where the first subscript always denotes the home country of the firm whereas the second one refers to the country of consumption. Xic denotes the consumption of X in country i, being a constant elasticity of substitution (CES) aggregate of the individual varieties. Consumer preferences are assumed to be a nest of the homogeneous Z-good and the differentiated X-good. The symmetry of varieties within a group of goods allows us to formulate utility of country i (Ui) as follows:

(1)

where α denotes the Cobb-Douglas expenditure share for differentiated products, and ε > 1 is the elasticity of substitution between varieties. Iceberg transport costs (τ) impede goods trade in both sectors. In quantity terms, one unit of consumption of an X-variety in country j requires a firm in i to send (1 + τ) units. Similarly, one consumption unit of a foreign Z-good in j requires a firm to ship, that is, to produce, (1 + τ) units. For convenience, quantities of X are defined as (both NE and vertical MNE) firm-specific production for the respective foreign market, whereas Zij and Zji are normalized to represent consumed rather than produced quantities. The consumer maximization problem can be solved in two steps. In the first , needs to be chosen so as to minimize the cost step, each variety , of attaining Xic, whatever the consumption of Xic is. In the second step, consumers allocate income between the Z-good and the composite X-good. Let be the price of an X variety in country i produced by a type-k firm headquartered in country j. The price for the homogeneous good, qi, is indexed once, since all (indigenous and foreign) homogeneous goods consumed at a single location i must face the same price qi. Further, si denotes the price aggregator, defined as the minimum cost of buying one unit of Xic at prices of an individual variety: s.t.

(2)

The first-stage budgeting problem leads to: (3) where Ei denotes total expenditures of consumers in country i. Identical price elasticities of demand and identical marginal costs (technologies) within

518 peter egger, mario larch, michael pfaffermayr, and hannes winner

a country ensure that the domestic price of a locally produced good is equal to the mill price for exports. Hence, mill prices of all goods produced in country i are equal in equilibrium (pi). Inserting these equilibrium conditions, the price aggregator si of differentiated goods consumed in country i can be written as .

(4)

The second-stage budgeting yields the division of expenditures between the two sectors: (5) (6)

where qi is the price of Z consumed in country i. The Z-sector is perfectly competitive, and we take qi as the numéraire. 2. Factor markets, production and income Let ωKi and ωLi denote the factor rewards for capital and labor, respectively. Assuming that Z-production only uses labor (L), variable unit costs (i.e., marginal costs) cZi satisfy (7)

where ωLi is the wage rate of unskilled workers in i and ⊥ indicates that at least one of the adjacent conditions has to hold with equality. This implies (8)

There is monopolistic competition in the X-sector, where each firm produces under a Leontief technology, using both capital and labor. The country-specific unit input coefficients of capital and labor are denoted by aKxi and aLxi, respectively. Additionally, national firms and horizontal and vertical MNEs require capital to set-up plants αKni, αKhi, αKvi, and they employ labor to produce firm-specific assets . In line with the literature, we assume that firmand blueprints specific fixed costs are highest for horizontal MNEs, slightly lower for vertical ones, and lowest for exporters: aKniωKi + aLniωLi < aKviωKi + aLviωLi < aKhiωKi + aLhiωLi and without loss of Specifically, we set generality. δ is the additional labor requirement to organize a multinational network, and 1 + γ are the fixed costs country i’s MNEs have to incur to set up a foreign plant in j. As mentioned above, horizontal MNEs also run domestic . production plants, which is reflected in Factor market clearing in country i for capital Ki and labor Li requires (9)

(10)

the impact of endogenous tax treaties on foreign direct investment 519

Variable unit costs of producing an X-variety in country i are given by . There is a fixed markup over variable costs, which is determined by the elasticity of substitution between varieties. Given that under CES-utility demand for all varieties is positive, we may write (11)

Free entry implies that firms earn zero profits, since operating profits are used to cover fixed costs. The corresponding zero-profit conditions determine the numbers of firms. and operating National firms in i have to bear fixed costs of profits are taxed at the rate ti (i.e., country i’s corporate tax rate), which is applied to profits earned in country i.6 The profit condition for NEs, therefore, implies (12)

Vertical and horizontal MNEs are able to finance fixed costs via operating profits of both the domestic and the foreign activities. We maintain that the foreign subsidiary fully (and immediately) repatriates its profits to the domestic parent (see Hartman [1985] and Sinn [1993] for a discussion of alternative assumptions). Without a tax treaty, operating profits from foreign affiliates denote the tax rate of country i applied are exposed to double taxation. Let to foreign-earned profits of domestic multinational firms (i.e., the taxation of the “permanent establishments” abroad indicated by superscript M), and withholding tax on repatriated profits levied by country j. The conditions for the number of MNEs are thus given by (13) (14)

From the perspective of an MNE with headquarters in country i, tj is the is equal to the overall corporate tax rate abroad, and, therefore, tax burden on foreign-earned profits under double taxation.7 Applying the exemption method in the tax treaty, becomes zero. In contrast, if both treaty

6. In terms of the OECD Model, national firms do not hold a “permanent establishment” (i.e., “a fixed place of business through which the business of an enterprise is wholly or partly carried on”; Art. 5 OECD Model) in foreign markets, and are therefore, not subject to taxation abroad. 7. Several OECD countries also apply various forms of unilateral methods to reduce the burden of double taxation, such as flat rates, deductions from the tax base or exemptions of foreign-earned profits. However, in most cases such practices do not completely eliminate double taxation.

520 peter egger, mario larch, michael pfaffermayr, and hannes winner

partners agree on the credit method, we have to replace by (assuming 8 that All factors are owned by the households, so that consumer income (i.e., GNP) in country i is given by (15) The equivalence of total factor income (Ei, Ej) and demand in each economy implicitly balances international payments. 3. The public sector The only source of income for the public sector are taxes on operating profits of firms. Tax revenues for country i can therefore be summarized as:

(16) We consider three alternatives of public expenditure. First, governments pay lump-sum transfers to consumers. Second, governments use tax revenues to finance public infrastructure. By assumption, the production of public infrastructure only uses labor and lowers the capital needed to set up a plant (for this approach see Egger and Falkinger 2003; Kellenberg 2003). Let PIi be the amount of public infrastructure provided by government in country i. The capital needed to set up a national firm in i is given by (17) where β is a scaling parameter. Similarly, the amount of capital needed to set up a plant for MNEs is lowered by the public infrastructure, therefore + γ and A third possibility is to finance a public good PGi, which directly enters the utility function of consumers (see, for example, Huizinga 1991; Lockwood 2003).

8. If tiM < tjM, a firm would accumulate an “excess foreign tax credit.” In this case, the credit is typically limited to tiM yielding tiM′ = 0. Consequently, the overall tax burden under the credit method is equal to that of the exemption method (see Hines 1999 for further details) The credit method is applied, for instance, by the United States, the United Kingdom, Japan, Greece, Ireland and Spain, while the exemption method is used by Germany, France and Italy. We do not refer to the deduction method, since it is not proposed in the OECD Model and it is typically applied unilaterally (see Devereux and Hubbard 2003 for a welfare analysis).

the impact of endogenous tax treaties on foreign direct investment 521

Again, it is assumed that the production of the public good only needs labor. The modified utility function in this case is given by:

(18)

where β is the scaling parameter introduced above. 4. Model parametrization For our simulations, we use the following parameter values. World factor endowments are set at L = 100 and K = 50. To study the effect of changes in the absolute bilateral country size, we triple both countries’ factor endowments. For the additional labor requirement to organize a multinational network, we assume δ = 0:01. Foreign-owned plants exceed the domestic capital requirement by γ = 0:1. The fixed input coefficients in the X-sector production are assumed at Lxi = 0:6 and Kxi = 0:4. We consider a value for the elasticity of substitution of ε = 4, which is similar to that one usually used in the literature (Markusen 2002). According to the United Nations World Trade Database, the share of manufacturing trade in the 1990s amounts to 70%–80% of total trade. Therefore, we assume an expenditure share for manufactures of α = 0:8. Motivated by the findings in Baier and Bergstrand (2001), the iceberg trade cost parameter is set at τ = 0:2. The scaling parameter used to model the importance of public expenditures for the capital usage in plant set-up is set at β = 0:4. Regarding tax rates, we assume , and tjω = tiω = 0.05 in the case of equal taxes. In the alternative scenario with unequal tax rates we .9 assume ti = tiM = 0.25 and B. Hypotheses To analyze the impact of tax treaties on FDI, two issues are of special interest: (i) welfare effects of tax treaties—under which conditions are countries inclined toward implementing a bilateral tax treaty?; and (ii) FDI effects of tax treaties–does FDI rise or fall after implementing a tax treaty, given that the treaty increases welfare? TAX TREATIES AND WELFARE : A bilateral tax treaty is implemented only if both countries gain in welfare.10 Figure 1 plots the signs of welfare effects of tax

9. The choice of these rates is motivated by the fact that withholding tax rates on cross-border payments of dividends are of similar size in most bilateral tax treaties. 10. Here, we rule out the possibility that the welfare gain of one party is used to compensate the welfare loss of the other. Further, it is reasonable to assume that marginal

522 peter egger, mario larch, michael pfaffermayr, and hannes winner figure 1. welfare change after tax treaty implementation (tax revenue is spent to finance public infrastructure) Capital in country 1 (% of world endowment)

100 90 80 70 60 50 40 30 20 10 10

20

30 40 50 60 70 80 Labor in country 1 (% of world endowment)

90

100

Positive welfare effect in both countries Non-Positive welfare effect in at least one country

treaties in the capital to labor endowment box of country i, in the case that governments spend their tax income on public infrastructure and countries levy identical tax rates.11 The welfare effect of a tax treaty is derived as the difference in the model outcome of two scenarios—one in which a treaty is effective, and a second one in which no treaty is implemented. The white areas in the figure indicate positive welfare effects of tax treaties. We see that positive welfare effects are more likely if relative factor endowments are dissimilar, and the difference in size (i.e., absolute factor endowment differences) is not too pronounced. The boundaries of the black area in the region around the center of the factor box are shaped like those of an ellipsoid. Hence, countries of more dissimilar size are likely to implement a treaty at smaller relative factor endowment differences, as

administrative costs of treaty formation rule out the implementation of a treaty at zero welfare gains. 11. Note that we focus on the endogeneity of bilateral treaties, leaving the unilateral setting of tax rates exogenous. Bond and Samuelson (1989), Janeba (1995), and Davies (2003) investigate the strategic tax rate setting under different assignment rules (i.e., credit, exemption, and deduction). The case of endogenous tax treaties under exogenous tax rates seems justified in the medium run. For instance, in the dataset used in the empirical analysis below, the number of newly implemented tax treaties is five times higher than that one of changes in the statutory corporate tax rates.

the impact of endogenous tax treaties on foreign direct investment 523

compared to equally sized economies.12 The welfare gains in both countries can be explained as follows: A tax treaty increases the number of firms in the capitalabundant country (specifically the number of horizontal MNEs), while the decrease in tax revenues and the induced reduction in public infrastructure provision does not produce countervailing effects of sufficient size due to the low capital price there. The capital-scarce country gains because it attracts more foreign plants. If the differences in country size become too large, there are no gains in welfare from implementing a treaty, irrespective of whether relative factor endowments are very different or not. Here, we have to consider two cases. If the differences in relative factor endowments are not big (i.e., the factor endowment points lie within the ellipsoid in center of Figure 1), the negative welfare effects of the treaty are driven by the reduction in tax revenue and the corresponding decrease in infrastructure. Dissimilarity in size and relative factor endowments (i.e., the black humps in the north-west and south-east of the factor box in Figure 1) implies negative welfare effects, since vertical MNEs in the capital-abundant country come into existence. This leads to a shift of production of the X-goods to the capital-scarce country. In sum, this suggests a nonlinear relationship between the relative factor endowment/country size differences and the likelihood of implementing a treaty. We infer the robustness of these results with respect to the alternative modes of government expenditures mentioned above (i.e., tax revenues are used to finance a public good, or they are distributed as a lump-sum transfer to consumers). Qualitatively, our theoretical findings on the relationship between welfare gains and factor endowment/country size differences remain unaffected. To study the impact of bilateral country size on the treaty-induced welfare effects, we triple the world factor endowments ceteris paribus. The result is illustrated in Figure 2. Now, there are islands with positive welfare effects along the main diagonal. Moreover, the black ellipsoid associated with non-positive welfare effects is narrower than in Figure 1. Hence, countries with similar relative factor endowments are more likely to implement a tax treaty, if both of them are larger. The reason for this lies in the rising dominance of horizontal MNEs in similarly endowed economies. If horizontal MNEs are prevalent in both countries, they will always gain from bilateral tax treaties. Notably, a change in world factor endowments does not affect the welfare effects of tax treaties under the alternative considered modes of public spending. TAX TREATIES AND FDI : As mentioned in the introduction, the reduction of a multinational’s tax burden through the abolishment of double taxation should

12. Graphically, imagine a relative factor price line with a negative slope. Two points on this line are relevant: the consumption point on the diagonal of the factor box, and a second point located at the border between the white and the black area. The segment of the factor price line between these two points is smaller for countries of dissimilar size than for countries with identical size.

524 peter egger, mario larch, michael pfaffermayr, and hannes winner figure 2. welfare change after tax treaty implementation, tripled world endowments (tax revenue is spent to finance public infrastructure) Capital in country 1 (% of world endowment)

100 90 80 70 60 50 40 30 20 10 10

20

50 60 70 80 30 40 Labor in country 1 (% of world endowment)

90

100

Positive welfare effect in both countries Non-Positive welfare effect in at least one country

promote FDI. In contrast, the treaty-induced restrictions on tax avoidance should lower FDI. In the theoretical model, we focus on the first issue. Nevertheless, as can be seen in Figure 3, even in this case a tax treaty may reduce FDI (see the black areas in the figure). The reason for a negative impact on outward FDI is that a treaty can lead to reduced tax revenues in the parent country. If governments use tax revenues to finance public infrastructure (thereby reducing the plant setup costs), a tax treaty induces higher fixed costs for MNEs. This effect may outweigh the direct FDI-enhancing effect of the reduction in tax burden due to the abolishment of double taxation. As a consequence, the domestic capital requirement for firm setup rises and the domestic price of capital relative to labor increases. The higher fixed costs for domestic MNEs reduce their competitiveness at the international goods market. In turn, they are partly replaced by foreign national firms. Even if governments use tax revenues to finance a public good, negative FDI effects can arise, but they are less likely under this alternative scenario. However, a negative effect on outward FDI does not occur under lump-sum transfers to consumers. Figure 4 combines the information in Figures 1 and 3, and illustrates under which factor endowment configurations a treaty increases (reduces) a country’s outward FDI, given that a treaty will be implemented (i.e., increases welfare in both countries). In all non-black areas in the figure, countries are willing to

figure 3. change in country 1’s outward fdi after tax treaty implementation (tax revenue is spent to finance public infrastructure) 100

Capital (% of RoW endowment)

90 80 70 60 50 40 30 20 10 10

20

30

FDI with tax treaty larger

50 60 70 40 Labor (% of RoW endowment) No Change

80

90

100

FDI without tax treaty larger

figure 4. change in country 1’s outward fdi after tax treaty implementation conditional on positive welfare effects (tax revenue is spent to finance public infrastructure) 100

Capital (% of RoW endowment)

90 80 70 60 50 40 30 20 10 10

20

30

40 50 60 70 Labor (% of RoW endowment)

Welfare and FDI Positive Welfare Positive and FDI negative

80

90

100

Welfare Positive and no change in FDI Others

526 peter egger, mario larch, michael pfaffermayr, and hannes winner

implement a treaty. In the white area, country 1’s outward FDI rises, while in the dark gray area country 1 does not conduct outward FDI so a treaty does not affect its welfare. In the light gray area to the north of the diagonal, there is a positive effect on welfare but a negative one on country 1’s outward FDI. Changing the model parametrization we identify β and ε as the most important determinants of the light gray area. For instance, lowering β (i.e., lowering the cost-reducing effect of public infrastructure) leads to an increase of the light gray area. In this case, the aforementioned revenue effect would be even more pronounced. Similarly, increasing ε (i.e., intensifying competition) also induces an increase of the light gray area. In sum, even in a setting with a focus on the double taxation effects of treaties, there is no clear-cut theoretical prediction on the FDI effects of tax treaties. ASYMMETRIC TAXATION : So far, we assumed identical tax rates in the two economies. With asymmetric tax rates, the results concerning the welfare and FDI effects of tax treaties do not differ much (see the corresponding figures in the supplement, available upon request from the authors). This holds true irrespective of whether the credit or the exemption method is applied. Accordingly, we conclude that the derived hypotheses are robust enough to warrant empirical implementation. C. Data and Difference-In-Difference Estimation of the Endogenous Tax Treaty Effect on Bilateral Outward FDI We estimate the effect of tax treaties on bilateral outward FDI of the OECD economies in the period from 1985 to 2000. Bilateral FDI stock data at an annual basis are from the OECD (Bilateral Investment Statistics Yearbook 2003). Detailed information on the implementation dates of tax treaties is available from the International Bureau of Fiscal Documentation (IBFD; Tax Treaties Database 2002, Release 3). Tables A1 and A2 in the Appendix provide details for those newly implemented tax treaties over the covered period that are part of the sample used in the empirical analysis. In order to isolate the effect of tax treaties on bilateral FDI, we focus on the marginal effect of implementing a new treaty on the change in FDI. Our theoretical model suggests tax treaty implementation as an endogenous event. Hence, there is self selection. To take account of this endogeneity, the proper specification of the decision to implement a new tax treaty (i.e., selection into treatment) is essential.13 Further, a difference-in-difference approach is recommended to guard against time-invariant, endogenous, unobserved effects. Such time-invariant sources of bias arise, for example, if country

13. Wooldridge (2002, 637) points out that endogeneity is an issue in panel data, if the treatment (in our case, implementing tax treaties) is correlated with time-varying unobservables that affect the response. The same argument applies for difference-in-difference analysis. However, this correlation takes place in a theoretical general equilibrium model such as the one outlined above.

the impact of endogenous tax treaties on foreign direct investment 527

table a1. number of observations of treated and untreated Implementation year 1985 1986 1987 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Total

Treated

Untreated

Total

3 1 3 3 4 3 5 4 3 10 6 6 7 6 3 67

20 31 26 34 35 40 46 56 62 57 68 73 70 60 41 719

23 32 29 37 39 43 51 60 65 67 74 79 77 66 44 786

pairs with newly implemented tax treaties and pairs without new treaties are geographically mismatched. SELECTION INTO TREATY FORMATION : We compute differences between the average log FDI in the two-year period after a new treaty was implemented and the average log FDI in the two-year period before it was implemented. The same difference is calculated for the group of untreated country pairs. Finally, the formation of an appropriate control group is important. For each implementation year, we construct a biannual pre-treatment and a biannual post-treatment period (the latter includes the implementation year). Only those country pairs that did not sign a treaty until the end of the post-treatment period are valid potential control units. We select the country pairs in the control group according to their similarity in the propensity score of the selection equation (a probit model). The effects of new tax treaties on FDI are estimated by several matching estimators (see below). We focus on the tax treaty effect on outward FDI for those country pairs that have actually implemented a new treaty. Hence, these estimates reflect the average treatment effect on the treated. As indicated in the summary of the theoretical hypotheses, the welfare effects from implementing a bilateral tax treaty at the beginning of period are determined by the following measures of average absolute and relative bilateral factor endowments in period t: total bilateral country size (approximated , the similarity in bilateral country by the log bilateral

528 peter egger, mario larch, michael pfaffermayr, and hannes winner table a2. newly implemented treaties between 1985–2000 (only the 67 treaties covered in the empirical analysis are reported) OECD economy

Partner economy

Australia Austria Austria Austria Canada Canada Canada Canada Canada Denmark Finland France France France France Germany Germany Germany Germany Iceland Iceland Iceland Italy Japan Korea, Rep. Korea. Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Korea, Rep. Netherlands

Indonesia Australia South Africa Slovenia Sweden Brazil Inda Hungary South Africa Greece Mexico Turkey Mexico Venezuela Panama Philippines Turkey Mexico Venezuela Canada Netherlands Poland Bulgaria. South Africa Philippines Indonesia Hungary Brazil Egypt Ireland Italy China Czech Republic Spain Romania Mexico Russia South Africa Portugal Mexico

Implementation year 1993 1989 1997 1999 1985 1988 1987 1995 1997 1993 1999 1990 1993 1994 1996 1985 1990 1994 1997 1998 1999 2000 1992 1998 1987 1990 1991 1992 1992 1992 1993 1995 1995 1995 1995 1996 1996 1997 1998 1995

the impact of endogenous tax treaties on foreign direct investment 529

OECD economy

Partner economy

Netherlands Netherlands New Zealand Norway Norway Poland Poland Portugal Portugal Portugal Portugal Switzerland Switzerland United Kingdom United Kingdom USA USA USA USA USA USA USA USA USA USA USA USA

Venezuela Argentina Thailand Greece Australia Ukraine Hungary USA Bulgaria Poland Romania Mexico Thailand Turkey Mexico Canada China Mexico Indonesia Spain Inda Israel Portugal Thailand Turkey South Africa Venezuela

Implementation year 1998 1999 1999 1992 1995 1995 1996 1996 1997 1999 2000 1995 1997 1989 1994. 1985 1987 1989 1990 1991 1991 1995 1996 1998 1998 1998 2000

size (approximated by the log similarity index, , the difference in relative factor endowments (approxiwith mated by the absolute bilateral log difference in capital-labor ratios, , and interaction terms of the latter two variables . GDP data in constant U.S. dollars are available from the World and Bank (World Development Indicators 2004), and capital-labor ratios are computed in Baier, Dwyer, and Tamura (2002).14 We regress on the initial level of these variables, that is, they are predetermined.

14. We are grateful to Scott Baier for kindly providing the data.

530 peter egger, mario larch, michael pfaffermayr, and hannes winner

Denoting the parameter estimates of the selection model by α’s, we expect the following coefficient signs from the above model. αG > 0 indicates that the probability of implementing a new treaty rises with bilateral country size. The direct effects of S and |R| reflected in the signs of αS and α|R| should not be interpreted without reference to the interaction terms. Starting from the center of the factor box in Figure 1, a smaller relative factor endowment difference |R| is necessary as S gets smaller to obtain a positive welfare effect. Since S < 0 by definition, we expect αSx|R| < 0. However, as S gets extremely small toward the lower left and upper right edges of the factor box, an increase in |R| is no longer associated with a higher likelihood of implementing a tax treaty. Accordingly, we expect α(Sx|R|)2 < 0, because S2 > 0. Hence, including S x |R| and (S × |R|)2 in addition to the main effects S and |R| helps to capture this nonlinear relationship between country size and relative factor endowment differences on the one hand and implementing a treaty on the other hand. In addition to these determinants, we include the parent and the host government expenditure-to-GDP ratios (git, gjt) as two separate controls (World Bank, World Development Indicators 2004). These two variables indicate whether the parent and the host are high-tax or low-tax countries.15 If the objective of tax treaties is primarily to prevent double taxation, we would expect a negative sign for the coefficient of gjt. We report summary statistics for all variables in the Appendix (Table A3). Formally, we specify selection into treatment (i.e., the probability of new tax treaty implementation) by the following probit model

(19)

where α0 is a constant and εij,t+1 ∼ iid N (o,σε2). Due to the availability of data for FDI and the controls, the sample used in the econometric analysis spans the period from 1985 to 2000. After eliminating missing values, data for 67 newly implemented treaties can be used. Given the length of the period, this corresponds to about 4:5 treaties per annum (see Table A1 in the Appendix for more details on the time pattern of tax treaty implementation and the relative sizes of the groups of the treated and the untreated country pairs). Since (parent and host) countries tend to sign more than a single treaty over the observed time span and there is usually more than a single treaty signed in each year, we are able to include parent country (μi) and host country (νj) dummies as well as

15. Alternatively, we used corporate tax rates available from the Office of Tax Policy Research (World Tax Database). However, the results are insensitive with respect to this choice.

the impact of endogenous tax treaties on foreign direct investment 531

table a3. descriptive statistics Variables

Mean

Std. dev.

Minimum

Maximum

Endogenous variables (changes):a New tax treaties implemented Log change in bilateral outward FDI

0.09 0.47

0.28 0.80

0.00 −3.17

1.00 5.97

Exogenous variables in selection equation (Initial period levels]:b) Log bilateral GDP: Gj 27.75 1.33 23.42 Log parent-to-host similarity 1.21 −1.88 −6.20 in GDP: Sj 2.28 1.83 0.00 Absolute difference in log parent-to-host capital-labor ratio: |Rj| 5.57 Interaction term: Sij × |Rij| −4.58 −39.40 2 Interaction term: (Sij × |Rij|) 51.93 141.37 0.00 Parent government expenditure18.66 3.85 9.75 to-GDP ratio: g Host government expenditure15.63 5.69 3.14 to-GDP ratio: g

30.27 −0.69 11.90

0.00 1552 44 25.84 38.23

Notes: 786 observations. - a) Change between 2-year period after the treaty implementation and the 2-year period before the implementation. i.e., the treaty dummy is set at 1. if a new treaty was implemented and 0, else. The change in FDI is the log difference between the 2-year average of the period after and that one before the treaty was implemented.

time dummies (λt) to guard against the bias from omitted country-specific and time-specific effects. Table 1 summarizes the parameter estimates of four different probit models. The first two reported models (referred to as the parsimonious ones) impose the restrictions α(Sx|R|) = α(Sx|R|)2 = 0, whereas the third and the fourth model relax them. The second and the fourth model include country and time dummies. The country dummies especially improve the model fit considerably. This indicates that the propensity of signing a new treaty differs systematically between countries. Further, the restrictions on α(Sx|R|) and α(Sx|R|)2 to be zero are rejected at conventional levels of significance. Accordingly, we use the results from the fourth model in Table 1 in the subsequent analysis. The parameter estimates of this model lend support to our theoretical hypotheses. First, the likelihood of implementing a . Further, < 0 and treaty rises with bilateral country size, reflected by < 0 support our expectations regarding the functional relationship between relative and absolute bilateral factor endowment differences on the one hand and

532 peter egger, mario larch, michael pfaffermayr, and hannes winner table 1. selection equation (ratified double taxation treaties) Parsimonious specification Explanatory variablesa

Log bilateral GDP: Gj Log parent-to-host similarity in GDP: Sj Absolute difference in log parentto-host capital-labor ratio: |Rj| Interaction term : Sij × |Rij| Interaction term: (Sj × |Rj])2 Parent government expenditure-to-GDP ratio: g Host government expenditure-to-GDP ratio: g Number of country-pairs Pseudo R2 Likelihood ratio tests (P>x2}: Parent countries (18) Host countries (31) Time [14)

Full model

Excl. Group Incl. Group Excl. Group Incl. Group effects Est. effects Est. effects Est effects Est. Coeff. Coeff. Coeff. Coeff. −0.04 −0.76 0.02 0.34 0.03 0.96 − − − − −0.04 *** −3.42 −0.03 *** −2.61

3.66 ** 2.25 1.61 * 1.96 0.06 0.62 − − − − 0.09 0.68 −0.12 ** −2.25

−0.06 −0.97 −0.05 −0.39 0.02 0.20 −0.03 −0.29 0.00 −0.71 −0.04 ** −2.23 −0.02 * −1.90

4.14 ** 2.46 2.24 *** 2.56 −0.33 * −1.84 −0.47 *** −3.06 −0.02 *** −2.80 0.06 0.48 −0.11 ** −2.11

786 0.04

786 0.25

786 0.03

786 0.27

− − −

0.00 *** 0 02 ** 0.81

− − −

0.00 *** 0.01 ** 0.78

a

t-values in italics below coefficients. ***significant at 1%; **significant at 5%; *significant at 10%

implementing a tax treaty on the other. Finally, the host country government expenditure variable enters significantly. A negative impact suggests that OECD countries tend to implement bilateral tax treaties more likely with countries that keep government expenditures low. Since a similar finding is obtained if we use corporate tax rates instead, we may conclude that parent countries use tax treaties to reduce the relative attractiveness of low-tax countries. TAX TREATIES AND FDI: One branch of econometric literature on the estimation of endogenous treatment effects suggests propensity score matching techniques to infer the treatment effect on the treated (see Rosenbaum and Rubin 1983; Heckman, Ichimura, and Todd 1998; Vella and Verbeek 1999). We follow this

the impact of endogenous tax treaties on foreign direct investment 533

line of research and apply several different matching estimators to our differencein-difference specification. For this, we need to compare the growth in FDI of those countries that implemented a treaty with those that did not but were similar in all other respects (as captured by similar propensity scores). A properly weighted sample of untreated country pairs serves as a group of control units. We apply several different matching estimators. Nearest-neighbor matching estimators select one (one-to-one matching) or more (in our case, five) untreated control observations for each treated observation according to the propensity score, estimated in the above probit model. Each untreated observation potentially serves as a control for several treated ones. Under radius matching, all control units with a difference in propensity scores within a given radius are used. Kernel matching estimators make use of the whole sample of untreated observations, but they weight them according to the estimated kernel density weights. The obtained kernel estimates may be sensitive to the assumed density function and, more importantly, to the chosen bandwidth of the kernel. Local linear regression matching applies local linear weights rather than kernel weights. In general, there is a trade-off between exact matches at the danger of low precision (under nearest-neighbor, nearest-five-neighbors, small radius matching) and high precision at the danger of inexact matching (under wide radius matching or kernel matching). For all matching estimators, it is important that the balancing property holds, that is, that the covariates are balanced between the treated and the control group. The balancing property is violated if the values of a covariate differ significantly between the treated and the matched control group. This is likely the case with many controls. Although the balancing property holds for our continuous variables according to t-tests, it is violated for the country and time dummies. However, associated problems can easily be avoided by using these dummies not only in the selection equation but also as controls in the treatment regression after matching. If the balancing property holds, the average treatment effect on the treated can always be estimated by regressing the dependent variable (in our case, the change in FDI stocks) on the treatment dummy and a constant in a weighted least-squares regression. The weights are determined by the chosen matching approach (for instance, in case of one-to-one matching all treated and control observations exhibit weight one and all other observations weight zero; under kernel matching, the kernel weights are used). If the balancing property is violated, the problematic controls should simply be included in the weighted least-squares regression in addition to the treatment dummy (see Blundell and Costa Dias 2002). In our case, this means regressing the change in log FDI stocks on the treatment dummy and the country and time dummies. The coefficient of the tax treaty dummy is an estimate of the average treatment effect on the treated. Table 2 summarizes our findings for the various matching estimators based on the fourth probit model in Table 1. We report three sets of average treatment effects on the treated. The first one does not condition on the controls of the

Estd. Effect Descriptive comparison (tax treaties exogenous) Standard error One-to-one matching Standard error Five nearest neighbor matching Standard error Radius matching Standard error Kernel matching Epanechnikov kernel: bandwidth = 0.06 Standard error Epanechnikov kernel; bandwidth = 0.01 Standard error Epanechnikov kernel; bandwidth = 0.8 Standard error Gaussian kernel; bandwidth = 0.06 Standard error Uniform kernel: bandwidth = 0.06 Standard error Local linear regression matching Standard error

% change

Estd. Effect

% change

Estd. Effect

% change

0.002 0.12 −0.40** 0.17 −0.26* 0.15 −0.21 0.14

−0.62 12.34 −34.17 10.81 −23.57 11.14 −19.99 10.77

−0.18 0.12 −0.53** 0.20 −0.30 ** 0.13 −0.21 ** 0.10

−16.86 10.24 −42.44 11.63 −26.22 9.60 −19.14 8.08

−0.19 0.12 −0.41** 0.17 −0.30** 0.12 −0.21** 0.09

−18.11 10.12 −34.45 11.32 −26.58 8.86 −19.15 7.34

−0.23 * 0.14 −0.24 0.15 −0.12 0.13 −0.21 0.13 −0.24 * 0.14 −0.22 0.14

−21.35 10.75 −22.45 11.51 −12.15 11.06 −19.63 10.72 −21.71 10.73 −20.69 11.09

−0.20 ** 0.10 −0.22 ** 0.11 −0.26 *** 0.10 −0.21 ** 0.10 −0.21 ** 0.10 −0.18* 0.10

−18.75 8.05 −19.90 8.73 −23.24 7.35 −19.32 7.99 −19.60 7.91 −17.16 8.54

−0.21** 0.09 −0.22** 0.10 −0.24*** 0.09 −0.21** 0.09 −0.22** 0.09 −0.20** 0.09

−19.20 7.28 −20.28 7.83 −21.82 7.00 −19.26 7.29 −20.13 7.17 −18.23 7.55

Note: The tangible and intangible investment to sales ratios are logistically transformed; the selection probit equation includes firm size in terms of employment in 1993, the foreign ownership dummy, and the average exports to sales ratio 1997–1998. Tangible and intangible investment is measured in percent of sales and averaged over the periods 1999–2001. ** significant at 5%.

534 peter egger, mario larch, michael pfaffermayr, and hannes winner

table 2. the impact of tax treaty ratification on bilateral outward fdi (matching estimates) (average treatment effect on the treated) Excl. country & Incl. country & Incl. all controls of time dummies time dumnies the probit

the impact of endogenous tax treaties on foreign direct investment 535

probit, the second one includes the country and time dummies, and the third one includes all controls of the probit (dummies and continuous variables). On top of the matching results, the descriptive comparison reflects the parameter estimate of a regression of the log difference in FDI on the change in the tax treaty dummy (all standard errors in Table 2 are heteroskedasticity-robust). In the absence of endogenous selection, this provides a consistent differencein-difference estimate of both the treatment effect and the standard error. In contrast, the standard error of the tax treaty parameter estimated in a fixedeffects panel is likely downward biased (Bertrand, Duflo, and Mullainathan 2004). The covariates of the probit are jointly significant in the second-step weighted least-squares regression of the log change in FDI on the tax treaty dummy. Accordingly, we consider the third block of results in Table 2 as the preferred one. Note that the tax treaty coefficients cannot be interpreted at face value. Instead, one should take account of the underlying semi-logarithmic regression framework (see Kennedy 1981). We compute the percent changes in FDI associated with these estimates following van Garderen and Shah (2002). This obtains the corresponding percent change in FDI. Treating tax treaties as an exogenous variable, we do not find any significant impact on FDI. However, the one-to-one matching estimates indicate that newly implemented tax treaties reduce FDI significantly. This points to endogenous selection into treatment leading to an underestimation of the tax treaty effect on FDI in absolute value. As long as we condition on the country and time dummies in the second-step weighted least-squares regression of FDI on the tax treaty dummy, tax treaty implementation is found to reduce FDI significantly. This effect must be interpreted as an average treatment effect on the treated. It indicates the change in outward FDI for two countries that have actually implemented a tax treaty. The general result of a negative tax treaty effect is neither affected by the choice of the matching procedure nor by different assumptions for bandwidth and density functions in case of kernel matching. The effect is highest under oneto-one matching in absolute value. There, the control units are closest to the treated ones (“exact” matching). The corresponding tax treaty effect is more reliable than that one estimated from “inexact” but potentially more efficient procedures such as kernel matching or radius matching. As we enlarge the control group by considering not only the nearest neighbor but also other control units, the weighted estimated impact gets smaller in absolute value. The smallest estimate is obtained from the local linear regression procedure. In sum, this indicates that the loss in matching quality from using more control units is relatively big as compared to the associated gain in efficiency in our sample. D. Discussion and extension How do the obtained estimates relate to the results in previous research? As in Blonigen and Davies (2004a), a log specification in our sample leads to an insignificant tax treaty effect on outward FDI, if we assume that treaties are exogenous

536 peter egger, mario larch, michael pfaffermayr, and hannes winner

(see the first line of results in Table 2).16 In contrast, we rigorously account for the endogeneity of tax treaties by specifying an empirical selection model, which is guided by the hypotheses derived in a general equilibrium model of trade and multinationals. In our case, ignoring self-selection leads to insignificant and much lower point estimates in absolute value. Further, the data are typically unbalanced and the implementation of new treaties is unequally spaced over time (i.e., the spell lengths before and after new treaties are implemented differ across country pairs with new treaties). This complicates the interpretation of the treatment coefficient, since the effect may implicitly reflect a weighted shortrun and long-run impact. Only if the pre-treatment and treatment periods are of the same length for all treated and untreated country pairs in the sample (in our design, two years for each) is a difference-in-difference interpretation justified (see also Kyriazidou 1997). Our estimates of the treatment effects on the treated in Table 2 indicate that the expectation of a negative effect of new tax treaties is justified as concluded in the inferior level-specifications in Blonigen and Davies (2004a, 2004b). However, our estimates underpin the importance of the endogeneity of tax treaties for an unbiased inference. In our sample of countries, we consider the oneto-one matching result of an expected reduction in FDI growth by 34% in the third block of results in Table 2 as most reliable among the reported results. As mentioned above, we have so far required the pre-treatment and the treatment period to be of equal length (two years). If the treatment effects do not take place immediately due to adjustment costs, varying the window size of the treatment period may shed further light on the short-run versus long-run effects of implementing a new tax treaty on the treated. To assess the accumulation pattern of the tax treaty impact on outward FDI over time, we change the window size of

16. In a sample of OECD parent countries and a knowledge-capital specification using data over the period 1983–1992, Blonigen and Davies (2004b) estimate a significant negative impact of new tax treaties on levels of bilateral outward FDI stocks and flows. Their preferred specification is a model with fixed country-pair effects, and the identified tax treaty effect is extremely large in absolute values. In Table 5 of Blonigen and Davies (2004b), they report a new tax treaty parameter of –2597:6 in the FDI outbound stock specification in column 1. Given the average level of OECD outbound stocks of 3378:13 reported in their Table 2, this means that FDI stocks should be expected to shrink by almost 77%. In column 3 of Table 5, they run a fixed country-pair specification which includes the lagged FDI stock as a regressor. The reported new tax treaty parameter in this model is –2212:3 and the coefficient of the lagged dependent is 0:308, implying a long-run treaty effect of –3197:0, that is, a reduction in FDI stocks by 97%. These results are also consistent with the levels regressions in Blonigen and Davies (2004a), finding new treatyinduced reductions in U.S. FDI of 77% for outward stocks and of 88% for inward stocks. However, the effect on U.S. FDI in a preferable specification in logs is insignificant.

the impact of endogenous tax treaties on foreign direct investment 537

figure 5. effect of tax treaties on average bilateral fdi in periods of different length after ratification (point estimates and 95% confidence bound based on epanechnikov kernel with bandwidth 0.06) 0 % change in bilateral FDI

−10 −20 −30

point estimate

−40 −50 −60 −70 −80 Years 0-2

Years 0-3

Years 0-4

Years 0-5

the treatment period from two years to three, four and five years, respectively.17 If adjustment costs are important, the two-year window is closest to the short-run impact, while the others reflect the medium-run to long-run impact.18 In Figure 5, we display the percent-change of FDI associated with a new tax treaty for these different windows of the treatment period. To estimate the effects, we choose the one-to-one matching estimates. Accordingly, the outer left value of the black line reflects the corresponding estimate reported in Table 3. The point estimate is displayed in black and the upper and lower 95% confidence bounds appear in gray. The figure indicates that adjustment costs are important and the accumulated negative impact of bilateral tax treaties is higher in absolute value over the longer periods.

conclusion Two primary goals of bilateral tax treaties are the elimination of double taxation of cross-border activities and the prevention of tax avoidance and evasion. Accordingly, the net impact of implementing a tax treaty on foreign direct investment is not clear-cut. 17. We are not able to consider longer periods because of the associated substantial loss of observations and, in particular, of tax treaty–implementing country pairs. 18. However, with an adjustment cost parameter of the size as estimated in Blonigen and Davies (2004b), more than 90% of the long-run impact would be already accumulated after only two years.

538 peter egger, mario larch, michael pfaffermayr, and hannes winner

This paper analyzes the influence of bilateral tax treaties on bilateral stocks of outward FDI, where we focus on the double taxation issue but not on tax evasion. We set up a general equilibrium model of trade and multinational firms to study the welfare and FDI effects of tax treaties. This model motivates an empirical specification to explain the implementation of tax treaties. The selection into treatment (i.e., a positive treaty-induced change in welfare) is considered as an endogenous event. Based on this selection model, we estimate the tax treaty implementation effect on FDI using various difference-in-difference propensity score matching approaches. In our sample of OECD outward FDI over the period from 1985 to 2000, we identify a significant negative and reasonable treatment effect on the treated, which is robust with respect to the matching estimator choice. A negative tax treaty effect on FDI can be explained by our general equilibrium model, if tax revenues are spent for public infrastructure to reduce plant setup costs. However, it may also indicate that the tax avoidance aspect of bilateral tax treaties is present in our sample.

references Arnold, B. (2004). “Tax treaties and tax avoidance: the 2003 revisions to the commentary to the OECD model,” Bulletin for International Fiscal Documentation, 58, pp. 244–260. Arnold, B.J., J. Sasseville and E.M. Zolt (2002). “Summary of the proceedings of an invitational seminar on tax treaties in the 21st century,” Bulletin for International Fiscal Documentation, 56, pp. 233–245. Baier, S.L. and J.H. Bergstrand (2001). “The growth of world trade: tariffs, transport costs, and income similarity,” Journal of International Economics, 53, pp. 1–27. Baier, S., Dwyer, G.P. and R. Tamura (2002). “How important are capital and total factor productivity for economic growth?,” unpublished manuscript, Clemson University. Bertrand, M., E. Duflo and S. Mullainathan (2004). “How much should we trust differences-in-differences estimates?,” Quarterly Journal of Economics, 119, pp. 249–275. Blonigen, B. A. and R. B. Davies (2004a). “The effects of bilateral tax treaties on U.S. FDI activity,” International Tax and Public Finance, 11, pp. 601–622. Blonigen, B.A. and R.B. Davies (2004b). “Do bilateral tax treaties promote foreign direct investment?,” in J. Hartigan, ed., Handbook of International Trade: Economic and Legal Analysis of Laws and Institutions (Oxford: Blackwell Publishers), forthcoming. Blundell, R. and M. Costa Dias (2002). “Alternative approaches to evaluation in empirical microeconometrics,” Portuguese Economic Journal, 1, pp. 91–115. Bond, E.W. and L. Samuleson (1989). “Strategic behaviour and the rules for international taxation of capital,” The Economic Journal, 99, pp. 1099–1111. Carr, D.L., Markusen, J.R. and K.E. Maskus (2001). “Estimating the knowledge-capital model of the multinational enterprise,” American Economic Review, 91, pp. 693–708. Dagan, T. (2000). “The tax treaties myth,” New York University Journal of International Law and Politics, Summer, pp. 939–996.

the impact of endogenous tax treaties on foreign direct investment 539

Davidson, R. and J.G. MacKinnon (1993). Estimation and Inference in Econometrics (New York: Oxford University Press). Davies, R.B. (2003a). “The OECD model tax treaty: tax competition and two-way capital flows,” International Economic Review, 44, pp. 725–753. Davies, R.B. (2003b). “Tax treaties, renegotiations, and foreign direct investment,” Economic Analysis and Policy, 32, pp. 251–273. Davies, R.B. (2004). “Tax treaties and foreign direct investment: potential versus performance,” International Tax and Public Finance, 11, pp. 775–802. Davies, R.B. and T.A. Gresik (2003). “Tax competition and foreign capital,” International Tax and Public Finance, 10, pp. 127–145. de Mooij, R.A. and S. Ederveen (2003). “Taxation and foreign direct investment: a synthesis of empirical research,” International and Tax Public Finance, 10, pp. 673–693. Devereux, M.P. and R.G. Hubbard (2003). “Taxing multinationals,” International Tax and Public Finance, 10, pp. 469–487. Dixit, A.K. and J.E. Stiglitz (1977). “Monopolistic competition and optimum product diversity,” American Economic Review, 67, pp. 297–308. Easson, A. (2000). “Do we still need tax treaties?,” Bulletin for International Fiscal Documentation, 54, pp. 619–625. Egger, H. and J. Falkinger (2003). “The role of public infrastructure for firm location and international outsourcing,” unpublished manuscript, University of Zürich. Hamada, K. (1966). “Strategic aspects of taxation on foreign investment income,” Quarterly Journal of Economics, 80, pp. 361–375. Hartman, D. (1985). “Tax policy and foreign direct investment,” Journal of Public Economics, 26, pp. 107–121. Gresik, T.A. (2001). “The taxing task of taxing transnationals,” Journal of Economic Literature, 39, pp. 800–838. Heckman, J., H. Ichimura, and P. Todd (1998). “Matching as an econometric evaluation estimator,” Review of Economic Studies, 65, pp. 261–294. Hines, J.R. (1999). “Lessons from behavioral responses to international taxation,” National Tax Journal, 52, pp. 305–322 Huizinga, H. (1991). “National tax policies towards product-innovating multinational enterprises,” Journal of Public Economics, 44, pp. 1–14. Janeba, E. (1995). “Corporate income tax competition, double taxation treaties, and foreign direct investment,” Journal of Public Economics, 56, pp. 311–326. Kellenberg, D. (2003). “The provision of public inputs and foreign direct investment,” Center for Economic Analysis Discussion Paper No. 03-11 (Department of Economics, University of Colorado at Boulder). Kennedy, P.E. (1981). “Estimation with correctly interpreted dummy variables in semilogarithmic equations,” American Economic Review, 71, p. 801. Kyriazidou, E. (1997). “Estimation of panel data sample selection model,” Econometrica, 65, pp. 1335–1364. Lockwood, B. (2003). “Imperfect competition, the marginal cost of public funds and public goods supply,” Journal of Public Economics, 87, pp. 1719–1746. Markusen, J.R. (2002). Multinational Firms and the Theory of International Trade (MIT Press: Cambridge, MA). Markusen, J.R. and K.E. Maskus (2002). “Discriminating among alternative theories of the multinational enterprise,” Review of International Economics, 10, pp. 694–707.

540 peter egger, mario larch, michael pfaffermayr, and hannes winner Mutti, J. and H. Grubert (2004). “Empirical asymmetries in foreign direct investment,” Journal of International Economics, 62, pp. 337–358. Rosenbaum, P.R. and D.B. Rubin (1983), “The central role of the propensity score in observational studies for causal effects,” Biometrika, 70, pp. 41–55. Sinn, H.-W. (1993.) “Taxation and the birth of foreign subsidiaries,” in H. Heber and N. van Long, eds., Trade, Welfare, and Economic Policies: Essays in Honor of Murray C. Kemp (University of Michigan Press: Ann Arbor, MI), pp. 325–352. van Garderen, K.J. and C. Shah (2002). “Exact interpretation of dummy variables in semilogarithmic equations,” Econometrics Journal, 5, pp. 149–159. Vella, F. and M. Verbeek (1999). “Estimating and interpreting models with endogenous treatment effects,” Journal of Business and Economic Statistics, 17, pp. 473–478. Wooldridge, J.M. (2002). Econometric Analysis of Cross Section and Panel Data (MIT Press: Cambridge, MA).

20. host-country governance, tax treaties, and u.s. direct investment abroad∗ henry j. louie and donald j. rousslang introduction Recent studies have shown that poor governance has a chilling effect on investors. Mauro (1995) showed that corruption and political instability discourage investment in the local economy in general, and Wei (2000) showed that these traits discourage incoming foreign direct investment (FDI).1 Attempts to show that poor governance discourages investment from U.S. companies have not met with much success, however. Wheeler and Mody (1992) could find no evidence that the quality of host-country governance had any effect on investment spending by the foreign subsidiaries of U.S. companies and when Wei (2000) regressed U.S. FDI by itself against corruption and political stability, the coefficients were not statistically significant at the usually accepted levels. During the time periods examined in these studies, the United States might have been a special case with respect to the response of its FDI to corruption: From 1977, when it enacted the Foreign Corrupt Practices Act (FCPA), until early 1999 it was the only country with laws to prohibit its companies from bribing foreign government officials. Hines (1995) found that in the five-year period after its enactment, the Act slowed the growth of U.S. FDI in corrupt host countries relative to the growth of FDI from other countries. Hines’ findings do not necessarily imply that corruption discourages investment (the Act might merely have discouraged U.S. investors from using bribes to overcome non-deliberate bureaucratic delay),2 but they do imply that the FCPA put U.S.-owned companies at a competitive disadvantage against their foreign rivals in host countries

∗ This chapter was reprinted with permission from International Tax and Public Finance. The chapter was originally published as “Host-country governance, tax treaties, and American direct investment abroad,” 15 International Tax and Public Finance 256 (2008). The views expressed are of the authors and do not reflect the official views of the Treasury Department or the State of Hawaii. The authors are indebted to Laura Clauser, Harry Grubert, Warren Hrung, Laura Power, William Randolph, the editor, and two anonymous referees for helpful comments and discussion. 1. A foreign investment is considered to be direct (as opposed to portfolio) if the foreign investor controls the local enterprise, where control usually is defined as ownership of 10 percent or more. 2. This possibility has been suggested by (among others) Leff (1964) and Huntington (1968).

542 henry j. louie and donald j. rousslang

rife with corruption. Thus, it is strange that previous studies failed to demonstrate that poor governance discourages U.S. FDI, since the effect shows up so prominently for global FDI.3 The United States historically has been the most important home country for direct investment outflows,4 so this failure weakens the argument that improving the quality of governance will help a country to attract FDI. The present chapter takes a different approach from the earlier studies to determine how host-country governance affects U.S. FDI. Instead of looking at the effect on the amount of investment, we look at the effect on the rates of return that companies require on their FDI. To do this, we compare the after-tax returns earned by the various foreign subsidiaries of individual U.S. corporations. The notion is that poor host-country governance increases the risks, and hence the required return associated with local investment. To gauge the quality of host-country governance, we rely mainly on indexes created by Business International (now a part of the Economist Intelligence Unit), as did all of the earlier studies. The indexes are built from subjective assessments, made by a network of BI correspondents and analysts, about the investment climate in the host country. We also use the approach to examine whether a bilateral income tax treaty with the United States encourages U.S. FDI in the treaty partner. The effect of a tax treaty on investment is technically ambiguous. Companies sometimes lobby for tax treaties in order to obtain lower host-country withholding taxes on crossborder payments of dividends, interest and royalties, but the U.S. Treasury sometimes seeks a tax treaty in order to more effectively curb tax avoidance by U.S. multinational companies. The only empirical work we have found on the topic are two studies by Blonigen and Davies (2000, 2004). In the first study, the authors used a cross-section regression and found that U.S. FDI in the host country was positively correlated with the age of the tax treaty. Recognizing that the correlation could result from reverse causality (it is natural to expect that tax treaties would be concluded first with host countries in which U.S. investment ties were strongest),5 in the second study they used pooled time-series and crosssection data to account for country-specific fixed effects and concluded that the 3. It is possible that the Act’s effectiveness ebbed over time as investors and corrupt foreign government officials found ways around it. Looking at investment stocks in 1993, Wei (2000) found no evidence that U.S. investors were any more discouraged by hostcountry corruption than investors from other home countries, but he also was unable to find a statistically significant effect of corruption on U.S. FDI. 4. According to Organization for Economic Cooperation and Development (2001), the United States accounted for 23% of the total FDI outflows of OECD countries in the decade from 1990 to 1999. The United Kingdom was the second largest home country with 16% of the total. 5. This is true, whether U.S. FDI causes companies to lobby for a tax treaty, or whether it causes the U.S. Treasury to seek one as a means to limit tax avoidance on the part of U.S. companies.

host-country governance, tax treaties 543

tax treaties established in the 1970s and 1980s did not encourage (and might actually have reduced) U.S. FDI in the treaty partners. Our results support this conclusion, and indicate that it applies to the older treaties as well. An advantage of our approach over the ones taken by the earlier studies is that the after-tax rates of return account automatically for a great many factors that should be controlled for in a study explaining FDI but that are hard to quantify with the available data. For example, they take account of transportation costs for inputs and outputs, of taxes other than income taxes, and of valuable government services that might ameliorate the effect of local tax burdens by providing a higher pretax return. There are some conceptual differences between our results and those from the earlier studies. For one thing, the earlier studies lumped together the direct costs imposed by poor governance (such as the cost of bribes, the cost of bureaucratic delays, or the cost of insuring against expropriation of assets) with the greater uncertainty that poor governance tends to add to investment outcomes.6 If the direct costs are forecasted accurately and are deducted from the taxable income, our estimates will not capture their effect: The deductible costs merely increase the required return before deductions, leaving no net effect on the observed after-tax profit.7 Assuming the bulk of the direct costs are small or are deducted from taxable income (i.e., illegal bribes are small or are disguised as legitimate expenses), our estimates will show the extent to which the cost of corruption exceeds the cost of ordinary taxation. Another difference is the way the models treat corruption in the case where corrupt government officials deliberately restrict access to the local market to create monopoly profits that they can sell. In this case, the successful U.S. entrants might both invest more and enjoy higher rates of return.8 This could cause a regression of U.S. FDI on corruption to fail, or even to produce a perverse result if U.S. investors were more successful than others at gaining entry in such markets. Our regressions, however, would still show that corruption adversely affects investment by raising the rate of return.

6. Shleifer and Vishny (1993) provide a good account of why corruption imposes costs in excess of the direct cost of bribes. 7. Of course, in some instances investors can avoid the cost of the uncertainty by diversifying their investments, although such diversification opportunities would appear to be rare. 8. Unfortunately, the available indexes do not allow us to distinguish between different types of corruption in ways that might help us to determine more narrowly the likely economic consequences, such as the distinctions that Rose-Ackerman (1999) makes between bribes that clear the market (when government allocates a scarce benefit), those that act as incentives to clear bureaucratic roadblocks, those that lower costs (payments made to reduce customs duties or taxes), and those that permit criminal activity.

544 henry j. louie and donald j. rousslang

A. The model We examine the long-run equilibrium and assume that markets are broadly competitive, that investments can be undertaken in small increments, and that each U.S. multinational corporation requires an expected after-tax rate of return from investment in a foreign subsidiary that includes a premium for the risk associated with overall foreign investment and a premium for the risk that is specific to the host country.9 Under such circumstances, the multinational corporation will arrange its affairs so that the risk-adjusted after-tax return from a marginal investment is the same in each of its foreign subsidiaries. The host-country-specific risk may be affected by the quality of local governance and by a U.S. bilateral income tax treaty, as well as other factors. Thus, the required after-tax rate of return on investment in a foreign subsidiary (r) is given as: rs = f(rap, Ch, Th, Zh),

(1)

where ra is the company’s required after-tax return on FDI in general; C is a vector of variables representing various aspects of the quality of governance in the host country, such as corruption or political instability of the government; T denotes whether the host country has a tax treaty with the United States; Z is a vector of variables for other factors that affect risk or the short-run profitability of FDI; and the subscripts denote the subsidiary (s), the U.S. parent company (p), and the host country (h). It is necessary to account for factors that influence short-run profitability, such as short-run changes in local income or in the foreign exchange value of the local currency, because they can cause the subsidiary’s earnings in a given year to misrepresent the expected long-run returns that drive the investment decisions. The reported return from investment in a subsidiary might systematically misrepresent the actual contribution to the parent company’s overall profits, even in the short run, because the parent has the incentive to shift income among its affiliates in order to save on income taxes. Income can be shifted by manipulating payments among related parties for royalties, interest, or goods or services. Thus, the observed after-tax rate of return on investment in a subsidiary (ro) is the sum of r and the income shifted to the subsidiary. The incentive for income shifting depends mainly on the host-country’s statutory corporate income tax rate (t).10 Therefore, we use the following reduced-form regression equation to

9. Even if it is not risk averse, the parent corporation may require a higher rate of return to compensate for the possibility that the subsidiary’s assets may be expropriated, or that they cannot be fully recovered in the event that the investment goes sour. 10. There are also other determinants of income shifting. For example, Horst (1971) describes how attempts to save on import duties result in income shifting. Companies may also shift income to circumvent restrictions that the host country imposes on royalty payments to the U.S. parent. However, these other incentives are hard to quantify, so they are included in the error term of our regression.

host-country governance, tax treaties 545

estimate the effects of tax treaties and the quality of host-country governance on the observed rate of return: ros = α + βrap + γ Ch + δTh + .eth + ζZh + us,

(2)

where u is an error term. If t adequately accounts for income shifting and if its effect is independent from that of the other variables in the regression, then the other coefficients in equation 2 should accurately portray the effects on r, as well as on ro. The corruption and tax rate variables, however, might interact if corruption leads to inefficient administration of taxes, so the coefficient of C might understate the effect on r if corruption makes it easier for companies to shift income away from the local tax jurisdiction and thereby reduce the reported income. We explore this possibility in the empirical application of the model. B. The Data The data on activities of the U.S. parent corporations and their foreign subsidiaries are drawn from confidential federal corporate tax returns. Data on the book value of the foreign subsidiary’s assets, on its earnings, on the taxes paid on its earnings, and on its payments of dividends, interest and royalties to the U.S. parent corporation are from IRS Form 5471, “Information Return of U.S. Parents with Respect to Certain Foreign Corporations.” The data are tabulated by the Statistics of Income (SOI) Division of the Internal Revenue Service. The assets, earnings, taxes and payments to the U.S. parent are all measured in U.S. dollars. At time of writing, data from Form 5471 needed for our analysis were available only for fiscal years 1992, 1994 and 1996, and only for the 7,500 largest subsidiaries (in terms of assets) of U.S. parents with assets of $500 million or more. Data on the parent’s foreign tax credits are from IRS Form 1118, “Foreign Tax Credit – Corporation,” which is also tabulated by SOI. The withholding tax rates under the U.S. tax treaties are from U.S. Department of Treasury (1998). The withholding tax rates for non-treaty countries are from information guides published by Price Waterhouse and by Ernst and Young, and from the descriptions of tax systems published by the International Bureau of Fiscal Documentation (2002 updates). The statutory corporate income tax rates are those constructed by Altshuler, Grubert, and Newlon (1998). They are the rates reported by Price Waterhouse, adjusted for various special cases, such as tax holidays. Data for corruption, political instability and bureaucratic inefficiency are from indexes constructed by Business International and Transparency International. (These data are described in more detail in section 3.3). The data used to calculate the 1-year and 10-year rates of growth of the host-country economy, as well as the variance of the growth rates for the decade preceding each sample year, are drawn from the International Monetary Fund publication International Financial Statistics (various issues).

546 henry j. louie and donald j. rousslang

C. Applying The Model 1. Inferring the required after-tax rates of return This section describes in detail the variables used to apply the model. Unless direct foreign investments are allor-nothing propositions, the multinational company will tend to bring into equality the marginal (rather than the average) after-tax returns to investment in its foreign subsidiaries. The tax return data, however, allow us to observe only the average after-tax rates of return. A disadvantage of this approach is that using the average returns in place of the marginal returns introduces measurement error in the dependent variable, which reduces the statistical fit of the regression and biases the coefficients toward zero. The United States generally taxes earnings of a foreign subsidiary only when the earnings are repatriated as dividends by the U.S. parent. Therefore, the U.S. parent’s after-tax return from its investment in a foreign subsidiary includes its share of the subsidiary’s retained earnings plus the net-of-tax values of the payments to the parent (or another controlled affiliate) of interest and dividends.11 The great bulk of our sample (over 80%) consists of wholly-owned foreign subsidiaries. For these entities, the distinction between the parent’s debt and equity positions in the foreign subsidiary is largely artificial and driven by tax considerations, so there is little reason to distinguish between the after-tax values of the dividend and interest payments.12 The parent may also receive royalty payments from the subsidiary. Technically, a royalty payment is a return on a domestic asset held by the parent, but it sometimes represents a return that the parent could not have garnered without the subsidiary’s operations, in which case it should be considered as part of the total return from the investment in the subsidiary. To explore the sensitivity of our results to the various possibilities, we measure the subsidiary returns both with and without the royalty payments. We assume that, with the exception of random error and the short-run cyclical effects that are explained by the variables in our regression equation, the observed after-tax rate of return (and the pattern of payments from the foreign subsidiary) represents the long-run equilibrium. The taxes on the returns from the investment in the subsidiary include the host-country income tax and the residual U.S. tax (the so-called “repatriation tax”) on the payments to the parent. We calculated the numerator of ro as the parent’s share of the subsidiary’s earnings, plus its payments of interest and royalties to the parent, minus the repatriation taxes. The repatriation taxes consist of the withholding taxes in the host country and any residual U.S. tax after the foreign tax credit, so they depend on the foreign tax credit position of the U.S. parent.

11. The parent’s share of the repatriation payments may come indirectly through other related affiliates, but as long as they leave the original host country, they are subject to current U.S. tax under subpart F of the Internal Revenue Code, even if they remain abroad. 12. The interest payments account for about seven percent of the total income generated by the foreign subsidiaries in our sample.

host-country governance, tax treaties 547

If the parent owes a residual U.S. tax on its foreign source income, it is said to have an excess-limit position. For excess-limit parents, we calculated the repatriation taxes as the statutory U.S. corporate income tax rate (tU.S.) multiplied by payments from the subsidiary to the U.S. parent of dividends (grossed-up to reflect the underlying earnings), interest and royalties, minus the foreign taxes imposed by the host country, including income taxes on corporate profits underlying the dividends and taxes withheld at source. This is represented algebraically as follows: Reptaxexcess limit = (tU.S. − tF)∗(div)/(1 − tF) + tU.S.∗(int + roy) − (whd∗div) − (whi∗int) − (whr∗roy),

(3)

where tU.S. is the applicable U.S. corporate income tax rate, tF is the average host-country income tax rate; whd is the host-country withholding rate on dividends; whi is the host-country withholding rate on interest; whr is the host-country withholding rate on royalties; div is the dividend payments from the subsidiary to the U.S. parent; int is the interest payments from the subsidiary to the U.S. parent; and roy is the royalty payments from the subsidiary to the U.S. parent. The foreign tax credit is limited to the latent U.S. tax, so if the parent pays more in foreign taxes than it can deduct from the U.S. tax on its foreign earnings, it is said to have excess foreign tax credits. For excess-credit parents, the repatriation taxes are simply the host country’s withholding taxes on the subsidiary’s payments to its parent, represented algebraically as follows:13 Reptaxexcess credit = (whd∗div) + (whi∗int) + (whr∗roy).

(4)

It is hard to measure the repatriation tax accurately, because a company’s foreign tax credit position often varies from year to year, and foreign tax credits that are not used in one year can be carried over into other years. Additionally, some discounted value of the tax may apply to earnings retained by the subsidiary. For our central case, the repatriation tax is calculated assuming the U.S. parent’s current year position (as reported on its tax return) is also the long-run position. To see how sensitive our results are to the treatment of this tax, we also considered several other possibilities: the repatriation taxes are ignored in the company’s assessment of the after-tax rates of return (so ro is measured gross of these taxes); the repatriation taxes are calculated as if all of the parent companies are in long-run excess credit; the repatriation taxes are calculated as if all of the parents are in long-run excess limit; and the repatriation tax was assumed to apply to the firm’s retained earnings. This last variation is akin to assuming that

13. All payments by a subsidiary to a related affiliate in another country are treated as payments to the U.S. parent, because the payments are subject to U.S. tax under the subpart F provisions of the Internal Revenue Code.

548 henry j. louie and donald j. rousslang

the current rate of residual tax represents the parent’s cost of keeping earnings abroad. The results are not much different for any of the specifications.14 The denominator of ro is measured as the total assets of the subsidiary, including assets funded by external debt. Thus, if the subsidiary has external debt, ro will understate the rate of return on assets, because the external interest payments are not included in the numerator. We measure ro this way, because we cannot accurately separate external and related-party debt owed by each foreign subsidiary. For the wholly owned subsidiaries that comprise the bulk of our sample, however, the ratio of external debt to assets should be the same as for the consolidated company, and differences in the ratios can be attributed to income shifting. Therefore, since ra is measured as the average of the returns that the U.S. parent earns from its other foreign subsidiaries (besides the one producing ro), this variable should normalize for differences in the external debt ratios among the subsidiaries. That is, if the regression adequately accounts for income shifting, both ro and ra should be understated by the same proportion and ra should capture the effect of differences among consolidated companies in the ratio of external debt to equity. The variable ra should also help to normalize for differences in overall profitability among different affiliated groups that might result from differences in the value of intellectual property rights. There may be big differences among industries in the responsiveness of FDI to poor governance. For example, financial services companies are likely to be less sensitive to risks of expropriation than a manufacturing company whose investments mean committing a substantial amount of physical capital. FDI in extractive industries may be less sensitive to costs of poor governance if the host country has monopoly power in the market for the raw materials and the companies can pass the costs forward to consumers. Also, if (as often happens) the U.S. parent has subsidiaries in more than one of these broad industry groups, ra becomes less effective at capturing company differences in debt-equity ratios, since the ratios vary widely by industry. For these reasons, we limit the sample to the manufacturing subsidiaries. The subsidiary’s reported earnings in a single year often fail to reflect the long-run annual contribution to the parent’s profits owing to reporting errors or random one-off events that are impossible to account for systematically. Examples are a large one-time profit or loss from the sale of an asset, or a write-off that reflects a large capitalized loss in the value of an intangible asset (a goodwill write-off). Therefore, the observed return from a foreign subsidiary was deleted from the sample if the reported data showed an obvious anomaly—nonexistent or negative sales, foreign taxes that are negative or that exceed foreign earnings, or an annual income or loss for the subsidiary that exceeds the total value of

14. Of the four alternative specifications, only the first is reported (in Table 3). The others are available from the authors on request.

host-country governance, tax treaties 549

its assets. We also ran regressions in which the profit or loss is limited to half of the value of assets. Curtailing the sample on the basis of the rates of return may provide more accurate results, because a profit or loss greater than all (or even half) of the subsidiary’s assets is much more likely to represent a one-time capitalization of events that have accumulated over a number of years, or a faulty measure of assets, rather than a true annual return. If no limits are placed on the rates of return, statistical noise completely dominates the regression results.15 2. Measuring the tax treaty variable A dummy variable was used to represent the presence of a bilateral income tax treaty between the United States (T) and the country where the subsidiary is located. In a few cases, the treaty was not in effect very long by the beginning of the sample period, and the FDI stocks might not have adjusted fully. It is hard to know when such adjustments occur, however, because negotiations typically have been underway for a number of years by the time the treaty enters into force and investors may anticipate the result. Therefore, separate regressions were run for two cases. In the first case, T took a value of unity if a treaty was in force or entered into force during our sample period. In the second case, T took a value of unity only if the treaty had been in force for at least five years prior to the sample period (i.e., by 1987). The results were substantively the same for both regressions, so we report only the second case, which produced a slightly more significant coefficient for T. The effective dates of U.S. bilateral income tax treaties are from U.S. Department of Treasury (1998). 3. Measuring the quality of host-country governance Our independent variables C1 and C2 are constructed from the BI indexes for corruption and political instability. BI provides separate indexes for these variables, but as Wei (2000) pointed out, they are interdependent. Political instability can make the tenure of government officials uncertain, causing them to discount the cost of being dismissed for malfeasance. In turn, widespread corruption can breed popular dissatisfaction with the existing government. BI also provides separate indexes for bureaucratic red tape, the efficiency of the judiciary system and corruption, but these variables, too, are interdependent. For example, excessive red tape and an inefficient judiciary increase the opportunity for mischief by corrupt officials, and officials seeking to gain from corrupt practices would encourage more red tape and discourage judiciary reform. The indexes measuring these characteristics are also highly correlated with each other and, as Shleifer (2000) has noted, it would be hard to distinguish statistically between their effects. Therefore, following Mauro, we combine all three of them into an index of “bureaucratic efficiency (variable C3)” that we used in 15. In the uncensored sample, the highest rate of return is 1,688,700% and the lowest is –8,900%. (The average rate of return in the censored sample is 8.9 percent.) When we performed the regressions shown in Table 2 using the raw data, none had an F-value greater than 1.2, indicating that none was significant at the 45% level.

550 henry j. louie and donald j. rousslang

place of the corruption index. In our sample of forty-six host countries, however, the simple correlations between the indexes of corruption, red tape and judiciary efficiency vary between 0.72 and 0.84, whereas the highest correlation between any of these indexes and the index for political stability is only 0.57. Therefore, we include political stability as a separate variable in the regressions. The BI indexes are based on subjective assessments of host-country investment conditions. Subjective perceptions are important ingredients in investment decisions, however, so the indexes might well be better for our purpose than more objective measures of corruption or political instability, such as the size or frequency of bribery payments, or the frequency with which assets are expropriated. We reversed the BI index values presented in Mauro’s study, so that on a scale from zero to one, a higher number indicates greater corruption, political instability or bureaucratic inefficiency. The indexes are shown in Table 1, along with the host-country means of other variables used in the analysis. The BI values for host-country governance are averages of the indexes created for the period from 1980 through 1983, which precedes by nine years the period for which we calculate rates of return on FDI (1992–1996).16 A similar lag is present in the earlier study by Wei (2000), who used the same corruption data to obtain his central estimates for the effect on foreign investment stocks in 1993. Wei supplemented his central results with those based on an average of corruption indexes created by the International Country Risk Group (ICRG) for the years 1991–1993, and with an index constructed by Transparency International (TI) that is based on an average of surveys of corruption “over a number of years,” and that he appears to regard as less reliable than the BI index. Both Wei and Mauro used averages of survey responses over a period of years, which implies that they viewed year-to-year changes with skepticism. At least two issues need be addressed when setting the lag between corruption perceptions and investment returns. The first issue is whether annual changes in the perceptions of corruption, as measured by the available surveys, are meaningful. If they are not, then it is probably better to use an average of survey results over several years, as we did (and as Wei and Mauro did). In this regard, we believe that perceptions of political stability are more likely to change rapidly than are perceptions of corruption. The second issue is the appropriate lag to use. We expect the investment stocks to adjust to perceptions of corruption more rapidly than the rates of return, because it takes time for new investment to produce its returns. The lag in our regressions, however, is about the same as that in Wei’s regression of capital stocks against corruption. 4. Measuring the other variables Two variables are included to capture the effect of short-run changes in income or exchange rates for the host country: Z1 is the growth of income in the host country in the current year and Z2 is the

16. The indexes are described in detail in Business International (1984).

table 1. means of host-country variables for the sample years (1992, 1994, and 1996) host country variables a Host Country

C1

40.3 97 15.7 66 90 209.3 7.3 17.3 16.7 1.3 1.3 4.7 205 195

0.234 0 0.2 0.025 0.425 0 0.075 0.55 0.075 0.45 0.675 0.05 0 0.05

C2

0.228 0.15 0.096 0.2 0.246 0.1 0.354 0.4 0.15 0.337 0.133 0.121 0.108 0.179

C3

0.323 0.025 0.175 0.092 0.483 0.042 0.142 0.458 0.042 0.442 0.575 0.067 0.175 0.133

T

0 1 1 1 0 1 0 0 1 0 1 1 1 1

t

0.272 0.36 0.359 0.396 0.314 0.367 0.15 0.335 0.357 0.25 0.34 0.289 0.346 0.565

Z1

0.105 0.079 0.069 0.065 0.066 0.002 0.17 0.234 0.076 –0.087 0.108 0.02 0.055 0.067

Z2

0.107 0.093 0.177 0.162 0.196 0.075 0.182 0.103 0.15 –0.096 0.017 0.093 0.138 0.209

Z3

1.22 0.012 0.019 0.017 0.098 0.005 0.012 0.022 0.019 0.039 0.048 0.027 0.016 0.02

ro μ

σ2

0.109 0.08 0.108 0.061 0.101 0.083 0.133 0.183 0.099 d d 0.102 0.073 0.072

0.018 0.012 0.013 0.012 0.019 0.008 0.010 0.012 0.006 d d 0.025 0.011 0.016 Continued

host-country governance, tax treaties 551

Argentina Australia Austria Belgium Brazil Canada Chile Colombia Denmark Ecuador Egypt Finland France Germany

No. of Subs.

Host Country

Hong Kong India Indonesia Ireland Israel Italy Japan Korea Malaysia Mexico Morocco Netherlands New Zealand Nigeria Norway

No. of Subs.

C1

26.3 13.3 8 29.3 9.7 139.7 102.7 26.3 25 78.3 0.7 97 12.7 0.3 6

0.2 0.475 0.85 0.025 0.075 0.25 0.125 0.425 0.4 0.675 0.434 0 0 0.7 0

C2

0.05 0.3 0.254 0.233 0.375 0.208 0.058 0.25 0.158 0.312 0.289 0.117 0.15 0.271 0.05

C3

0.075 0.45 0.775 0.133 0.108 0.367 0.092 0.392 0.3 0.517 0.412 0 0 0.567 0.033

T

0 0 0 1 0 1 1 1 0 0 1 1 1 0 1

t

0.174 0.493 0.327 0.1 0.121 0.509 0.563 0.338 0.322 0.343 0.38 0.35 0.33 0.3 0.28

Z1

0.137 0.062 0.115 0.107 0.112 0.071 0.025 0.1 0.148 0.115 0.077 0.062 0.081 0.428 0.07

Z2

0.272 0.061 0.118 0.188 0.169 0.155 0.218 0.352 0.166 0.118 0.111 0.141 0.11 0.138 0.107

Z3

0.004 0.008 0.007 0.017 0.014 0.021 0.016 0.008 0.007 0.026 0.014 0.016 0.016 0.185 0.008

ro μ

σ2

0.135 0.069 0.088 0.149 0.109 0.06 0.062 0.095 0.123 0.125 d 0.108 0.046 d 0.063

0.014 0.008 0.022 0.026 0.015 0.011 0.006 0.019 0.014 0.021 d 0.016 0.013 d 0.009

552 henry j. louie and donald j. rousslang

table 1. means of host-country variables for the sample years (1992, 1994, and 1996) host country variables a (cont’d...)

a

1.3 7 4 15 21.3 38 12.3 94 22 32.3 19.7 14 259.7 2.3 23.3 1

0.6 0.5 0.275 0.55 0.325 0 0.2 0.3 0.075 0 0.85 0.4 0.075 0.2 0.425 0.125

0.467 0.246 0.396 0.392 0.246 0 0.35 0.333 0.1 0.075 0.437 0.183 0.167 0.1 0.229 0.35

0.567 0.367 0.341 0.525 0.442 0 0.3 0.358 0.075 0 0.733 0.489 0.1 0.317 0.458 0.2

1 0 0 1 0 0 0 0 1 1 0 0 1 0 0 0

0.49 0.09 0.3 0.35 0.4 0.1 0.41 0.35 0.287 0.18 0.3 0.414 0.33 0.3 0.309 0.423

0.067 0.078 0.147 0.154 0.109 0.15 0.028 0.05 0.065 0.039 0.127 –0.039 0.057 0.114 0.037 0.011

0.077 0.059 0.106 0.111 0.294 0.307 0.089 0.198 0.127 0.144 0.253 0.205 0.125 0.188 0.001 0.023

0.003 0.008 1.574 0.006 0.033 0.006 0.017 0.023 0.018 0.021 0.005 0.051 0.012 0.016 0.026 0.013

d 0.14 0.116 0.217 0.129 0.126 0.089 0.08 0.087 0.135 0.111 0.061 0.077 d 0.091 d

d 0.012 0.008 0.011 0.010 0.014 0.008 0.016 0.014 0.021 0.011 0.02 0.012 d 0.033 d

C1 and C2 are the BI indexes for corruption and political stability. C3 is the average of the BI indexes for corruption, bureaucratic red tape, and judiciary efficiency. All of the indexes are averages for the years 1980 through 1983 and all of them have been divided by 100. T is the treaty dummy; a “1” means that a bilateral income tax treaty with the United States was in force by 1987. Z1 is the average annual growth in GDP for the sample years, Z2 is the average of the rates of growth in GDP for the decade immediately preceding each sample year and Z3 is the variance of the growth rate in GDP for the decade preceding the sample year. A “d” means that the cell value is not given to prevent disclosing confidential business information.

host-country governance, tax treaties 553

Pakistan Panama Peru Philippines Portugal Singapore South Africa Spain Sweden Switzerland Thailand Turkey United Kingdom Uruguay Venezuela Zimbabwe

554 henry j. louie and donald j. rousslang

annual average rate of income growth in the most recent ten years. Each growth rate is measured in U.S. dollars, so it is the sum of the local rate of inflation (measured in local currency), the percentage change in the local currency’s dollar exchange rate, and the rate of growth of real GDP. Thus, the variables should normalize for the effect on subsidiary earnings caused by temporary fluctuations in the exchange rate, in local inflation, or in real GDP growth. Some countries may experience greater annual variability in income owing to factors other than the quality of local governance, such as a reliance on natural resource exports. This may cause subsidiaries that produce for the local market to experience greater uncertainty in expected returns. To take account of such uncertainty, we include a variable (Z3) that is measured as the variance of the host country’s growth rates for the decade preceding each sample year. The North American Free Trade Agreement (NAFTA) came into effect in 1994, in the midst of our sample period. The agreement contains some provisions that might have reduced the required rates of return over the longer run by reducing uncertainty for investments in Canada or Mexico (such as the provisions that reaffirm the principle of national treatment of foreign subsidiaries, those that provide better protection for intellectual property rights, and those that provide new mechanisms for resolving investment disputes),17 but the agreement also provided greater profits for U.S. FDI already in place in the countries by reducing tariffs on their trade with the United States. Canada and Mexico account for an important share (14%) of our subsidiary observations, so we include a dummy variable to account for the effects of NAFTA. D. The Empirical Results The observations are pooled for the three sample years, because we could not reject the hypothesis that the coefficients of the independent variables were the same in all of the years, either individually or all together. Table 2 presents the regression results from our central case: the royalty payments are included in the subsidiary’s returns, the parent’s current-year foreign tax credit position is used to calculate the repatriation taxes, as shown in equations (3) and (4), and profits or losses that exceed half of the subsidiary’s assets are excluded from the sample. The coefficients of ra are highly significant in all of the regressions, indicating that there are important differences in the average rate of return among the companies. As already explained, given the way ro and ra are measured, the company differences may reflect differences between affiliated groups in debt-equity ratios, or differences in profits contributed by valuable intellectual property rights. The coefficient of Z1 has the expected sign, and together with the coefficient of Z2 indicates that a temporary growth spurt is accompanied by

17. The provisions of NAFTA are described in United States Government (1993).

host-country governance, tax treaties 555

table 2. tax treaties, the quality of host-country governance, and the rate of return to u.s. fdi pooled regression a , 1992, 1994, and 1996

Dependent Variable: after-tax rate of return from the foreign subsidiary (ro), measured gross of related-party interest and royalty payments Independent variable

(2i)

(2ii)

Intercept

0.095∗∗ 0.088∗∗ (14.40) (14.92) U.S. parent’s rate of return 0.464∗∗ 0.473∗∗ on FDI (15.71) (16.24) Host country current year 0.064∗ 0.058∗ GDP growth (Z1) (2.29) (2.21) Host country growth in GDP in –0.013 –0.018 prior ten years (Z2) (–0.33) (–0.47) Variance of annual host country –0.007 –0.003 growth in GDP (Z3) (–0.71) (–0.32) Dummy for 1992 0.000 –0.000 (0.02) (–0.05) Dummy for 1994 0.001 0.001 (0.27) (0.35) Host country statutory tax –0.108∗∗ –0.127∗∗ rate (t) (–8.48) (–10.70) Treaty (T) –0.013∗∗ –0.000 (–2.86) (–0.02) NAFTA 0.007 0.005 (1.50) (0.95) Corruption (C1) – 0.036∗∗ (3.47) Political instability (C2) – – Bureaucratic inefficiency (C3) Average of BIindexes) Corrected observations Adjusted R2 a

–

–

6467 0.09

6361 0.10

(2iii) 0.083∗∗ (10.58) 0.476∗∗ (16.29) 0.061∗ (2.20) –0.012 (–0.27) –0.006 (–0.69) –0.000 (–0.04) 0.001 (0.28) –0.121∗∗ (–9.65) –0.006 (–1.01) 0.008 (1.54)

0.057∗∗ (3.76) – 6361 0.10

(2iv) 0.093∗∗ (13.93) 0.475∗∗ (16.18) 0.066∗∗ (2.59) –0.021 (–0.51) –0.007 (–0.75) –0.001 (–0.14) 0.001 (0.29) –0.117∗∗ (–8.69) –0.009 (–1.65) 0.006 (1.28) – – 0.013 (1.44) 6361 0.10

(2v) 0.083∗∗ (11.20) 0.474∗∗ (16.62) 0.056∗ (2.05) –0.014 (–0.32) –0.004 (–0.36) –0.000 (–0.00) 0.001 (0.33) –0.129∗∗ (–11.00) 0.007 (0.16) 0.006 (1.13) 0.027 (1.52) 0.03 (0.97) – 6361 0.10

The t-statistics are based on Newey-West standard errors. A single asterisk (∗) indicates significance at the five percent level, and a double asterisk (∗∗) indicates significance at the one percent level.

556 henry j. louie and donald j. rousslang

higher observed rates of return. The coefficients for Z3 and for the year dummies are insignificant. The coefficient of NAFTA dummy was not statistically significant at the usually accepted levels, but it indicates that the agreement increased the observed after-tax rates of return to U.S. FDI in Canada and Mexico. The negative coefficients of t imply that income is shifted away from host countries that impose higher taxes. The coefficients were highly significant in all the regressions. Evaluated at the sample mean for ro (= 0.10) and using regressions 2ii through 2v, the coefficients imply that a percentage point increase in t would reduce reported after-tax profits of the subsidiary by 1.2%–1.3%.18 The coefficient of the tax variable, which is supposed to represent an income shifting incentive in our model, falls in absolute value in the altered specification (becomes a smaller negative number). There are at least two reasonable explanations for this result. First, if the original specification overstates the value of retained earnings by ignoring the repatriation tax on them, then the calculated after-tax rate of return will be too high for countries that have a low tax rate, particularly because they also tend to have large retained earnings. This would cause a spurious negative correlation between our dependent variable and the tax rate variable and cause its coefficient to overstate the effect of income shifting. On the other hand, if the altered specification understates the value of retained earnings by overstating the repatriation tax on them, then the calculated after-tax rates of return will be too low for countries that have a low tax rate, causing the coefficient of the tax rate variable to understate the effect of income shifting. The coefficient of the treaty dummy (T) is insignificant when the regression includes a variable for host-country corruption or political stability (regressions 2ii, 2iii and 2v). This result also holds when ro is measured before subtracting any repatriation taxes (regression 3iii), implying that the reductions in withholding tax rates that the treaties provide do not have much influence on investment decisions. This result is consistent with the theoretical predictions. Hartman (1985) has shown that the withholding taxes should not affect the required rate of return on investment financed from a mature foreign subsidiary’s retained earnings, and Sinn (1991, 1993) has shown that they should not affect the longrun, steady-state investment levels, even for a new foreign venture. It is also consistent with the results reported by Blonigen and Davies (2000), who found

18. The estimates in Hines and Rice (1994) imply that this tax increase would reduce the reported before-tax rate of return (r0B) by about 1.7 percent. To compare our results with theirs, first note that in our model, rOB is equal to r0/(1 - t), where t is the average effective tax rate, and (dr0B/dt)/r0B = (dr0/dt)/r0 + (dt/dt)/(1 - t), where dr0/dt is given by the coefficient of t from our regressions. Thus, assuming that dt/dt is positive and less than one (recall that t is the rate of tax on reported, rather than actual, income), and evaluating (dr0B/dt)/r0B at the sample means t0 = 0.21 and r0 = 0.10, the estimates imply that a percentage point increase in t would reduce r0B by between 1.2 percent and 2.5 percent.

host-country governance, tax treaties 557

that the coefficients of the treaty variables in their regressions did not change much when they took account of withholding taxes. Reverse causality between T and the required rate of return might be present in our cross-section regressions, because a lower rate of return might reflect larger investment and, consequently, greater interest in a tax treaty. The coefficient of T is insignificant, though, indicating that reverse causality, if it exists, is not important enough to matter (unless the tax treaties actually discourage U.S. FDI). However, the coefficient of T is highly significant, both statistically and absolutely, in regression 2i, which includes no measure of the quality of hostcountry governance. This demonstrates that it is important to take account of the quality of host-country governance when exploring the effect of the tax treaties on U.S. FDI. Our results are consistent with the notion that good governance attracts both FDI and a U.S. income tax treaty, but they offer no support for the conclusion that a tax treaty encourages U.S. FDI. The variables for corruption (C1) and political instability (C2) both have positive and significant coefficients when included separately in the regression explaining ro (regressions 2ii and 2iii). The coefficients imply that corruption or political instability increase the observed after-tax rates of return. When both variables are included in the regression, however, both become less significant and the coefficient of C2 is insignificant at the usually accepted levels.19 Thus, we are unable to identify the separate effect of political instability. According to the coefficients of C1 in regression 2ii and of C2 in regression 2iii, the effect on the rate of return of moving from the most corrupt to the least corrupt host country in our sample is about the same as the effect of moving from the least politically stable to the most politically stable host country.20 The coefficient of C1 implies that compared to Singapore, corruption raised the average after-tax rates of return to U.S. FDI in Mexico by 2.4 percentage points, or by the equivalent of a tax of 15% on the mean pre-tax return from the Mexican subsidiaries.21 This effect is lower than the bottom of the range reported by Wei (2000), who estimated that the difference in the level of corruption between these countries has the effect on investment of an increase in the income tax rate 19. This result also occurred with each of the alternative specifications of ro described in section IV. 20. The index C1 ranges from 0 to 0.85. Multiplying this range times the coefficient of C1 (=.036) yields an increase in the after-tax return of 3.1 percentage points. The index C2 ranges from 0 to 0.467. Multiplying this range times the coefficient of C2 (=0.57) yields an increase in the after-tax rate of return to 2.7 percentage points. 21. The coefficient of C1 in the regression is 0.036. The value of C1 is 0.675 for Mexico and 0 for Singapore. The mean pretax rate of return of the Mexican subsidiaries was 16.7%. Thus, the estimated effect of the difference in corruption as a percent of this mean return is [0.036∗0.675∗100]/0.167 = 14.6%. Using one standard deviation around the coefficient for the corruption variable and choosing the resulting highest and lowest values implied by 2ii, 3i, 3ii, or 3iii yields a low estimate of 7.7 percent and a high estimate of 20.4%.

558 henry j. louie and donald j. rousslang

of from 18–50 percentage points. Recall, however, that Wei’s estimates include the effect of the direct cost of corrupt payments that can be deducted from taxable income, whereas ours reflect only the cost of bribes that cannot be deducted from taxable income and of the uncertainty associated with corruption.22 Table 3 presents a variety of modifications for regression 2ii that test the robustness of our central results. Regression 3i uses weaker criteria for excluding outliers. Namely, it excludes a subsidiary from the sample if its annual profit or loss exceeds its total assets. In regression 3ii the subsidiary returns are measured net of royalty payments to related parties. As explained previously, in regression 2ii the royalty payments are considered as part of the subsidiary’s contribution to the U.S. parent’s total profits. The assumption implicit in this specification is that the parent could not successfully exploit the intangible assets for which the royalty was paid without its investment in the subsidiary. At the other extreme, the U.S. parent could license the intangibles to an unrelated party and receive the same return, so none of the royalty payments should be considered as part of the subsidiary’s contribution to the parent’s total profits. Regression 3ii provides results for this alternative case. In regression 3iii the subsidiary returns are measured gross of repatriation taxes and in regression 3iv the parent is assumed to pay repatriation tax in the current year on all of the subsidiary’s income, including its retained earnings.23 As explained earlier, it is difficult to determine the repatriation tax with precision. Regressions 3iii and 3iv test the robustness of our central results for reasonable upper and lower bounds for the repatriation tax. Finally, regression 3v includes dummies for the parent companies to test for company-specific effects.24 The results from all of these alternative specifications are qualitatively the same as those for regression 2ii. 22. As noted in section II, the coefficient of C1 might understate the effect of corruption on the required rate of return if corruption makes it easier for companies to reduce the local reported income by bribing local tax officials. One might expect the opportunity for bribing tax officials to increase with the general level of corruption and the incentive for such behavior to increase with the local statutory tax rate. To investigate the possible interaction of these variables, we added a new term that is their product to regression 2ii. The results (not shown) provide no evidence of an interaction, as the coefficient of the new term was statistically insignificant and the significance of the regression equation (the F value) also declined. Thus, we could find no evidence that corruption among tax officials matches the general level of corruption in the host country. 23. The results were also substantially the same when the repatriation taxes were measured assuming all of the parents are in long-run excess limit and when they were measured assuming all of the parents are in long-run excess credit. 24. Our programs were unable to accommodate a regression that simultaneously accounts for heteroskedasticity, serial correlation and parent dummies for the full sample of 392 U.S. parents and their 6361 foreign subsidiaries. We were, however, able to run such a regression for a subset that contains the 53 largest U.S. parents and their 3,177 foreign subsidiaries, which is half of the full sample. The results from the smaller sample are qualitatively the same as those from regression 3vi, which is based on ordinary standard errors.

0.089∗∗ (13.40)

0.063∗∗ (12.03)

U.S. parent’s rate of return on FDI Host country current year GDP growth (Z1) Host country growth in GDP in prior ten years (Z2) Variance of annual host country growth in GDP (Z3) Dummy for 1992

0.502∗∗ 18.09 0.040 (1.16) 0.031 (0.72) –0.009 (–0.93) 0.000 (0.12)

0.435∗∗ (16.95) 0.053∗ (2.13) –0.013 (–0.35) 0.000 (0.01) –0.000 (–0.04)

0.5∗∗ (17.37) 0.064∗ (2.50) –0.036 (–0.95) –0.001 (–0.16) –0.001 (–0.27)

0.357∗∗ (14.51) 0.038 (1.93) –0.022 (–0.73) –0.008 (–1.31) 0.000 (0.28)

–0.066 (–1.80) 0.049∗∗ (2.46) –0.118 (–0.50) –0.006 (–0.69) –0.005 (–1.22)

(3vii) Country averages for ro and ra

(3iv) Assume immediate repayment of repatriation taxes

0.083∗∗ (15.61)

(3vi) TI corruption variable for 2001

(3iii) ro measured gross of repatriation taxes

0.092∗∗ (14.93)

(3v) Parent company dummies

(3ii) ro measured net of related-party payment

Intercept

Independent variable

0.091∗∗ (8.67)

0.116∗∗ (3.54)

0.474∗∗ (15.95) 0.065 (1.26) –0.018 (–0.61) –0.01 (–1.66) –0.000 (–0.06)

0.249 (1.25) 0.138 (1.70) 0.02 (0.27) 0.006 (0.38) –0.006 (–0.47) Continued

host-country governance, tax treaties 559

(3i) Sample limited to -1
table 3. tax treaties, the quality of host-country governance, and the rate of return to u.s. fdi variations on regression 2ii pooled regression a , 1992, 1994 and 1996 Dependent variable: after-tax rate of return from the foreign subsidiary (ro), measured gross of related-party interest and royalty payments

a

(3vii) Country averages for ro and ra

Corrected observations Adjusted R2

(3vi) TI corruption variable for 2001

Corruption (C1)

(3v) Parent company dummies

NAFTA

(3iv) Assume immediate repayment of repatriation taxes

Treaty (T)

(3iii) ro measured gross of repatriation taxes

Host country statutory tax rate (t)

(3ii) ro measured net of related-party payment

Dummy for 1994

(3i) Sample limited to -1
Independent variable

–0.000 (–0.07) –0.141∗∗ (–7.82) –0.004 (–0.61) 0.006 (1.06) 0.032∗ (2.42) 6483 0.10

0.002 (0.55) –0.123∗∗ (–9.46) –0.003 (–0.60) 0.004 (0.93) 0.040∗∗ (3.56) 6385 0.09

0.001 (0.41) –0.123∗∗ (–11.62) –0.000 (–0.03) 0.005 (0.95) 0.035∗∗ (3.75) 6350 0.11

0.002 (0.80) –0.087 (–12.47) 0.004 (0.95) 0.000 (0.05) 0.038∗∗ (6.13) 6361 0.08

–0.001 (–0.14) –0.128∗∗ (–8.16) –0.003 (–0.47) 0.012∗ (2.08) 0.029∗ (2.53) 6361 0.18

0.001 (0.15) –0.119∗∗ (–4.11) –0.009 (–0.75) 0.008 (1.54) 0.000 (0.62) 6359 0.10

–0.006 (–0.36) –0.046 (–1.41) –0.005 (–0.64) –0.001 (–0.06) 0.029∗ (2.04) 135 0.05

The t-statistics are based on Newey-West standard errors for regression 3i through 3v and 3vii. The t-statistics for regression 3vi are based on ordinary standard errors. A single asterisk (∗) indicates significance at the five percent level, and a double asterisk (∗∗) indicates significance at the one percent level.

560 henry j. louie and donald j. rousslang

table 3. tax treaties, the quality of host-country governance, and the rate of return to u.s. fdi variations on regression 2ii pooled regression a , 1992, 1994 and 1996 (cont’d...)

host-country governance, tax treaties 561

In addition to the tests for robustness, we also ran two other regressions. Regression 3vi uses the corruption index constructed by TI in place of the BI index. The results were poor and indicated that the TI index has almost no power to explain the required rates of return. This might be because the index was constructed for 2001, which is after our data on rates of return. Also, the index is an average of disparate surveys, so it might produce inconsistent cross-country comparisons. Regression 3vii uses country averages for ro and ra (calculated as the dollar-weighted averages of the corresponding disaggregate variables). The results are weaker than those from regression 2ii, which demonstrates the advantage of using the disaggregate data. E. Summary of Main Findings We find that poor host-country governance, as indicated by indexes measuring corruption or political instability, causes U.S. companies to require a significantly higher rate of return on their FDI. For example, our estimates imply that an increase in the level of corruption from that present in Singapore to that present in Mexico causes the companies to raise the required rate of return by about 2.4 percentage points. This amounts to about 15% of the observed mean pretax returns of the U.S. subsidiaries in these countries. The effects of corruption that we find do not reflect the added returns that companies require to compensate them for the cost of bribes that are structured as deductible expenses. Instead, they reflect only the cost of non-deductible (i.e., illegal) bribes and the premium for the added uncertainty associated with corruption. If non-deductible bribes are inconsequential, the estimates show the amount by which the cost of corruption exceeds the cost of ordinary taxation. We find no evidence that a bilateral income tax treaty with the United States reduces the required rates of return to U.S. FDI. This finding is consistent with theoretical predictions, particularly those of Sinn (1991, 1993), and with the conclusions reached by Blonigen and Davies (2001). We also find that failing to include a variable for the quality of host-country governance can cause a simple cross-section regression to yield the misleading implication that a tax treaty encourages U.S. FDI.

references Altshuler, Rosanne, Harry Grubert and T. Scott Newlon (1998). “Has U.S. investment abroad become more sensitive to tax rates?” NBER Working Paper No. 6383 (Cambridge, MA: NBER). Blonigen, Bruce A., and Ronald B. Davies (2000). “The effects of bilateral tax treaties on U.S. FDI activity,” NBER Working Paper No.7929 (Cambridge, MA: NBER). Blonigen, Bruce A. and Ronald B. Davies (2004). “The effects of bilateral tax treaties on U.S. FDI activity,” International Tax and Public Finance, 11, pp. 601–622. Business International Corporation (1984). Managing and Evaluating Country Risk (New York, NY).

562 henry j. louie and donald j. rousslang Ernst and Young (various years) Doing Business in (various countries). (Ernst and Young International). Grubert, Harry (1998). “Taxes and the division of foreign operating income among royalties, interest, dividends and retained earnings,” Journal of Public Economics, 68, pp. 269–290. Hartman, David (1985). “Tax policy and foreign direct investment,” Journal of Public Economics, 26, pp. 107–121. Hines, James R. Jr. (1995). “Forbidden payments: foreign bribery and American business after 1977,” NBER Working Paper No. 5266 (Cambridge, MA: NBER). Hines, James R. Jr. and Eric M. Rice (1994). “Fiscal paradise: foreign tax havens and American business,” The Quarterly Journal of Economics, 109, pp. 149–182. Horst, Thomas (1971). “The theory of the multinational firm: optimal behavior under different tariff and tax rates,” Journal of Political Economy, 79, pp. 1059–1072. Huntington, Samuel P. (1968). Political Order in Changing Societies (New Haven, CT: Yale University Press). International Bureau of Fiscal Documentation (1978 (2002 update)). Taxes and Investment in Asia and the Pacific. (Amsterdam: IBFD). —— (1987 (2002 update)). Taxation in Latin America. (Amsterdam: IBFD). International Monetary Fund (Various years). International Financial Statistics. (Washington, D.C.: IMF). Leff, Nathaniel (1964). “Economic development through bureaucratic corruption,” American Behavioral Scientist, 8, pp. 8–14. Mauro, Paulo (1995). “Corruption and growth,” The Quarterly Journal of Economics, 110, pp. 681–712. Organisation for Economic Co-operation and Development (2001). International Direct Investment Statistics Yearbook (Paris: OECD). —— (2000). Model Tax Convention on Income and on Capital (Paris: OECD Committee on Fiscal Affairs). Price Waterhouse (various years). Doing Business in (various countries). Price Waterhouse World Firm Services. Rose-Ackerman, Susan (1999). Corruption and Government: Causes, Consequences and Reform (Cambridge: Cambridge University Press). Shleifer, Andrei (2000). “Comment and discussion,” Brookings Papers on Economic Activity, pp. 347–354. Shleifer, Andrei and Robert W. Vishny (1993). “Corruption,” Quarterly Journal of Economics, 108, pp. 599–617. Sinn, Hans-Werner (1991). “The vanishing Harberger triangle,” Journal of Public Economics, 45, pp. 271–300. —— (1991). “Taxation and the birth of foreign subsidiaries,” in H. Herberg and N.V. Long, eds., Trade, Welfare, and Economic Policies: Essays in Honor of Murray C. Kemp (Ann Arbor, MI: University of Michigan Press). Transparency International, http://www.transparancy.org. U.S. Department of Treasury (1998). “U.S. tax treaties,” Internal Revenue Service Publication 901 (Washington, D.C.: U.S. Government Printing Office). United States Government (1993). NAFTA. (Washington, D.C.: U.S. Government Printing Office). Wei, Shang-Jin (2000). “How taxing is corruption on international investors?” Review of Economics and Statistics, 82, pp. 1–11. Wheeler, David and Ashoka Mody (1992). “International investment location decisions: the case of U.S. firms,” Journal of International Economics, 33, pp. 57–76.

21. tax treaties for investment and aid to sub-saharan africa: a case study∗ allison christians introduction The United States is committed to increasing trade and investment to less developed countries (LDCs),1 particularly those in Sub-Saharan Africa, where povertyrelated conditions are extreme and foreign trade and investment minimal.2 This commitment is demonstrated in U.S. efforts to negotiate agreements to eliminate ∗

This chapter was reprinted with permission from the Brooklyn Law Review. The chapter was originally published as “Tax Treaties for Investment and Aid to Sub-Saharan Africa,” 71 Brook. L. Rev. 639 (2005). 1. There is no uniform convention for the designation of a country as “less developed.” The term is generally used to reflect a country’s economic status or growth potential. In the context of taxation, these labels may be used to distinguish “in a general way between countries with highly developed, sophisticated tax systems and those whose tax systems are at an earlier stage of development.” Victor Thuronyi, Tax Law Design and Drafting, xxvii n.1 (1996). In the United States, the Central Intelligence Agency (CIA) delineates three categories in a hierarchy, consisting of 34 “developed countries,” 27 “former USSR/ Eastern Europe[an]” countries, and 172 “less developed countries” (all other recognized countries, including all of Sub-Saharan Africa except South Africa). See Central Intelligence Agency, Office of Public Affairs, The World Factbook 2005 (GPO 2005), (available at http://www.cia.gov/cia/publications/factbook, however, the online version varies in format and content from the print version and is updated frequently, therefore this chapter references the data found in the print version except where otherwise noted) [hereinafter World Factbook] (defining LDCs in Appendix B). As a rough guide to U.S. foreign policy, this chapter incorporates the CIA terms. For a discussion of the arbitrary and often unyielding nature of these designations despite changes in a particular country’s economic status or prospects, see What’s in a name?, Economist, Jan. 17, 2004, at 11. 2. Since the late 1980s, increasing trade with and investment in LDCs has become a preferred means of providing aid to such countries. See, e.g., Paul B. Thompson, The Ethics of Aid and Trade 2 (1992); see also Bruce Zagaris, The Procedural Aspects of U.S. Tax Policy Towards Developing Countries: Too Many Sticks and No Carrots?, 35 Geo. Wash. Int’l L. Rev. 331, 384 (2003) (stating that the “official policies” of the U.S. “are to mobilize private capital rather than foreign aid”). For an overview of poverty conditions and foreign investment in African nations, see, for example, UN Conference on Trade and Development, Foreign Direct Investment in Africa: Performance and Potential 1-2, 21 UN Doc. UNCTAD/ITE/IIT/Misc. 15 (1999) (hereinafter UNCTAD) (stating that foreign investors typically associate Africa with “pictures of civil unrest, starvation, deadly diseases and economic disorder,” and foreign investment “inflows into Africa have increased only modestly” since the 1980s).

564 allison christians

trade barriers such as tariffs and quotas with many of these countries.3 U.S. officials also consistently proclaim a commitment to enter into tax treaties with LDCs,4

3. The main agreement is the African Growth and Opportunity Act (AGOA), a trade preference agreement, discussed infra note 248. The U.S. is also currently negotiating a free trade agreement with the South African Customs Union (consisting of South Africa, Botswana, Lesotho, Namibia, and Swaziland). See United States Trade Representative, Background Information on the U.S.-SACU FTA (2003), available at http://www.ustr.gov/ Trade_Agreements/Bilateral/Southern_Africa_FTA/Background_Information_on_ the_U.S.-SACU_FTA.html. 4. See, e.g., The Japanese Tax Treaty (T. Doc. 108-14) and the Sri Lanka Tax Protocol (T. Doc. 108-9); Hearing Before the Comm. on Foreign Relations, 108th Cong. 29 (2004) (“[w]e are trying to expand the scope of these treaties to developing countries.”); Joseph H. Guttentag, An Overview of International Tax Issues, 50 U. Miami L. Rev. 445, 450 (1996) (“[t] ax treaty expansion in this area is a high Treasury priority.”); Press Release, U.S. Treasury Dep’t, Treasury Welcomes Entry into Force of U.S.–Sri Lanka Income Tax Treaty (July 22, 2004), available at http://www.treas.gov/press/releases/js1809.htm [hereinafter Treasury Press Rel. JS-1809] (“The Treasury Department is committed to continuing to extend and broaden the U.S. tax treaty network, including new agreements with emerging economies”). The U.S. currently has 16 tax treaties with LDCs: Barbados, China, Cyprus, Egypt, India, Indonesia, Jamaica, Korea, Morocco, Pakistan, Philippines, Sri Lanka, Thailand, Trinidad and Tobago, Tunisia, and Venezuela. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.– Barb., Dec. 31, 1984, T.I.A.S. No. 11,090 [hereinafter U.S.–Barbados Treaty]; Agreement for the Avoidance of Double Taxation and the Prevention of Tax Evasion with Respect to Taxes on Income, U.S.–P.R.C., Apr. 30, 1984 T.I.A.S. No. 12065 [hereinafter U.S.–China Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Cyprus, Mar. 19, 1984, 35 U.S.T. 4737 [hereinafter U.S.–Cyprus Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Egypt, Aug. 24, 1980, 33 U.S.T. 1809 [hereinafter U.S.–Egypt Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–India, Sept. 12, 1989, S. Treaty Doc. No. 101-5 (1990) [hereinafter U.S.–India Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Indon., July 11, 1988, T.I.A.S. No. 11593 [hereinafter U.S.–Indonesia Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Jam., May 21, 1980, 33 U.S.T. 2865 [hereinafter U.S.–Jamaica Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and the Encouragement of International Trade and Investment, U.S.–S. Korea, June 4, 1976, 30 U.S.T. 5253 [hereinafter U.S.-Korea Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Morocco, Aug. 1, 1977, 33 U.S.T. 2545 [hereinafter U.S.–Morocco Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Pak., July 1, 1957, 10 U.S.T. 984; Convention with Respect to Taxes on Income, U.S.–Phil., Oct. 1, 1976, 34 U.S.T. 1277 [hereinafter U.S.–Philippines Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Sri Lanka, Mar. 14, 1985, S. Treaty Doc. No. 99-10 (2004) [hereinafter, U.S.–Sri Lanka

tax treaties for investment and aid to sub-saharan africa: a case study 565

on the theory that tax treaties can eliminate excessive taxation and therefore help to increase trade and investment between the partner countries.5 As such, tax treaties appear to be a perfect complement to trade agreements in furthering U.S. efforts to increase trade and investment in LDCs. Yet there are currently no tax treaties in force between the United States and any of the LDCs in Sub-Saharan Africa.6

Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.-Thail., Nov. 26, 1996, S. Treaty Doc. No. 105-2 (1998) [hereinafter U.S.-Thailand Treaty]; Convention for the Avoidance of Double Taxation, the Prevention of Fiscal Evasion with Respect to Taxes on Income, and the Encouragement of International Trade and Investment, U.S.–Trin. & Tobago, Jan. 9, 1970, 22 U.S.T. 164 [hereinafter U.S.– Trin. & Tobago Treaty]; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Tunis., June 17, 1985, as amended by Prot., 29 I.L.M. 1580 [hereinafter U.S.–Tunisia Treaty]; and Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital, U.S.–Venez., Jan. 25, 1999, 38 I.L.M. 1707 [hereinafter U.S.–Venezuela Treaty]. 5. This theory has been officially propounded since the first independent U.S.-LDC treaty was contemplated. See Letter from John F. Dulles to the President (July 9, 1956), in Staff of the Joint Committee on Internal Revenue Taxation, Legislative History of United States Tax Conventions 1445 (1962) (proclaiming that a treaty with Honduras would increase U.S. investment in that country because “[b]y eliminating double taxation . . . [tax treaties] have contributed much to the trade and investment flowing between [partner] countries and the United States”). For a recent restatement of the theory, see The Japanese Tax Treaty (T. Doc. 108-14) and the Sri Lanka Tax Protocol (T. Doc. 108-9): Hearing Before the Comm. on Foreign Relations, 108th Cong. 11 (2004) (statement of Barbara M. Angus, International Tax Council, U.S. Dep’t of Treasury) (in regards to a proposed treaty with Sri Lanka, “[t]he goal of the tax treaty is to increase the amount and efficiency of economic activity” between the partner countries). 6. The U.S. tax treaty network at one time included ten LDCs in Sub-Saharan Africa, pursuant to extensions of existing tax treaties with the United Kingdom and Belgium. Press Release, U.S. Treasury Dep’t, Treasury Dep’t Announces Termination of Extensions of Income Tax Conventions Between the U.S. and the U.K. and the U.S. and Belgium to 18 Countries and Territories (July 1, 1983). All of these treaties were subsequently terminated. Id. Today, the only Sub-Saharan African country with a U.S. tax treaty is South Africa. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital Gains, U.S.–S. Afr., Feb. 17, 1997, S. Treaty Doc. No. 105-9 (1997). The United States considers South Africa to be a developed country. See supra note 1. Ethiopia, Ghana, and Liberia each have a treaty with the U.S. that deals solely with the taxation of income from shipping and aircraft activity. Agreement to Exempt from Income Tax, on a Reciprocal Basis, Income Derived from the International Operation of Aircraft and Ships, U.S.–Eth., Oct. 30–Nov. 12, 1998, State Dept. No. 98-183; Agreement to Exempt from Income Tax, on a Reciprocal Basis, Certain Income Derived from the International Operation of a Ship or Ships and Aircraft, U.S.–Ghana, Nov. 12, 2001, State Dept. No. 02-01; Agreement for Reciprocal Relief from Double Taxation on Earnings from Operation of Ships and Aircraft, U.S.–Liber., July 1–Aug. 11, 1982 34 U.S.T. 1553.

566 allison christians

The lack of tax treaties between the United States and the LDCs of SubSaharan Africa cannot be explained by disinterest or lack of support on the part of academics, practitioners, or lawmakers—representatives from all of these sectors have urged the importance of entering into these agreements.7 Neither can the omission be attributed to disinterest on the part of the LDCs in SubSaharan Africa themselves.8 Many of these nations have long pursued tax treaties with the United States,9 and a few have gone so far as to formally and publicly express their interest in commencing negotiations with the United States.10 Nevertheless, these agreements are largely unnecessary due to parallel provisions in U.S. domestic tax law. See I.R.C. § 883 (2005). 7. See, e.g., The U.S.–Africa Partnership: Hearing Before the S. Comm. on Foreign Relations, 108th Cong. (2004) (statement of Sen. Lugar) (supporting legislation that “directs the Secretary of the Treasury to seek negotiations regarding tax treaties with [AGOA] eligible countries”); Statement by Michael A. Samuels to the Subcomm. on Trade Comm. on Ways and Means (July 16, 1996), available at 1996 WL 433282 (F.D.C.H.) (“Given the vital role that investment must play in the development of African countries . . . the new policy must contain several key investment-related priorities, including an emphasis on bilateral investment and tax treaties.”); Hon. Charles B. Rangel, The State of Africa: The Benefits of The African Growth and Opportunity Act – Next Steps (July 14, 2003), available at 149 Cong. Rec. E1464-01 (stating that the U.S. should negotiate tax treaties with AGOA countries); Calvin J. Allen, United States Should Expand Tax Treaty Network in Sub-Saharan Africa, 34 Tax Notes Int’l 57, 58 (2004) (same); Karen B. Brown, Missing Africa: Should U.S. International Tax Rules Accommodate Investment in Developing Countries?, 23 U. Pa. J. Int’l Econ. L. 45, 46, 69 (2002) (arguing for multilateral tax treaties between the U.S. and countries in Sub-Saharan Africa). 8. All of the LDCs in Sub-Saharan Africa are in urgent (if not desperate) need for foreign capital, and most are responding to the need by implementing measures to make their countries more attractive to foreign investors. See, e.g., James Gathii, A Critical Appraisal of the NEPAD Agenda in Light of Africa’s Place in the World Trade Regime in an Era of Market Centered Development, 13 Transnat’l L. & Contemp. Probs. 179 (2003). Given the powers ascribed to tax treaties in increasing trade and investment between partner countries, most LDCs would pursue the opportunity to commence negotiations with the U.S. (provided that the concessions required to secure such agreements are not too great). 9. For example, Nigeria began pursuing a tax treaty with the United States in 1978, after Nigeria unilaterally withdrew from its coverage under an extension of the 1945 tax treaty between the United Kingdom. and the United States (as a former U.K. territory). Nigeria to Terminate Tax Treaty with U.S., Seek Renegotiated One, Wall St. J., Aug. 24, 1978, at 23. See supra note 6 (discussing the treaty extension); I.R.S. Announcement 78-147, 1978-41 I.R.B. 20 (Oct. 10, 1978) (terminating the treaty). Although the tax treaty was apparently negotiated at length, it was never completed. 10. Calvin J. Allen, Botswana, Burundi Wish to Negotiate Tax Treaties with United States, 26 Tax Notes Int’l 1264 (2002). This announcement is a rather unusual event, since tax treaties are generally commenced and negotiated in secret. Their existence is usually made public after negotiations have concluded and the treaty has been signed by the respective countries, pending ratification. Richard E. Andersen & Peter H. Blessing, Analysis of United States Income Tax Treaties ¶ 1.04[1][a][i], [ii] (2005). Thus, countries

tax treaties for investment and aid to sub-saharan africa: a case study 567

Finally, the lack of tax treaties cannot be charged to a lack of commitment on the part of the United States to conclude agreements that will increase trade and investment to assist in the economic growth of the countries of Sub-Saharan Africa.11 The United States has demonstrated its commitment by making significant concessions in the context of trade and aid agreements, in the form of direct aid as well as reduced tariffs.12 That the lack of tax treaties cannot be explained by a lack of support or commitment on the part of scholars, policymakers, or governments suggests that there must be some other reason or reasons that tax treaties have not been concluded between the United States and the LDCs of Sub-Saharan Africa. This chapter explores many of these reasons by presenting as a case study a hypothetical tax treaty between the United States and Ghana, one of the LDCs of SubSaharan Africa.13 Hypothesizing the structure and operation of a tax treaty between these two countries provides a vehicle for measuring the potential effect of such a treaty on international commerce. While there has been some discussion among scholars and policymakers regarding the paucity and inefficacy of tax treaties between the United States and LDCs, much of the discussion has focused on abstract principles of international tax law. By examining the effects a U.S. treaty with Ghana might have on investors, this chapter analyzes these legal principles in the context of current global tax conditions for investment in LDCs. This case study demonstrates that in today’s global tax climate, a typical tax treaty would not provide significant tax benefits to current or potential investors. Consequently, there is little incentive for these investors to pressure the U.S. government to conclude tax treaties with many LDCs.

don’t usually issue public proclamations regarding their desire to enter into tax treaties. Similarly, since there is little public disclosure regarding progress in treaty-making by the U.S. Treasury Department, there is little means to determine the reaction, if any, that the Treasury has had to these or other requests to initiate negotiations. 11. See Richard Mitchell, United States–Brazil Bilateral Income Tax Treaty Negotiations, 21 Hastings Int’l & Comp. L. Rev. 209, 225 (1997) (“[t]he United States displays eagerness to enter into tax treaties with developing nations”); Miranda Stewart, Global Trajectories of Tax Reform: The Discourse of Tax Reform in Developing and Transition Countries, 44 Harv. Int’l L. J. 139, 148–149 (2003) (pointing to the number of U.S. treaties with other emerging economies as evidence that “the lack of U.S. treaty-making with [LDCs in Sub-Saharan Africa] cannot be explained solely by a general reluctance to enter into tax treaties with less developed or non-capitalist countries”). 12. The main agreements are the African Growth and Opportunity Act (AGOA), a preferential trade regime, discussed infra note 248, and the recently introduced Millennium Challenge Act, an aid package tied to countries’ demonstrated commitment to growth through investment and trade, discussed infra note 249. 13. Ghana was chosen as a subject for this case study for several reasons, including its existing commercial ties to the U.S. These reasons are described infra Part III.A.

568 allison christians

There are, of course, any number of other reasons why tax treaties may not be concluded between the United States and the LDCs of Sub-Saharan Africa, including competing priorities for the U.S. government, either for tax treaties with other countries or for other domestic or international tax matters. Undoubtedly, socio-political factors play an important role as well.14 However, this chapter argues that since tax treaties with LDCs like Ghana would not provide major tax benefits to the private sector, even if concluded, these treaties would not have a significant impact on cross-border investment and trade. Accordingly, the main justification so consistently proclaimed to support the pursuit of tax treaties between the United States and LDCs is misguided. If the United States is truly committed to increasing investment and trade to the LDCs of Sub-Saharan Africa, an examination of how the global tax climate has changed since tax treaties were first implemented is in order. We must acknowledge that tax treaties cannot deliver the promised benefits, and we must examine the factors that prevent them from so doing. An overview of the background and function of tax treaties and their proclaimed benefits is discussed in Section A of this chapter. Section B presents the case study of a hypothetical tax treaty between the United States and Ghana. It shows that such a treaty would produce few tax benefits to current or potential investors and would therefore be largely ineffective in stimulating trade and investment between these two countries. Section C concludes that after decades of adherence to the promise of tax treaties, we must acknowledge their failure to deliver and search for alternative ways to achieve the goal of promoting aid through the vehicles of investment and trade. A. Background: Tax Treaties, Investment, and Trade This section provides the context for a discussion of the role of tax treaties in delivering investment and aid to LDCs. Section A.1 describes some of the strategies employed by the United States to assist LDCs, and how tax treaties comport with these strategies. Section A.2 explains the role tax treaties play as the locus of international tax law by outlining the purposes and goals surrounding the origin and evolution of these agreements. Section A.3 discusses the limitations that arise because international tax law concepts are embodied in a network of overlapping, varying, and mostly bilateral agreements between select nations.

14. For example, there may be national interests at stake, such as security, defense, or energy supply issues, that may contribute to the prioritization of concluding tax treaties with LDCs in other areas of the world, such as Sri Lanka (concluded in 2004) and Bangladesh (currently pending ratification). See John Venuti et al., Current Status of U.S. Tax Treaties and International Agreements, 34 Tax Mgmt Int’l J. 653 (2005) (updating on a monthly basis the status of current U.S. tax treaties and international agreements). The various foreign policy goals that motivate the agenda for treaty-making is a subject that deserves much attention, but is beyond the scope of this chapter.

tax treaties for investment and aid to sub-saharan africa: a case study 569

This section introduces some of the problems faced by the LDCs of Sub-Saharan Africa, which operate largely outside of this network. 1. United States strategy for assistance to LDCs The United States has adopted a foreign aid strategy towards Sub-Saharan Africa that centers on the idea that creating investment and trade opportunities for LDCs will most effectively boost economic growth in these countries, thereby lifting them out of poverty through commercial interaction with the global community.15 A key component of this foreign aid strategy is the identification and elimination of barriers to trade and investment. Among the most significant potential barriers are double taxation, which occurs when two countries impose similar taxes on the same taxpayer in respect to the same item of income, regulatory barriers, such as currency exchange and other market controls, and tariffs. These barriers have historically been addressed in very different ways. Regulatory barriers and tariffs have been addressed by most countries in a generally uniform manner through regional and global trade agreements.16 The main multilateral agreement is the General Agreement on Tariffs and Trade (GATT), to which 147 countries are signatories through the World Trade Organization (WTO).17 Additional tariff and regulatory barrier reduction between the United States and Sub-Saharan Africa has been accomplished through the African Growth and Opportunity Act (AGOA), an agreement that seeks to increase growth and alleviate poverty through the elimination of tariffs and quotas for selected imports from designated Sub-Saharan African nations.18

15. See supra note 2. 16. Regulatory barriers are also addressed, to a lesser extent, in bilateral investment treaties (BITs), as discussed infra Part IV.E. 17. 37 of the 47 LDCs in Sub-Saharan Africa are members of the WTO. See the GATT agreement and accompanying agreements, available at Understanding the World Trade Organization, http://www.wto.org/english/thewto_e/whatis_e/tif_e/org6_e.htm (last visited Jan. 24, 2006). 18. The Trade and Development Act of 2000, Pub. L. No. 106-200, 114 Stat. 251. (codified as amended at 19 U.S.C. § 3721 (2000)) (more commonly known as the African Growth and Opportunity Act (AGOA I)), amended by Trade Act of 2002 Pub. L. No. 107210 § 3108, 116 Stat. 933, 1038–1040 (AGOA II), amended by AGOA Acceleration Act of 2004, Pub. L. No. 108–274, 118 Stat. 820 (AGOA III) [hereinafter referring to the three acts collectively as AGOA]. AGOA eliminates “competitive need limitations” (quotas) and tariffs on over 1,800 items that would otherwise be subject to such constraints under the Harmonized Tariff Schedule of the United States (unless otherwise exempt under a free trade agreement). See AGOA § 103(2) (2004). See also United States International Trade Commission, Harmonized Tariff Schedule of the United States 67–68 (2004), available at http://www.usitc.gov; AGOA Faq, http://www.agoa.gov/faq/faq.html (last visited Oct. 9, 2005). AGOA is a preferential trade regime, rather than a free trade agreement. For a discussion of AGOA and other trade and aid initiatives entered into with Sub-Saharan Africa over the past several years up to 2002, see Brown, supra note 7, at 49–51. As of March, 2005, 36 of the 47 LDCs in Sub-Saharan Africa were eligible for AGOA benefits.

570 allison christians

Another barrier reduction device is the Millennium Challenge Act of 2003 (MCA), a new official direct assistance initiative that will direct foreign aid only to countries demonstrating a commitment to poverty reduction through economic growth.19 According to the Organization for Economic Cooperation and Development (OECD),20 the harmful effects of double taxation on cross-border trade and investment “are so well known that it is scarcely necessary to stress the importance of removing the obstacles that double taxation presents to the development For a list of currently-eligible countries, see the AGOA, Country Eligibility, http://www. agoa.gov/eligibility/ country_eligibility.html (last visited Oct. 9, 2005). 19. Millennium Challenge Act of 2003, Pub. L. No. 108-199, 118 Stat. 3, 211–226; see S. Res. 571, 108th Cong. (2003) (enacted). For information on the MCA, see the MCA website, http://www.mca.gov; see also Colin Powell, Welcome Message from the Honorable Colin L. Powell, http://www.mca.gov (last visited Oct. 9, 2005) (stating that the MCA “reflects a new international consensus that development aid produces the best results when it goes to countries that adopt pro-growth strategies for meeting political, social and economic challenges”). The MCA is not solely directed at Sub-Saharan Africa, but approximately half of its currently identified recipient countries are located in this region. See Millennium Challenge Corporation: Eligible Countries, available at http://www.mca.gov/ countries/eligible/index.shtml (last visited Nov. 3, 2005). 20. The OECD is an international organization consisting of thirty member countries, all of which are considered by the United States to be developed countries (with the nominal exception of South Korea, which the United States specifically classifies as less developed, even though all OECD countries are deemed to be developed under the “developed country” listing). World Factbook, supra note 1, at 628, 639, 641. It is perhaps worthy of note that other organizations, such as the IMF, diverge from the views of the United States in some of these classifications. For instance, the IMF classification of “developing countries,” while similar in most respects to the United States’ classification of LDCs, diverges by including in its list both Mexico and Turkey. Id. at 628. Mexico is not independently listed as a developed country under the United States’ classification system, and the Czech Republic, Hungary, and Slovakia are separately categorized as “former USSR/Eastern European” countries, but the term developed countries is defined as including all of the OECD member countries. See OECD, Ratification of the Convention on the OECD and OECD Member Countries, http://www.oecd.org/document/58/0,2340, en_2649_201185_1889402_1_1_1_1,00.html; World Factbook, supra note 1, at 628, 639, 641 (Appendix B provides a listing of LDCs that includes the “Four Dragons,” a group that includes South Korea. Country data on South Korea provides GDP and poverty statistics, available at http://www.cia.gov/cia/publications/factbook/geos/ks. html). The CIA still considers South Korea an LDC despite its 2004 estimated per capita GDP of $19,200, well over the typical $10,000 threshold separating developed from less-developed, and despite the fact that just 4% of the population is considered to be living in poverty conditions. World Factbook, supra note 1, at 304, 628, 639. In contrast, the inclusion of Mexico as a developed country is anomalous, given its per capita GDP of $9,600. Id. at 365, 628. As discussed below, the OECD developed and continually updates a model income tax convention that both encapsulates and sets international tax standards.

tax treaties for investment and aid to sub-saharan africa: a case study 571

of economic relations between countries.”21 The U.S. Government mirrors this sentiment, identifying the eradication of “tax barriers” as a major component of its dedication “to eliminating unnecessary barriers to cross-border trade and investment.”22 Yet, unlike other barriers to trade and investment, double taxation has not been reconciled on a global scale. Instead of a world tax organization to coordinate efforts and resolve disputes,23 relieving double taxation remains the purview of individual countries.24 Nevertheless, a consensus has emerged regarding the 21. OECD Committee on Fiscal Affairs, Model Tax Convention on Income and on Capital 7 (2005) (hereinafter OECD Model). 22. Treasury Press Release JS-1809, supra note 4 (“This new tax treaty relationship will serve to eliminate tax barriers to cross-border trade and investment between the two countries… [by providing] greater certainty to taxpayers with respect to the tax treatment of their cross-border activities and [reducing] the potential for double taxation of income from such activities.”); Press Release, U.S. Treasury Dep’t, Remarks by Treasury Secretary John W. Snow at the Signing Ceremony for the U.S.-Barbados Income Tax Protocol (July 14, 2004), available at http://www.treas.gov/press/releases/js1786.htm [hereinafter Treasury Press Rel. JS-1786]; see also Treasury Press Release JS-1267, Treasury Welcomes Senate Approval of New U.S.-Sri Lanka Tax Treaty (March 26, 2004), available at http:// www. treas.gov/press/releases/js1267.htm (hereinafter Treasury Press Rel. JS-1267) (stating that the new treaty with Sri Lanka is “an important step in our ongoing efforts to broaden the reach of our tax treaty network”). 23. See What is the WTO?, http://www.wto.org/english/thewto_e/whatis_e/whatis_e. htm (last visited Oct. 9, 2005) (noting that the WTO is “the only international organization dealing with the rules of trade between nations”). The several international organizations concerned with standardizing and coordinating global taxation do not approach the level of member country participation in the WTO. For example, the OECD is one of the primary international organizations that concerns itself with setting standards for international taxation, but it has only 30 members, few new members are added (the latest addition was the Slovak Republic, in 2000), and many countries with rapidly growing economies, such as Brazil, Russia, India, and China, are not members. OECD, Ratification of the Convention on the OECD and OECD Member Countries, http://www.oecd.org/ document/58/0,2340,en_2649_201185_1889402_1_1_1_1,00.html (last visited Jan. 12, 2006). 24. The vast majority of international agreements that address the problem of double taxation are bilateral. See, e.g., Reuven Avi-Yonah, International Tax as International Law, 57 Tax L. Rev. 483, 497 (2004) (noting that there are over 2,000 bilateral tax treaties). However, there are a few regional multilateral tax treaties currently in force, including the Andean Pact Income Tax Convention between Bolivia, Colombia, Ecuador, Peru, and Venezuela (November 16, 1971); the Arab Economic Unity Council Tax Treaty between Egypt, Iraq, Jordan, Kuwait, Sudan, Syria, and Yemen (Y.A.R.) (December 3, 1973); the Agreement Among the Governments of the Member States of the Caribbean Community for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion With Respect to Taxes on Income, Profits, or Gains and Capital Gains and for the Encouragement of Regional Trade and Investment between the Caribbean Community (CARICOM) countries of Antigua, Belize, Dominica, Grenada, Guyana, Jamaica, Montserrat, St. Christopher and Nevis, St. Lucia, St. Vincent and the Grenadines, and Trinidad and

572 allison christians

appropriate tax treatment of cross-border investment activity.25 Under this consensus, double taxation is addressed primarily by tax treaties, which allocate tax revenue among jurisdictions based on concepts of residence and source.26 Thus, the United States, along with the rest of the developed world, has a network of tax treaties, spanning most of its major trading partners across the globe.27 Expanding the tax treaty network has been termed by the Treasury Department as a commitment, an ongoing effort,28 and the “primary means” for the elimination of tax barriers to international trade and investment.29 Officials from other countries echo these sentiments.30 From the perspective of LDCs, a major problem with embodying international tax laws aimed at preventing double taxation in tax treaties is that LDCs typically have few of these treaties in place. But the tax treaty network, with its central role in the evolution of international tax law, directly affects these countries regardless of their level of inclusion. To demonstrate the extent of this influence, the following section discusses why and how tax treaties became the source of international tax law, and explores how this international tax system has impacted tax treaties between the United States and the LDCs of Sub-Saharan Africa.

Tobago (July 6, 1994); the Tax Convention Between the Member States of the West African Economic Community (C.E.A.O.) between Burkina Faso, Côte D’Ivoire, Mali, Mauritania, Niger, and Senegal (October 29, 1984); the Agreement on the Avoidance of Double Taxation on Personal Income and Property, signed by Bulgaria, Czechoslovakia, Germany (G.D.R.), Hungary, Mongolia, Poland, Romania, and the Soviet Union (still in force with respect to various successor states) (May 27, 1977); and the Convention Between the Nordic Countries for the Avoidance of Double Taxation With Respect to Taxes on Income and on Capital, between Denmark, the Faeroe Islands, Finland, Iceland, Norway, and Sweden (September 23, 1996) (generally based on the OECD Model). 25. See generally Reuven S. Avi-Yonah, The Structure of International Taxation: A Proposal For Simplification, 74 Tex. L. Rev. 1301 (1996) (providing an introduction to the evolution of international taxation from decisions and compromises made by the U.S. and the League of Nations in the 1920s to the “flawed miracle” that exists today). 26. See id. at 1306. 27. See discussion infra notes 96 and 97. 28. Treasury Press Release JS-1267, supra note 22. 29. Treasury Press Release JS-1809, supra note 4; Treasury Press Release JS-1786, supra note 22; see also Letter from Gregory F. Jenner thanking Sen. Susan M. Collins for her Comments on a Possible Chile–U.S. Tax Treaty, U.S. Treasury Thanks Senator for Comments on Possible Chile–U.S. Tax Treaty (Apr. 22, 2004), 2004 WTD 83-16 (“Income tax treaties can serve the important purpose of addressing tax-related barriers to cross-border trade and investment”). 30. For example, Bangladeshi officials assert that when the new treaty between the U.S. and Bangladesh enters into force, it “will encourage U.S. investment in the education, highway, and communication sectors in Bangladesh.” U.S. Treaty Update, PwC In & Out, 15 J. Int’l Tax’n 4–5 (Dec. 2004).

tax treaties for investment and aid to sub-saharan africa: a case study 573

2. Origins of tax treaties as international law Every country establishes its jurisdiction to impose income taxation under sovereign claim of right. In the United States, the taxation of income from international transactions turns on whether the income is earned by a resident31 or a nonresident.32 In the case of residents, the United States purports to tax “all income from whatever source derived.”33 In the case of nonresidents, the U.S. taxing jurisdiction is generally limited to income derived from investments and business activities carried out in the United States (known as source-based taxation).34 Most developed countries similarly impose worldwide, or residence-based, income taxation on residents, and source-based taxation on income earned within their borders.35 As a result, ample potential exists for double taxation of international transactions involving two developed countries.36 Therefore, the United States and most of

31. Whether individual or entity. See I.R.C. § 7701(a), (b) (2005). 32. I.R.C. § 7701(a), (b) (2005). 33. I.R.C. § 61(a) (2005) (“gross income means all income from whatever source derived”); see also I.R.C. §§ 1, 11(a) (2005) (imposing tax on incomes of individuals and corporations, respectively); Treas. Reg. §§ 1.1-1(b), 1.11-1(a). The authority to extend its jurisdiction in this broad fashion is confirmed by Cook v. Tait. 265 U.S. 47, 56 (1924): The basis of the power to tax was not and cannot be made dependent upon the situs of the property in all cases, it being in or out of the United States, nor was not and cannot be made dependent upon the domicile of the citizen, that being in or out of the United States, but upon his relation as citizen to the United States and the relation of the latter to him as citizen. Id. 34. I.R.C. §§ 871, 881–882 (2005). 35. OECD countries generally impose some form of worldwide taxation, although a few (Australia, Austria, and Switzerland) provide certain statutory exemptions, and many provide for exemption under treaty, as discussed below. See Ernst & Young, Worldwide Corporate Tax Guide 29–53, 894–910 (2005), available at http://www.ey.com/global/ download.nsf/Ireland/WorldWideCorporateTaxGuide/$file/WW_Corporate_Tax_ guide_2005_.pdf (describing the tax systems of, and treaty benefits provided by Australia, Austria, and Switzerland, respectively). Some countries such as France are generally source-based, or territorial systems, which generally refrain from taxing the foreign income earned by their residents. See id. at 240–52. Nevertheless, these countries enforce worldwide taxation of certain kinds of income earned in low-tax jurisdictions in order to prevent capital flight. Id. Thus, France imposes worldwide taxation on certain low-taxed foreign income. See generally id. (providing tax system features and rates). 36. The most common form of double taxation occurs when there is a residence-source overlap, as a taxpayer’s country of residence (the home country) imposes residence-based tax on income earned in a foreign (source or host) country, while the host country imposes source-based tax on the same item. Overlaps can also occur when countries have overlapping or conflicting rules for determining the source of an item of income or the residence of a taxpayer. For example, while the United States assigns corporate residence according to country of incorporation, the United Kingdom assigns corporate residence according to the seat of management and control. See I.R.C. § 7701(a)(4), (5) (2005) (assigning corporate residence to country of incorporation).

574 allison christians

the other countries that impose worldwide taxation provide a foreign tax credit,37 which essentially removes the residence-based layer of tax while preserving the source-based layer. Thus, the United States and most other countries imposing worldwide income taxation generally relieve double taxation on a unilateral basis under statutory law. The same result is attained under treaties. Tax treaties are contracts, generally between two countries,38 under which the signatory countries agree to the taxation each will impose on the activities carried out between their respective jurisdictions.39 Because the United States unilaterally provides a mechanism to prevent U.S. taxation in the event foreign taxation applies, treaties aimed at relieving double taxation would appear to be duplicative.40 Treaties might seem unnecessary ab initio, because the United States provided the foreign tax credit mechanism almost immediately following the inception of the income tax itself, decades before any tax treaties were ever negotiated.41 Nevertheless, the

37. See generally Ernst & Young, supra note 35. 38. But see supra note 24 (noting that some treaties are multilateral). 39. In the U.S., treaties have the same effect as acts of Congress, and are equivalent to any other U.S. law. U.S. Const. art. VI, cl. 2; see American Trust Co. v. Smyth, 247 F.2d 149, 152 (9th Cir. 1957); Samann v. Comm’r, 313 F.2d 461, 463 (4th Cir. 1963). As such, they are subject to and may be overridden by subsequent revisions in domestic law (“treaty override”) under the “last in time” rule of I.R.C. § 7852(d) (2005). See Cherokee Tobacco, 78 U.S. (11 Wall.) 616, 621 (1871) (“a treaty may supersede a prior act of Congress, and an act of Congress may supersede a prior treaty.”); Edye v. Robertson, 112 U.S. 580, 597–600 (1884) (“A treaty, then, is a law of the land as an act of Congress is . . . [so a] court resorts to the treaty for a rule of decision for the case before it as it would to a statute.”); Whitney v. Robertson, 124 U.S. 190, 194 (1888) (“a treaty is placed on the same footing, and made of like obligation. . . . If the two are inconsistent, the last one in date will control the other”); see also Philip F. Postlewaite & David S. Makarski, The ALI Tax Treaty Study – A Critique and a Modest Proposal, 52 Tax Law. 731, 740 (1999) (arguing that treaty override is seen as a “serious problem” because it potentially places the U.S. in violation of existing international obligations); Richard L. Doernberg, Overriding Tax Treaties: The U.S. Perspective, 9 Emory Int’l L. Rev. 71, 131 (1995) (discussing treaty override in the U.S. and concluding that “these provisions embody an important contractual principle”: that breach of an obligation is desirable when “what is gained from the party that breaches exceeds what is lost by the party against whom the breach occurred,” thus a breach might be appropriate as long as the United States compensates the aggrieved party). 40. See generally Elisabeth Owens, United States Income Tax Treaties: Their Role in Relieving Double Taxation, 17 Rutgers L. Rev. 428 (1963) (arguing that treaties play a relatively small role in relieving double taxation, owing to the U.S. foreign tax credit); see also Tsilly Dagan, The Tax Treaties Myth, 32 N.Y.U. J. Int’l L. & Pol. 939 (2000) (showing that tax treaties are not needed to relieve double taxation, since each country would find it in its own best interest to unilaterally relieve double taxation on its citizens and residents). 41. After a brief and limited stint during the Civil War, the income tax was re-introduced in 1913. See Steven R. Weisman, The Great Tax Wars 5, 278 (2002). The foreign tax credit was enacted quickly thereafter, in 1918. See Revenue Act of 1918, ch. 18, §§ 222(a)(1),

tax treaties for investment and aid to sub-saharan africa: a case study 575

United States began entering into tax treaties in 1932 and the practice continues to the present.42 One of the original reasons to enter into treaties was that before they existed, there was no international standard for relieving double taxation: the United States was alone in providing a comprehensive foreign tax credit that unilaterally relieved residence-based taxation.43 The provision of unilateral relief of double taxation was seen as a “present of revenue to other countries,” for which the possibility of source-based taxation was preserved.44 Other European nations, especially Italy and France, relied heavily on source-based taxation and therefore vigorously defended the U.S. position of ceding residence-based taxation to that of source.45 In stark contrast, Britain imposed worldwide taxation and provided a foreign tax credit that was extremely limited and generally preserved its residence-based taxation.46

238(a), 240(c), Pub. L. No. 65-254, 40 Stat. 1057, 1073, 1080–82 (1919). Section 222(a)(1) was applicable to individuals, 238(a) to corporations, and 240(c) defined the taxes for which credit would be allowed. 42. The first U.S. tax treaty was signed with France in 1932 and entered into force on April 9, 1935. Convention on Double Taxation, U.S.-Fr., Apr. 27, 1932 S. Exec. K, 72-1, (1935). Since then, the U.S. tax treaty network has grown by an average of one treaty per year, based on the entry-in-force dates of all U.S. tax treaties ever entered into force. The most recent treaty to enter into force is with Sri Lanka. See U.S.–Sri Lanka Treaty, supra note 4 (entered into force Jun. 13, 2004). The most recently signed is with Bangladesh, which was signed on September 26, 2004, but as of the time of publication has not yet entered into force. See Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion With Respect to Taxes On Income, U.S.-Bangl., Sept. 26, 2004, S. Treaty Doc. No. 109-5; see also Muhammad Kibria, Bangladesh, United States Sign Tax Treaty, 2004 WTD 188-3 (Sept. 28, 2004). 43. See H. David Rosenbloom & Stanley I. Langbein, United States Tax Treaty Policy: An Overview, 19 Colum. J. Transnat’l. L. 359 (1981); Michael J. Graetz & Michael M. O’Hear, The “Original Intent” of U.S. International Taxation, 46 Duke L. J. 1021, 1023 (1996). 44. Edwin R. A. Seligman, Double Taxation and International Fiscal Cooperation, 132, 135 (1928). Source-based taxation was even enhanced to the extent the foreign country’s tax rates were lower than that of the U.S. In such cases, foreign countries could raise their tax rates to the U.S. level with the assurance that these taxes would be creditable in the U.S., leaving the investor indifferent as to the higher foreign rate. See Richard E. Caves, Multinational Enterprise and Economic Analysis 190 (1996) (“Neutrality depends on who pays what tax, not which government collects it”). 45. Graetz & O’Hear, supra note 43, at 1072. 46. Britain’s view was supported by the Netherlands. See Ernst & Young, supra note 35, at 631–32. Both countries were primarily capital-exporting nations, and thus the importance of preserving residence-based taxation was high. The U.S. was also a capital-exporting nation at the time, but arguably regarded source-based taxation as having the superior claim. Graetz & O’Hear, supra note 43, at 1046.

576 allison christians

The conflicting views of the United States and the United Kingdom regarding the proper method for relieving double taxation prompted several years of debate out of which a consensus emerged in the early 1920s.47 Under this consensus, “personal taxation” was to be preserved for residence-based taxation, and “impersonal taxation” was to be preserved for source.48 How these terms would be defined and implemented in the context of the then vastly differing tax systems depended on long and contentious negotiations, held under the auspices of the League of Nations, in which the United States played a large part.49 Ultimately, the League of Nations promulgated a model tax treaty under which countries would reciprocally restrict source-based taxation of passive income items, such as dividends and interest, in favor of preserving residence jurisdiction over these items,50 and reciprocally relieve residence-based taxation on foreign-source business income, as had been done unilaterally by the United States through the foreign tax credit.51 By subsequently entering into tax treaties following the League of Nations model, the United States retreated from its position of unilaterally providing foreign tax credits. The tax concessions thereby obtained from treaty partners reduced the revenue cost of the foreign tax credit,

47. Discussions began in the newly formed International Chamber of Commerce (ICC) in 1920. In 1921 the ICC adopted a resolution that taxing jurisdiction turned on the nature of the tax, with distinctions being made between “super” and “normal” taxes. The U.S., however, rejected this resolution and endorsed closer adherence to the U.S. system, with exceptions made for particular kinds of income, including that from international shipping (as to which residence-based taxation was to be preserved) and that from sales of manufactured goods (to be apportioned under formula). The ICC synthesized the views of the U.S. and 14 other countries and produced a new resolution in Rome, in 1923. The League of Nations began to take over the discussions in 1923, using the Rome resolutions as a basis for discussion. The compromise of the ICC as to “super” and “normal” taxes resurfaced in League of Nations discussions. See id., at 1067–70; Mitchell B. Carroll, International Tax Law: Benefits for American Investors and Enterprises Abroad, 2 Int’l Law. 692, 696 (1968). 48. Graetz & O’Hear, supra note 43, at 1080. 49. See Carroll, supra note 47, at 693, 698 (stating that in the early 1920s the U.S. had been invited by the League of Nations to participate in forming tax treaty policy, but the Department of State had not responded because of the Senate’s rejection of membership in the League (by virtue of its failure to consent to ratification of the Treaty of Versailles). Nevertheless, interest in tax treaties grew in the U.S. and the League planned subsequent Committee meetings to “facilitate attendance by Americans”). 50. Graetz & O’Hear, supra note 43, at 1086–87 (citing Britain’s strong role in producing this result); Avi-Yonah (1996), supra note 25, at 1306. 51. See Graetz & O’Hear, supra note 43, at 1023. The League of Nations first produced a model treaty in 1927. See Carroll, supra note 47, at 698.

tax treaties for investment and aid to sub-saharan africa: a case study 577

which had been the main goal of U.S. involvement in first negotiating these instruments.52 The concepts embodied in the League of Nations model treaty evolved into a model treaty developed by the OECD in 1963, which has been updated periodically since then (the OECD Model).53 The OECD Model has become the standard upon which most of the over 2,000 tax treaties currently in force are based.54 Following the League of Nations and OECD standards, tax treaties minimize source-based taxation of income derived from passive investment activity, such as dividends, interest, and royalties, while preserving residence-based taxation of these items. Once activities increase to a sufficiently significant level of engagement, however, source-based jurisdiction again takes precedence.55 As a member of the OECD, the United States participated in the development of the OECD model, but also developed its own model to reflect specific policies (the U.S. Model).56 First published in 1977 and most recently updated in 1996, the U.S. Model is based on the OECD Model in most respects.57 One notable difference between the models, however, is that the OECD Model allows for the alleviation of double taxation either via a foreign tax credit or by providing that the residence country will exempt the income earned in the source country

52. See Carroll, supra note 47, at 693–94 (interest in pursuing tax treaties grew because these instruments “would reduce the amount of foreign taxes that could be credited against the United States tax . . . and possibly leave something for the Treasury to collect”). 53. The OECD Model was itself based on a series of model treaties promulgated by the League of Nations. It has since been updated several times to cope with the changing nature of business, culminating with the most recent update on February 1, 2005. Unless otherwise noted, references in this chapter to the OECD Model refer to the 2005 version, which is available at http://www.oecd.org. See OECD Model, supra note 21. 54. Compiled in February 2005 from Ernst & Young, supra note 35, and the LexisNexis Tax Analysts Worldwide Tax Treaties database, http://w3.nexis.com/sources/scripts/info. pl?250064 (last visited Nov. 11, 2005). 55. The required level of engagement is defined as a “permanent establishment” as discussed infra Part II.C.2. 56. See U.S. Model Income Tax Convention of Sept. 20, 1996, available at http://www. treas.gov/offices/tax-policy/library/model996.pdf (hereinafter U.S. Model); U.S. Model Income Tax Convention of Sept. 20, 1996: Technical Explanation, available at http:// www.irs.gov/pub/irstrty/usmtech.pdf. 57. The Joint Committee on Taxation compares provisions of both the U.S. and OECD models when analyzing and describing new tax treaties entered into by the U.S. See, e.g., George Yin, Chief of Staff, Joint Comm. on Taxation, Testimony of the Staff of the Joint Committee on Taxation before the Senate Comm. on Foreign Relations Hearing on the Proposed Tax Treaties with Japan and Sri Lanka (Feb. 25, 2004) (explaining the use of the U.S. and OECD models in treaty negotiations and describing ways in which the new Japan–U.S. Treaty deviates from each model), available at http://www.house.gov/jct/ x-13-04.pdf.

578 allison christians

(known as the exemption method).58 The U.S. Model, in keeping with its historical preference to impose worldwide taxation and alleviate double taxation via the foreign tax credit mechanism, allows only the credit method.59 All modern U.S. tax treaties are based on the U.S. Model, with modifications made to reflect changes in law or policy since the release of the latest model.60 The consensus forged through the original model treaties has remained constant: in general, residence-based, or worldwide taxation is accorded primary status in the case of dueling tax jurisdictions, with treaties serving to set the limited boundaries within which source-based taxation will continue to take precedence. Residence-based, or worldwide income taxation is typically justified on the grounds that it promotes capital export neutrality, an efficiency principle dictating that taxpayers will not differentiate on tax grounds between locating activities domestically or abroad on tax grounds, since in either case the income generally will be subject to tax at the same rate.61 Thus, if taxation is imposed by a source country, the United States as home country generally provides the foreign tax credit against the U.S. tax imposed on the same item of income, leaving the 58. OECD Model, supra note 21, arts. 23A (exemption method), 23B (credit method). For example, among OECD countries, Belgium, Denmark, Finland, Germany, and Poland have treaties in which they completely relinquish their residual taxation of income derived by a permanent establishment. See generally Ernst & Young, supra note 35. For a recent example, see the Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and on Capital, Belg.-Ecuador, Dec. 18, 1996, 2248 U.N.T.S. 676 (entered into force Mar. 18, 2004). 59. See U.S. Model, supra note 56, art. 23. 60. A revised U.S. Model is apparently forthcoming from the Treasury Department. It was originally scheduled for release in December, 2004. See Kevin A. Bell, New Model Treaty Won’t Provide for Zero Dividend Withholding, 2005 TNT 70-7 (Apr. 13, 2005); Lee A. Sheppard, Angus Talks Treaty Policy, 2004 TNT 232-3 (Dec. 2, 2004) (stating that Treasury will issue an updated model treaty to reflect clauses in recently negotiated treaties). 61. See generally Peggy Musgrave, United States Taxation of Foreign Investment Income: Issues and Arguments (1969). The concept of capital export neutrality and its converse, capital import neutrality, were first developed by Peggy Musgrave in 1969 and they have been vigorously analyzed and debated ever since. For an overview of these norms, and an argument that capital export neutrality is generally the best principle for international taxation of both portfolio and direct investment, see Avi-Yonah, infra note 394, at 1604. See also Staff of Joint Comm. on Taxation, 108th Cong., Background Materials on Business Tax Issues Prepared for the House Committee on Ways and Means Tax Policy Discussion Series 53–54 (JCX-23-02) (Comm. Print 2002) (arguing that a worldwide tax system promotes economic efficiency, because investment location decisions will be governed by business considerations rather than tax considerations, and equity, because domestic and multinational activities are treated alike, and suggesting that worldwide taxation in some form is requisite to preserve the tax base from erosion by flight of activities to tax havens); Caves, supra note 44, at 190 (stating that all relevant taxes taken together are neutral if domestic and overseas investments that earn the same pre-tax return also yield the same after-tax return).

tax treaties for investment and aid to sub-saharan africa: a case study 579

U.S. investor in the same tax position as if the investment had been subject only to domestic tax.62 Nevertheless, most countries, including the United States, do not completely adhere to principles of capital export neutrality, regardless of the existence of tax treaties. Because the United States generally does not tax the foreign income of foreign companies, it is a relatively simple matter to avoid U.S. tax on much foreign income by placing the income stream in a foreign entity.63 In so doing, U.S. persons may defer U.S. taxation until the foreign earnings are repatriated in the form of dividends or capital gains.64 Deferral of this kind is the equivalent of a statutorily optional exemption of foreign income from U.S. taxation, as U.S. tax can be suspended indefinitely, according to the needs and desires of the shareholders.65 Thus, deferral allows

62. If tax credits perfectly offset foreign taxes paid, the taxpayer is indifferent to the allocation of the tax. See Caves, supra note 44, at 190. Most foreign tax credit systems are not perfectly offsetting but impose limitations as to creditability of taxes based on type or source of income and amount paid relative to domestic tax otherwise imposed. In the U.S., foreign taxes are currently segregated among several baskets according to the type of income that gave rise to the tax for purposes of applying a limit on the allowable tax credit. I.R.C. §§ 901–904 (2005). As a result, pooling of income from low-tax countries may be advantageous to taxpayers who have paid foreign taxes in excess of the allowable tax credit. See, e.g., David R. Tillinghast, Tax Treaty Issues, 50 U. Miami L. Rev. 455, 477 (1996). 63. See, e.g., Julie A. Roin, United They Stand, Divided They Fall: Public Choice Theory and the Tax Code, 74 Cornell L. Rev. 62, 113 (1988) (discussing the ease of avoiding U.S. tax through foreign entities); Avi-Yonah, supra note 25, at 1324–25 (arguing that as a result of the distinction between foreign and domestic companies in I.R.C. § 7701(a)(4) and (5) (2005) and the ensuing difference in taxation under I.R.C. §§ 11(d), 881, and 882 (2005), “taxpayers can easily choose between classification as foreign or domestic according to the formal jurisdiction of their incorporation”). 64. Deferral is limited to some extent, as discussed infra Part III.B. However, a U.S. person that earns active foreign income through a foreign corporation is generally not subject to U.S. tax until profits are repatriated as a dividend or the stock is sold, under the rules of Subpart F, I.R.C. §§ 951–964. 65. To allow deferral is therefore to provide incentives for active business operations to be located outside of the U.S., in low-tax jurisdictions. See Robert J. Peroni, Back to the Future: A Path to Progressive Reform of the U.S. International Income Tax Rules, 51 U. Miami L. Rev. 975, 987 (1997) (arguing that deferral “undercuts the fairness and efficiency of the U.S. tax system” by allowing profits earned overseas in low-tax jurisdictions to escape tax while equivalent domestic activities would be subject to tax). As a tax expenditure that costs the U.S. approximately $7.5 billion per year, deferral may be viewed as a subsidy, or tax incentive, for foreign business activities. See Office of Mgmt. & Budget, Executive Office of the President, Analytical Perspectives: Budget of the United States Government Fiscal Year 2005, at 287, available at http://www.whitehouse.gov/omb/budget/fy2005/ pdf/spec.pdf. Capital gains may be avoidable in the context of a conversion or liquidation of a subsidiary. See Dover Corp. v. Comm’r, 122 T.C. 324, 338, 353 (2004) (allowing the conversion of a foreign corporate subsidiary to disregarded entity status to avoid creation

580 allison christians

taxpayers to convert U.S. residence-based taxation to source-based taxation when it suits their purposes.66 To protect revenues, the United States has responded with a series of anti-deferral rules to prevent the easy escape of capital to foreign jurisdictions.67 To date, these anti-deferral measures have largely been restricted to passive income items so that deferral is still available for active income (residual taxation of which the United States might forego, under the foreign tax credit, if foreign taxes are in fact imposed). Despite the significance of deferral in curtailing the imposition of worldwide income taxation, the concept of residence-based taxation is the default system of most developed countries. The protection of residence-based taxation, by scaling back the need for foreign tax credits, was (and is) given as a reason—perhaps the primary reason—for entering into tax treaties. The OECD Model, as the baseline for the majority of the world’s tax treaties, thus represents an international consensus that the appropriate jurisdiction to tax income arising from cross-border activity is primarily the residence jurisdiction.68 This consensus, however, has not eliminated the limitations inherent in using tax treaties as the primary mechanism for the international coordination of tax matters. of subpart F income on its subsequent sale). The IRS, however, disagrees with this conclusion. See I.R.S. Chief Couns. Tech. Adv. Mem. 199937038 (Sept. 17, 1999), available at http://www.irs.gov/pub/irs-wd/9937038.pdf (holding that proceeds from sale of subsidiary after change in classification to disregarded entity did not escape subpart F); I.R.S. Chief Couns. Field Serv. Adv. Mem. 200049002 (Dec. 8, 2000), available at http://www. irs.gov/pub/irs-wd/0049002.pdf; I.R.S. Chief Couns. Field Serv. Adv. Mem. 200046008 (Aug. 4, 2000), available at http://www.irs.gov/pub/irs-wd/0046008.pdf (same, with sale made to related party). 66. See Peroni, supra note 65, at 987. 67. See, e.g., S. Rep. No. 99-313, at 363 (1986) (“[I]t is generally appropriate to impose current U.S. tax on easily movable income earned through a controlled foreign corporations since there is likely to be limited economic reason for the U.S. person’s use of the foreign corporation . . .”). In practice, current taxation applies to a significantly lesser extent than is contemplated under the subpart F rules, as these rules are apparently “not fully effective in meeting their objectives.” Harry Grubert, Tax Planning by Companies and Tax Competition by Governments: Is There Evidence of Changes in Behavior, in International Taxation & Multinational Activity 113, 137 (James R. Hines, Jr., ed. 2001) (Less than 50% of after-tax income of subsidiaries located in three Caribbean tax havens was subject to current tax under subpart F); see also Robert J. Peroni et al., Getting Serious about Curtailing Deferral of U.S. Tax on Foreign Source Income, 52 SMU L. Rev. 455, 464 (1999) (“[A]ntideferral provisions can be readily circumvented . . . ”). For a discussion of the deferral privilege and its inconsistency with U.S. international tax principles including the norm of capital export neutrality, see Peroni, supra note 65. 68. See Avi-Yonah, supra note 25, at 1303 (stating that a “coherent international tax regime exists that enjoys nearly universal support”); Reuven S. Avi-Yonah, International Taxation of Electronic Commerce, 52 Tax L. Rev. 507, 509 (1999) (arguing that the worldwide network of tax treaties constitutes an international tax regime with definable, common principles).

tax treaties for investment and aid to sub-saharan africa: a case study 581

3. Limitations on the use of treaties as international tax law Treaties are the traditional mechanism used for relieving double taxation on cross-border activity. Nevertheless, they have several significant limitations that render them an inefficient and unsatisfactory means of achieving their goals. This section discusses some of these limitations, including the incomplete coverage and restricted scope of tax treaties, their lack of uniformity, and their reliance on the assumption of reciprocal capital flows, and therefore reciprocal tax regimes, among contracting states. a. Limited coverage, scope, and uniformity. Not all countries have tax treaties, and no country has tax treaties with all the other countries of the world. The average individual tax treaty network comprises just 17 treaty partners, and over half of all countries have tax treaty networks of five or fewer treaty partners.69 In addition, the benefits of treaties are typically limited to activities conducted between the two signatory countries.70 As a result, there would have to be over 32,000 bilateral tax treaties to cover every possible cross-border transaction.71 The United States would have to enter into new treaties with over 160 countries to ensure that its coverage spanned the globe.72 At its current average rate of expansion of one new treaty per year since its first treaty was concluded with France in 1935, the prospect of completing a universal U.S. tax treaty network in a timely fashion appears slight.73

69. About 30% of countries have no tax treaties in force. For the 35 countries considered by the U.S. to be developed, the average network is about 49 treaties; for OECD countries, the average is 60. For less developed countries, the average is eight. Compiled in February 2005 from Ernst & Young, Worldwide Corporate Tax Guide (2004) and the LexisNexis Tax Analysts Worldwide Tax Treaties database, supra note 55. 70. This is almost universally true when the U.S. is a party. See U.S. Model, supra note 56, art. 22, at 31–33. 71. This figure is based on the assumption that there are approximately 255 independent nations in the world today—a figure that is an estimate because sovereignty of nations is a matter of foreign policy that varies from nation to nation. A currently prominent example is the case of Taiwan. See, e.g., Chen Redux: Inside the Rhetoric, There are Hints of a Thaw All Round, The Economist, May 22, 2004, at 37 (discussing China’s tight grip and world response). See also World Factbook, supra note 1, at 610–13 (country data on Taiwan), available at http://www.cia.gov/cia/publications/factbook/geos/tw.html. 72. The United States currently has 56 comprehensive income tax treaties in force which cover 64 countries. See John Venuti et al., Current Status of U.S. Tax Treaties and International Agreements, supra note 14 (listing all countries covered by tax treaties). The United States formally recognizes a total of 233 nations. See World Factbook at 628, 630, and and 639 (acknowledging the existence of 34 developed countries, 27 former USSR/ Eastern European countries, and 172 less developed countries). 73. Compiled by averaging the first entry-in-force dates of all comprehensive U.S. income tax treaties ever in force (on file with author).

582 allison christians

In addition, the OECD Model is aimed at only income taxation, to the exclusion of other kinds of taxes.74 Thus the term “double taxation” refers more particularly to double income taxation, and the term “relief of double taxation” refers particularly to the alleviation of circumstances in which two countries assert income taxation on the same item of income.75 Yet, there are a number of other taxes applied on businesses and individuals. Increasingly prominent throughout the world are consumption and trade taxes, and, primarily in developed countries, social security and other payroll taxes. As these taxes increase in application, tax treaties may cover a shrinking portion of revenues collected by countries. Finally, as contracts forged through negotiation, individual treaties deviate to various degrees from the standards set in the OECD Model.76 Treaties among

74. For reasons owing to historical distinctions that may be less clear today, income taxes have generally been attended to in tax treaties, while trade taxes are addressed in trade agreements. See generally Reuven S. Avi-Yonah & Joel Slemrod, Treating Tax Issues Through Trade Regimes, 26 Brook. J. Int’l L. 1683 (2001); Paul R. McDaniel, Trade and Taxation, 26 Brook. J. Int’l L. 1621 (2001); Alvin C. Warren, Income Tax Discrimination Against International Commerce, 54 Tax L. Rev. 131 (2001). 75. The OECD Model describes double taxation as “the imposition of comparable taxes in two (or more) states on the same taxpayer in respect of the same subject-matter and for identical periods.” OECD Model, supra note 21, at 7. 76. Even if their language is similar or identical, tax treaties may also vary due to differing interpretations under the domestic law of each country, or, in the case of U.S. treaties, pursuant to the agreement of the competent authorities. This is authorized under art. 3, ¶ 2 of the OECD, U.S., and UN Models, which state that any term not defined in the treaty is defined under the laws of each country as of the time the treaty is applied—i.e., “internal law, as periodically amended.” Postlewaite & Makarski, supra note 39, at 741 (adding that “[w]hen countries take different approaches to treaty interpretation, serious consequences may result, such as double taxation or the avoidance of any taxation”). The U.S. Model adds, “or the competent authorities agree to a common meaning pursuant to the provisions of Article 25 (the Mutual Agreement Procedure).” U.S. Model, supra note 56, art. 3, ¶ 2. Variation among treaties is also authorized under Article 25 of the OECD, U.S., and UN Models, which states that the competent authorities “shall endeavor to resolve by mutual agreement any difficulties or doubts arising as to the interpretation or application” of the treaty, and that the competent authorities “may also consult together for the elimination of double taxation in cases not provided for” in the treaty. U.S. Model, supra note 56, art. 25, ¶ 3; OECD Model, supra note 21, art. 25, ¶ 3; UN Model, infra note 78, art. 25, ¶ 3. The U.S. Model adds that “[t]he competent authorities also may agree to increases in any specific dollar amounts . . . to reflect economic or monetary developments.” U.S. Model, supra note 56, art. 25, ¶ 4. Finally, treaties may deviate from the international consensus even if they closely follow the model treaties due to periodic updates to the models and commentary thereto. For example, recent revisions to the OECD Model commentary with respect to the definition of a permanent establishment potentially broadens the scope of such provisions and may ultimately lead to a revision of Article 5 of the OECD Model. See, e.g., Richard M. Hammer, The Continuing Saga of the PE: Will the OECD Ever Get it Right?, 33 Tax Mgmt. Int’l J. 472 (2004) (suggesting that the current

tax treaties for investment and aid to sub-saharan africa: a case study 583

OECD member countries generally adhere to the pattern and main provisions of the OECD Model.77 Treaties between developed and less-developed countries, however, often contain nonstandard provisions. These provisions generally derive from a third model tax convention, first promulgated by the United Nations in 1980 (the UN Model). The UN Model was the product of a series of discussions and meetings of an Ad Hoc Group of Experts formed in 196778 to address concerns that the OECD Model (and, by association, the U.S. Model) was not appropriate for tax treaties involving nonreciprocal cross border activity.79 b. Assumption of reciprocal activity. The U.S. and OECD Models are directed at and work most effectively between two nations that export capital and transfer services in roughly reciprocal amounts. When treaty countries export and import capital to each other, each acts as a source country to investors from the other. Under these circumstances, tax treaties coordinate taxation without necessarily causing an imbalance in revenue allocation between the two countries: revenues given up by countries in their “source” role are recouped in their “residence” role.80 Consequently, such treaties are expected to have little revenue effect on either country.81 commentary should be revised because it is “murky and ambiguous,” and arguing for the incorporation of a clear de minimus rule in the OECD Model itself). 77. See OECD Model, supra note 21, at 10. Nevertheless, improvements and advances in international business and tax practices contribute to increased deviation even among OECD countries. Recently, so-called “double non-taxation” provisions have been introduced in new treaties. These provisions directly contravene existing OECD provisions. See, e.g., Michael Lang, General Report, in Int’l Fiscal Assoc., Double Non-Taxation, 89a Cahiers de Droit Fiscal International 77 (2004). 78. The Group of Experts included members from Latin American, North American, African, Asian, and European countries. The group also had observers from the IMF, the International Fiscal Association, the OECD, the Organization of American States, and the International Chamber of Commerce. See United Nations, Commentary on the Articles of the 1980 United Nations Model Double Taxation Convention Between Developed and Developing Countries 2 (Jan. 1, 1980); See also United Nations Dept. of Economic and Social Affairs, United Nations Model Double Taxation Convention Between Developed and Developing Countries viii (2001) (hereinafter UN Model). 79. See Leif L. Mutén, Double Taxation Conventions Between Industrialised and Developing Countries, in Double Taxation Treaties Between Industrialised and Developing Countries; OECD and UN Models, A Comparison 3 (Kluwer Law and Tax’n Pubs. 1990). 80. For example, while the U.S. may give up revenue by refraining from taxing dividends paid to foreign persons under a treaty, it recoups the loss by collecting the full tax on dividends paid by the foreign country to U.S. residents (without reduction under the foreign tax credit provisions, since under the treaty, the foreign country does not tax the dividend). See I.R.C. §§ 61 (U.S. persons taxed on income from whatever source derived) and 901 (foreign tax credit generally allowed only when foreign tax has been paid or accrued). 81. See, e.g., Staff of the S. Comm. on Foreign Rel., 105th Cong., Report on the Tax Convention with Ireland 17 (Comm. Print 1997) (“The proposed treaty is estimated to cause a negligible change in . . . Federal budget receipts”); Staff of the S. Comm. on Foreign

584 allison christians

If instead the flow of capital moves primarily from one country to another, reciprocity is lost. One country becomes primarily the source or host country, while the other becomes primarily the residence or home country. Because LDCs are typically capital importing countries, their primary role under tax treaties is as a source country.82 Residence jurisdiction will therefore be minimally exercised by LDCs.83 In such cases, a tax treaty shifts tax revenues inversely to the

Rel., 108th Cong., Report on the Tax Convention with the United Kingdom 16, 17 (Comm. Print 2003) (same). The balance apparently holds even in the case of complete exemption of source-country taxation. See, e.g., Staff of the Joint Comm. on Taxation, 108th Cong., Explanation of Proposed Protocol to the Income Tax Treaty Between the United States and Australia 39 (Comm. Print 2003) (suggesting that the new zero-rate for tax on direct dividends “would provide immediate and direct benefits to the United States as both an importer and an exporter of capital[,]” and that “[t]he overall revenue impact of this provision is unclear, as the direct revenue loss to the United States as a source country would be offset in whole or in part by a revenue gain as a residence country from reduced foreign tax credit claims with respect to Australian withholding taxes”). 82. The flow of capital between the U.S. and an LDC typically originates from the former and flows to the latter, although this is less true with respect to the “advanced developing” countries: Hong Kong, Korea, Singapore, Taiwan, and Brazil. See, e.g., Yoram Margalioth, Tax Competition, Foreign Direct Investments and Growth: Using The Tax System to Promote Developing Countries, 23 Va. Tax Rev. 161, 198 (2003) (stating that LDCs are generally not typical destinations for portfolio investment); see also Conventions and Protocols on Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital; Treaty Doc. 103-29, Sweden; Treaty Doc. 103–30, Ukraine; Treaty Doc. 103-31, Mexico; Treaty Doc. 103-32, France; Treaty Doc. 103–33, Kazakhstan; Treaty Doc. 103-34, Portugal; Treaty Doc. 104-4, Canada: Hearing Before the S. Comm. on Foreign Relations, 104th Cong. 18 (1995) (statement of Leslie B. Samuels, Assistant Secretary for Tax Policy, U.S. Dep’t of the Treasury) (capital flows are typically nonreciprocal between the U.S. and LDCs). Most multinationals are residents of developed countries. Of the top 100 multinational companies (as measured by foreign assets), just five are residents in LDCs (Hong Kong, Korea, Malaysia, Mexico, and Venezuela). United Nations Conference on Trade and Dev. [UNCTAD], World’s 100 Non-financial TNCs, Ranked By Foreign Assets, 2000, http://www.unctad.org/templates/Download.asp?d ocid=3812&lang=1&intItemID=2443 (hereinafter UNCTAD World 100 Non-Financial TNCs) (last visited Nov. 11, 2005). Of the top fifty multinational companies from developing economies, none are based in the LDCs of Sub-Saharan Africa. UNCTAD, The Top 50 TNCs from developing economies, ranked by foreign assets, 2000, http://www.unctad.org/templates/Download.asp?Docid=3811&lang=1&intItemID=244 3 (last visited Nov. 11, 2005). 83. It may be minimally exercised even in the absence of treaties, since few LDCs in Sub-Saharan Africa assert worldwide taxation on their residents. Among the exceptions are Angola, Ethiopia, Lesotho, Mauritius, Mozambique, Nigeria, Tanzania, and Uganda. See generally Ernst & Young, supra note 35. Ghana, the subject of the case study presented in Part III of this Chapter, generally exercises territorial taxation but imposes tax on certain repatriated earnings. See Republic of Ghana, Internal Revenue Act, 2000, Act 592, Part III, Div. I, § 6(1)(a) (hereinafter G.I.R.A.) (residents’ assessable income includes that “accruing in, derived from, brought into, or received in Ghana”).

tax treaties for investment and aid to sub-saharan africa: a case study 585

flow of capital. As a result, while the contraction of taxing jurisdictions is technically reciprocal in the treaty document, the one-sided flow of capital towards the LDC as source-country ensures that only that country experiences a true contraction of its taxing jurisdiction. The provider of the capital, namely the developed country, preserves its rights as the country of residence just as if the treaty had never been concluded. Nonreciprocal contraction by the LDC occurs in the context of portfolio investment as its role as the source country requires it to reduce its tax rates on dividends, interest, and royalties, while the residence country preserves the right to impose full taxation on these items. Nonreciprocal contraction also occurs in the context of active business income, as threshold rules for taxing business income prevent source-country taxation of certain activities, such as storing and displaying goods or building and construction activities.84 These threshold rules are embodied in the concept of the “permanent establishment.” The permanent establishment rules are found in Article 5 of each of the U.S., OECD, and UN model treaties. Under these rules, the source country agrees to refrain from taxing business income unless it is attributable to business activities that meet physical presence requirements, and even then, in some cases, only if the activities are conducted for a given duration or rise to a substantial enough level.85 Accordingly, under the U.S. and OECD Models, a permanent establishment is generally deemed to exist and therefore create taxing jurisdiction in the source country if business activities are conducted through a fixed place of business and consist of more than “peripheral or ancillary activities.” Certain activities, such as building and construction, however, must last more than a year in order to be deemed “permanent establishments.”86 Responding to the non-reciprocal aspects of relationships between developed and less developed countries, the UN Group of Experts sought to preserve source-country taxation in tax treaties in its Model. Thus, the UN Model provides for lower thresholds by shortening duration and presence requirements and including certain activities not included in the OECD and U.S. Models.87 84. See U.S. Model, supra note 56, art. 5; OECD Model, supra note 21, art. 5; UN Model, supra note 78, art. 5. 85. Id. 86. See U.S. Model, supra note 56, art. 5, ¶ 3. Peripheral and ancillary activities include exploratory or preparatory functions such as research and development, as well as activities considered incidental to the economic source of the income, such as storage, display, or delivery of goods. The U.S. Model is virtually identical to the OECD Model. 87. It otherwise adheres in large part to the OECD Model, and the two have become more similar. Indeed, the relevance of the UN Model has diminished significantly and it may be seen as irrelevant to the extent developed countries agree to higher source-based tax in their tax treaties, which they have done to a significant extent. See, e.g., John F. Avery Jones, Are Tax Treaties Necessary?, 53 Tax L. Rev. 1, 2 (1999) (“There seems [to be] little need for a separate model for developing countries”).

586 allison christians

For example, under the UN Model, a permanent establishment may arise after a duration of as low as six months for certain activities,88 fewer ancillary activities are excluded,89 and more income is attributed to permanent establishments via a force of attraction rule.90 Nevertheless, the UN Model limits source-country taxation simply by using the permanent establishment concept at all. In the absence of the treaty, the source country would typically provide little or no threshold to taxation.91 In addition, the UN Group of Experts determined that in treaties between developed and less developed countries, higher source-based taxation of passive items is appropriate. Just how high, however, has not been determined. While the OECD Model provides recommended maximum source-country tax rates for dividends (5% on “direct dividends”—those paid to corporate shareholders holding at least 10% of the paying company’s stock—and 15% on “regular

88. UN Model, supra note 76, art. 5, ¶ 3. In paragraph 3(a), building and construction activities and related supervisory activities are a permanent establishment if they last more than six contiguous months; in paragraph 3(b), consulting services are a permanent establishment if such services continue for a cumulative (even if non-contiguous) six months. In the OECD model, building and construction activities must continue for more than twelve months to constitute a permanent establishment, related supervisory activities are not included, and there is no parallel provision regarding consulting services. For a comparison of the OECD and UN Model permanent establishment provisions, see Bart Kosters, The United Nations Model Tax Convention and Its Recent Developments, Asia-Pacific Tax Bulletin, January/February 2004, at 4, available at http://unpan1.un.org/intradoc/ groups/public/documents/other/unpan014878.pdf. 89. For example, in the OECD and U.S. Models, the use of facilities or maintenance of a stock of goods for delivery is specifically excluded from the definition of permanent establishment, while in the UN Model it is not. Compare U.S. Model, supra note 56, art. 5, ¶ 4, and OECD Model, supra note 21, art. 5, ¶ 4, with UN Model, supra note 78, art. 5, ¶ 4. 90. The OECD and U.S. Models provide source-country taxation only of profits that are attributable to the permanent establishment. The UN Model includes profits attributable to the sale of the same or similar goods or merchandise as those sold through the permanent establishment and profits from the same or similar business activities as those conducted through the permanent establishment. Compare U.S. Model, supra note 56, art. 7, ¶ 1, and OECD Model, supra note 21, art. 7, ¶ 1, with UN Model, supra note 78, art. 7, ¶ 1. 91. For an argument that thresholds are appropriate, should be used even in the absence of a treaty, and should be made more uniform (in the current models, there are different thresholds for different activities), see Brian J. Arnold, Threshold Requirements for Taxing Business Profits Under Tax Treaties, in The Taxation of Business Profits Under Tax Treaties 55 (Brian J. Arnold et al. eds., 2003). The permanent establishment concept has been revised and updated to adapt to changes in business and technology over the years, but generally remains consistent with the original version introduced in the first OECD Model Tax Convention, which was released in 1963. OECD, Income and Capital Draft Model Convention, Draft Convention For the Avoidance of Double Taxation With Respect to Taxes on Income and Capital, art. 5, ¶¶ 1-2 (July 30, 1963).

tax treaties for investment and aid to sub-saharan africa: a case study 587

dividends”—all other shareholders), interest (10%), and royalties (0%),92 and the U.S. Model is virtually identical (but provides zero source-country taxation of interest),93 the UN Model leaves the source-country taxation of these items to be established through bilateral negotiations.94 Thus, the UN Model implies that higher tax rates are appropriate in tax treaties with LDCs, but declines to recommend exactly what rate is appropriate.95 The United States has frequently used the provisions and concepts of the UN Model in its tax treaties with developed as well as less developed countries.96 For example, the U.S. income tax treaties with Barbados, Canada, China, Cyprus, Egypt, Estonia, India, Indonesia, Jamaica, Kazakhstan, Korea, Latvia, Lithuania, Mexico, Morocco, the Philippines, Thailand, Tunisia, Turkey, the Ukraine, and Venezuela each provide for lower permanent establishment duration requirements, narrower definitions of ancillary and preparatory activities, higher sourcecountry tax rates on passive income items, or a combination of these features.97

92. See OECD Model, supra note 21, arts. 10, 11, 12. 93. U.S. Model, supra note 56, art. 11. 94. UN Model, supra note 78, art. 11 (including a blank line and a parenthetical that states “the percentage is to be established through bilateral negotiations”). 95. See generally U.S. Model, supra note 56, arts. 10, 11, 12; OECD Model, supra note 21, arts. 10, 11, 12; UN Model, supra note 78, arts. 10, 11, 12. 96. Kosters, supra note 88, at 9. 97. See U.S.-Barbados Treaty, supra note 4, art. 5; Convention with Respect to Taxes on Income and on Capital, U.S.–Can., art. 5, Sep. 26, 1980, T.I.A.S. No. 11087 (hereinafter, U.S.–Canada Treaty); U.S.–China Treaty, supra note 4, art. 5; U.S.–Cyprus Treaty, supra note 4, art. 9; U.S.–Cyprus Treaty, supra note 4, art. 5; U.S.–Egypt Treaty, supra note 4, art. 5; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Est., art. 5, Jan. 15, 1998, S. Treaty Doc. No. 105-55 (1999); U.S.–India Treaty, supra note 4, art. 5; U.S.–Indonesia Treaty, supra note 4, art. 5; U.S.–Jamaica Treaty, supra note 4, art. 5; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital, U.S.–Kaz., art. 5, Oct. 24, 1993, S. Treaty Doc. No. 103-33 (1996); U.S.–Korea Treaty, supra note 4, art. 9; Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Lat., art. 5, Jan. 15, 1998, S. Treaty Doc. No. 105-57 (1999); Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Lith., art. 5, Jan. 15, 1998, S. Treaty Doc. No. 105-56 (1999); Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.– Mex., art. 5, Sep. 18, 1992, S. Treaty Doc. No. 103-7 (1993) [hereinafter U.S.–Mexico Treaty]; U.S.–Morocco Treaty, supra note 4, art. 4; U.S.–Philippines Treaty, supra note 4, art. 5; U.S.–Thailand Treaty, supra note 4, art. 5; U.S.–Tunisia Treaty, supra note 4, art. 5; Agreement for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Turk., art. 5, Mar. 28, 1996, S. Treaty Doc. No. 104-30 (1997); Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital, U.S.–Ukr., art. 5, Mar. 4, 1994, S. Treaty Doc. No. 104-30 (2000); U.S.–Venezuela Treaty, supra note 4, art. 5.

588 allison christians

The consequence of preserving source-country taxation to overcome nonreciprocal capital flows, however, is that it undermines the relief of double taxation ostensibly sought as the primary purpose for entering into the treaty in the first place. This has been a source of problems for drafters and negotiators of tax treaties and treaty models, who appear to have difficulty determining whether it is better for LDCs to preserve source-country taxation so as to allow the source country to collect the maximum amount of revenues, or to relieve source-country taxation so as to attract the maximum amount of foreign investment.98 As discussed in Part IV, this choice is one of the main reasons tax treaties have become obsolete for many investors in LDCs. Yet new U.S. tax treaties with LDCs continue to be sought, and, when concluded, they continue to provide for higher source-country taxes on passive income items, even, on occasion, when the treaty rate exceeds that of the internal laws of the LDC.99 The importance of reciprocity as requisite to make a tax treaty appropriate is demonstrated in the current composition of the U.S. tax treaty network. Like all developed countries, the United States has tax treaties in place with all of its major reciprocal trading partners100 and with the bulk of its foreign direct investment sources and destinations.101 Yet, with just 55 comprehensive tax treaties covering 62 countries, the U.S. network is comparatively small relative to the other major economies of the world,102 and it excludes more than 20% of U.S. foreign direct investment.103 Moreover, just 16 U.S. tax treaties are with LDCs,104

98. Tsilly Dagan eloquently illustrated the conundrum and presented a game theory rationale that explains why many LDCs have opted for the latter. See generally Dagan, supra note 40. 99. See discussion infra at note 369. 100. Major trade partners include Canada, China, Germany, Japan, Mexico, and the U.K. World Factbook, supra note 1, at 577 (looking at entries listing major import and export partners). The most glaring exception in the U.S. tax treaty network is probably Brazil, with whom negotiations have been stalled since 1992. See infra note 537. 101. The tax treaty network currently covers approximately 78% of U.S. foreign direct investment, as valued at historical cost (book value of U.S. direct investors’ equity in and net outstanding loans to foreign affiliates). See Maria Borga & Daniel R. Yorgason, Direct Investment Positions for 2003: Country and Industry Detail, Surv. of Current Bus., July 2004, at 40, 49 (hereinafter Borga & Yorgason), http://www.bea.doc.gov/bea/ARTICLES/2004/ 07July/0704dip.pdf. 102. In contrast, the United Kingdom and France each have tax treaties with over 100 countries; Canada and the Netherlands with over 80. See Ernst & Young, supra note 35, at 134–36, 263–65, 641–42, 984–85. 103. See supra note 69. 104. See supra note 4. When the tax treaty with Sri Lanka (signed in 1985) entered into force in July of 2004, it was the first new country added to the tax treaty network since the treaty with Slovenia entered into force in 2001, and the first new LDC since Venezuela was added in 1999. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital, U.S.-Slovn., June 21, 1999,

tax treaties for investment and aid to sub-saharan africa: a case study 589

as compared with an average of 22 in other leading economies.105 To the extent that tax treaties influence the flow of trade and investment between the United States and the rest of the world, they may impact U.S. foreign investment, trade, and aid efforts to LDCs. The following part explores whether more complete U.S. tax treaty coverage could impact these flows by considering a hypothetical tax treaty with Ghana, an LDC in Sub-Saharan Africa. B. U.S. Tax Treaties with LDCs: Case Study of Ghana This part presents as a case study a hypothetical tax treaty between the United States and Ghana, based on current U.S. tax treaty standards with respect to LDCs. The case study demonstrates that the lack of tax treaties between the United States and the LDCs of Sub-Saharan Africa may be explained in large part by the fact that in today’s global tax climate, these agreements would not significantly impact the global tax burden that current or potential international investors are facing. As a result, even if governments are committed to concluding them, and even though they are supported by academics, practitioners, and lawmakers, tax treaties between the United States and the LDCs of Sub-Saharan Africa would nevertheless be largely ineffective in stimulating cross-border investment and trade. 1. Ghana as case study subject The pursuit of a tax treaty with Ghana, a nation of 20 million people in West Africa, would support current U.S. commercial and noncommercial interests in this country. Noncommercial interests of the United States in Ghana include longstanding diplomatic ties,106 an interest in fostering economic stability in this region of the world for humanitarian reasons, and recognition that conditions of extreme poverty like those found in Ghana are a potential breeding ground for terrorism.107

S. Treaty Doc. No. 106-9 (2001); U.S.-Venezuela Treaty, supra note 4; U.S.–Sri Lanka Treaty, supra note 4. 105. 17 of the 30 OECD countries have larger LDC tax treaty networks. For example, the United Kingdom and France each have tax treaties with 60 LDCs, Canada has 40, Germany has 36, Norway has 35, and Italy and Sweden each have 32. Compiled from Ernst & Young, supra note 35, at 129–31, 250–52, 287–88, 426–28, 652–53, 855–57, 938–39 and the LexisNexis Tax Analysts Worldwide Tax Treaties database, http://w3.nexis.com/ sources/scripts/info.pl?250064 (last visited Nov. 11, 2005). 106. See U.S. Department of State, Background Note: Ghana, http://www.state.gov/r/ pa/ei/bgn/2860.htm (last visited Nov. 11, 2005) (“The United States has enjoyed good relations with Ghana at a nonofficial, personal level since Ghana’s independence. Thousands of Ghanaians have been educated in the United States. Close relations are maintained between educational and scientific institutions, and cultural links, particularly between Ghanaians and African-Americans, are strong”). 107. Embassy bombings in Kenya and Tanzania in 1998, allegedly linked to the international terrorist organization al-Qaeda, provide perhaps the most illustrative reason for U.S. interests in brokering peace and stability in Sub-Saharan Africa. The U.S. also has

590 allison christians

U.S. commercial interests in Ghana include both trade and investment relationships. Several large foreign investments in Ghana are owned by U.S. companies,108 and U.S. companies continue to express interest in pursuing business opportunities in this country.109 U.S. investment in and trade with Ghana is generally facilitated by a number of factors. For instance, as a former colony of the United Kingdom,110 Ghana’s official language is English,111 its laws are a blend of customary law and English common law, and its regulatory state derives

interests in Sub-Saharan Africa for social justice reasons, including the extreme poverty faced by a majority of the population in this region. For a discussion of the importance of pursuing tax treaties in response to these issues, see Brown, supra note 7, at 48–51. 108. These include the Volta Aluminum Company, Ltd (Valco), a Ghanaian aluminum manufacturing company that is jointly owned by Kaiser Aluminum Corp. (a Texas corporation owning 90%) and Alcoa Inc., (a Pennsylvania corporation owning 10%); Regimanuel Gray, a construction company jointly owned by Regimanuel Ltd. (a Ghanaian company) and Gray Construction (a Texas corporation); and Equatorial Bottlers, a bottling company wholly owned by the Coca Cola Company (a Delaware corporation). See the company websites of Valco at http://www.alcoa.com/ghana/en/home.asp, Regimanuel Gray at http://www.regimanuelgray.com/about.asp, and Equatorial Bottlers at http://www.ghana. coca-cola.com (each describing the respective companies’ U.S. ownership and Ghanaian operations). 109. See, e.g., Newmont to Start up in Ghana, The Daily Telegraph (Sydney, Australia), Dec. 22, 2003, at 59 (discussing the purchase by Newmont Mining Corp., a Delaware Corporation, of the Ahafo gold mine in Ghana); Elinor Arbel, AMR, Pier 1 Imports, Sun Microsystems: U.S. Equity Movers Final, Bloomberg News, Aug. 16, 2004 (discussing plans by Alcoa Inc., a Pennsylvania corporation, to buy and restart an aluminum smelter in Ghana); G. Pascal Zachary, Searching for a Dial Tone in Africa, N.Y. Times, July 5, 2003, at C1 (quoting a former senior executive of Microsoft who surveyed Ghana as a potential regional hub for an information-technology industry, who stated that Ghana “has the potential to become for Africa what Bangalore became for India,” and discussing Rising Data Solutions, a Maryland corporation that recently introduced a call center in Ghana and Affiliated Computer Services, a Dallas company that began doing business in Ghana in 2001 and is looking to expand its operations). 110. 17 LDCs in Sub-Saharan Africa are former colonies of the United Kingdom: Botswana, Gambia, Ghana, Kenya, Lesotho, Malawi, Mauritius, Nigeria, Seychelles, Sierra Leone, Somalia, Sudan, Swaziland, Tanzania, Uganda, Zambia, and Zimbabwe; all but Somalia and Sudan designate English as their official language; an additional four countries list English among their official languages. See World Factbook, supra note 1, at 73, 204, 214, 295, 318–19, 339–40, 359–60, 406, 487–90, 502, 515–16, 522–23, 536, 562–63, 605–06, 608–09. 111. The use of English may be an important factor for the foreign investment location decisions of U.S. multinational firms. See Irving B. Kravis & Robert E. Lipsey, The Location of Overseas Production and Production for Export by U.S. Multinational Firms 32 (Nat’l Bureau of Econ. Research, Working Paper No. 0482, 1982), available at http://ssrn.com/ abstract=262704 (stating that identity of language may significantly reduce the costs of foreign investment activities).

tax treaties for investment and aid to sub-saharan africa: a case study 591

much from the British system, thus providing a familiar framework for commercial relations.112 U.S. trade and aid initiatives specifically identify Ghana as regionally significant to U.S. trade interests due to its central location in an international business corridor that stretches from Nigeria to Côte d’Ivoire.113 As is the case for many LDCs,114 the United States is one of Ghana’s principal trading partners, although U.S. goods comprise a small portion of Ghana’s total imports.115 As a result, like most of the LDCs in Sub-Saharan Africa, Ghana is a relatively untapped market for U.S. exports.116

112. U.S. multinational companies may prefer to invest in countries with which they have economic, political, language, or cultural ties. John H. Dunning, The Globalization of Business: The Challenge of the 1990s 37–43 (1993) (discussing geographical clustering of multinational companies). 113. President of the United States, 2003 Comprehensive Report on U.S. Trade and Investment Policy Toward Sub-Saharan Africa and Implementation of the African Growth and Opportunity Act 5 (2003), http://www.agoa.gov/resources/annual_3.pdf. Ghana is also poised to be the financial hub of a West African monetary zone that was expected to be established in July 2005 but is now targeted for implementation in 2009. See, e.g., Hon. Yaw Osafo-Maafo, M.P., Minister of Finance and Economic Planning, The Budget Statement and Economic Policy of the Government of Ghana for the 2004 Financial Year § 2(43) (2004); Peter J. Obaseki, The Future of the West African Monetary Zone Programme, 5 W. African J. of Monetary and Economic Integration 2, 4 (2005). When established, the West African monetary zone will facilitate commerce in the region by introducing a single currency in the Economic Community of West African States, which includes Benin, Burkina Faso, Cape Verde, Côte d’Ivoire, the Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Niger, Nigeria, Senegal, Sierra Leone, and Togo. For information,, see the Economic Community of West African States Home Page, http://www.sec.ecowas.int (last visited Nov. 11, 2005). 114. The U.S. is a principal export partner to 55% of LDCs, and a principal import partner to 40%. Compiled from World Factbook, supra note 1, at 16, 93, 99, 104, 109, 111, 126, 129, 131, 139, 156, 174, 176, 182, 203, 205, 215, 236, 296–97, 320, 322, 339, 341, 360, 382, 385, 404, 407, 491, 524, 564. 115. See U.S. Department of State, Ghana Country Commercial Guide FY2002, ch. 1, available at http://strategis.ic.gc.ca/epic/internet/inimr-ri.nsf/en/gr-80318e.html (stating that “[i]n the past, Ghana conducted most of its external trade with Europe, but trade with the United States is becoming increasingly significant”). Ghana’s import market is currently dominated by Nigeria, contributing 21.3% of all imports, followed by China, with 8.7% and the United Kingdom with 6.7%. World Factbook, supra note 1, at 215. The U.S. is its fourth largest partner, contributing 5.6% of total imports. Id. In comparison, the U.S. is currently a principal exporter to 18 other LDCs in Sub-Saharan Africa, contributing 50% of imports in Namibia, 42.3% in Eritrea, 31% in Equatorial Guinea, and between 12% and 19% in Angola, Chad, and Ethiopia. Id. at 16, 111, 174, 176, 182, 385. 116. As a potential export market, Ghana and other LDCs in Sub-Saharan Africa are also important to the U.S. labor market. See John Cochran, Not Out of Africa: Bush Visits Africa—But Why Now?, ABC News Report, July 8, 2003, available at http://abcnews.go.

592 allison christians

Trends in U.S. trade and investment interests in Ghana support the notion that increasing investment in this country is a viable goal, which is being advanced by current efforts in executing international agreements. For example, U.S. trade with Ghana increased following the enactment and implementation of AGOA.117 Nevertheless, U.S. investment in Ghana remains relatively slight by global standards.118 Low levels of investment in Ghana may be explained by a number of factors including several nontax barriers to investment. For instance, Ghana’s low level of infrastructure has been blamed as a major impediment to increased investment.119 Examples of Ghana’s infrastructural shortcomings include obvious physical burdens such as poorly maintained roads,120 interruptions in electricity,121

com/sections/wnt/World/africa030708_bush.html (last visited June 11, 2003) (“Over 100,000 U.S. jobs depend on exports to Africa . . .”). 117. Since 2000, when AGOA was first implemented, U.S. exports to Ghana have grown steadily, but imports from Ghana have decreased. United States International Trade Commission, U.S. Trade and Investment with Sub-Saharan Africa (2000), http:// www.usitc.gov/secretary/fed_reg_notices/332/I0516x3.HTM. 118. UNCTAD, World Investment Report 14 (2003), http://www.unctad.org/ en/docs/ wir2003_en.pdf (hereinafter WIR 2003). 119. See U.S. Department of State, Ghana Country Commercial Guide FY2003, ch. 7, § A1, available at http://strategis.ic.gc.ca/epic/internet/inimr-ri.nsf/en/gr109073e.html (stating that infrastructure shortcomings have impeded domestic productivity and discouraged foreign direct investment). Along with the rest of Sub-Saharan Africa, which experienced a large and continuing decline in foreign direct investment (FDI) in tandem with the global financial crisis of the late 1990s, Ghana’s share of global foreign investment has dropped significantly over the past few years, and it is considered an underperformer with respect to its FDI potential. Id. Its 40% decline in FDI from 2001 to 2002 mirrors the experience of the continent, to which FDI declined as a whole from $19 billion in 2001 to $11 billion in 2002 (a 41% decline). These declines are sharp when compared to that for global FDI, which declined as a whole by 21% in the same period. See WIR 2003, supra note 118, at 3, 14. 120. As John Torgbenu, a taxi driver in Accra, describes the multitude of certifications needed to obtain a cab license in Ghana: “Cars must be road-worthy, but the roads need not be car-worthy.” Interview with John Torgbenu, Taxi Driver in Accra, Ghana (2003) (on file with author). See also Memorandum of Economic and Financial Policies of the Government of Ghana for 2003–05, ¶ 8 (March 31, 2003) (hereinafter MEFP) (“Ghana’s basic infrastructure continues to remain in very poor shape. The building of roads, ports, and communication networks . . . have been driving forces behind the government’s efforts to secure a predictable flow of external financing for infrastructure development”). 121. Despite the presence of West Africa’s largest hydro-electric plants at Volta Lake in northern Ghana, electricity outages are such a frequent phenomenon that individuals, businesses, and institutions that can afford generators have them, and put them to use on a regular basis. Fueling the modernization process is one of the key developments sought in connection with Ghana’s requests for IMF funding. See MEFP, supra note 120.

tax treaties for investment and aid to sub-saharan africa: a case study 593

a lack of clean water,122 and a paucity of institutions such as schools and hospitals.123 Equally problematic are Ghana’s excessive administrative requirements and bottlenecks, as well as other barriers to the entry and operation of businesses by foreign persons.124 For example, Ghana continues to struggle with land and property protection,125 restricts foreign ownership of real property,126 and has only recently dismantled regulations that completely closed several industries to foreign investors.127 As part of its approach to poverty reduction and economic growth through the creation of a business-friendly environment, Ghana’s current administration has pledged to make significant improvements to its infrastructure.128 The reduction 122. Ghana is among the majority of LDCs in the world that have not developed an improved water supply. See World Health Organization, Water Supply, sanitation and hygiene development, http://www.who.int/water_sanitation_health/hygiene/en/ (last visited Nov. 11, 2005). 123. Low levels of education, literacy, and health care issues are among the institutional issues Ghana faces. See, e.g., George Gyan-Baffour, The Ghana Poverty Reduction Strategy: Poverty Diagnostics and Components of the Strategy 4 (2003), http://www. casmsite.org/Documents/Elmina%202003%20-%20Workshop%20-%20Poverty%20 Reduction%20-%203.pdf (last visited Nov. 11, 2005). 124. Much of these administrative regimes are a lasting legacy of colonization, under which the European nations imposed severe market controls to preserve the resources of their colonies for their exclusive use. See, e.g., Francis Agbodeka, An Economic History of Ghana 126–27 (1992). For an overview of ease of entry issues for LDCs generally, see Jeffrey C. Hooke, Emerging Markets: A Practical Guide for Corporations, Lenders, and Investors (2001) (discussing the entrenched obstacles to entry in LDCs); see also Leora Klapper et al., Business Environment and Firm Entry: Evidence from International Data 16 (Nat’l Bureau of Econ. Research, Working Paper No. 10380, 2004), available at http:// papers.nber.org/papers/w10380.pdf (finding that bureaucratic entry regulations are a significant burden that hampers the entry of firms into foreign markets). 125. Courts in Ghana are overwhelmed with land disputes. See, e.g., Joseph Coomson, Country Achieves Below 40 Percent Delivery, The Ghanaian Chronicle, Aug. 18, 2005 (discussing “many land disputes among traditional authorities” and stating that there are currently “more than 62,000 land disputes . . . being heard at the courts”). 126. The inability to own land translates to an inability to use land as collateral for financial transactions, which is a key to economic growth. See Enrique Gelbard & Sérgio Pereira Leite, Measuring Financial Development in Sub-Saharan Africa 18 (Int’l Montetary Fund, Working Paper No. 99/105, 1999), available at http://www.imf.org/external/pubs/ ft/wp/1999/wp99105.pdf. 127. See WIR 2003, supra note 118, at 36. 128. Hon. Yaw Osafo-Maafo, M.P., Minister of Finance and Economic Planning, The Budget Statement and Economic Policy of the Government of Ghana for the 2003 Financial Year ¶ 22 (2003) (pledging the government’s “commitment to deliver a macroeconomic environment that underpins our strategy for growth and poverty reduction”); Hon. Yaw Osafo-Maafo, M.P., supra note 113, at ¶ 4 (pledging to continue to “create an enabling environment for wealth creation”). See also various documentation in connection with IMF lending, including the MEFP, supra note 120.

594 allison christians

of administrative obstacles, combined with greater certainty with regard to the legal and regulatory regime, is credited with a recent surge in foreign investment from South Africa to other countries in Sub-Saharan Africa.129 It is hoped that this surge will be followed by increased investment from other countries, including the United States. An increased share of foreign investment is also expected to lead to spillover effects that would remedy some of the current deficiencies in physical infrastructure. Limited spillover effects have been achieved recently in connection with Ghana’s gold mining operations, which have provided funding to improve transportation routes.130 In Nigeria, one of Ghana’s close neighbors, investors in the telecommunications industry funded the installation of communication networks throughout the country.131 Ghana’s growing telecommunications industry

129. Nicole Itano, South African Companies Fill a Void, N.Y. Times, Nov. 4, 2003, at W1 (“It’s safer to go in, it’s easier to get materials in and out, easier to repatriate your profits,” according to Keith Campbell, a managing director of a South African risk management firm and vice-chairman of the South Africa–Angola Chamber of Commerce). The overhaul of economies has often been initiated by the international lending organizations, which have faced much criticism and been the subject of much debate in the face of the apparent failure of many of their reform efforts. However, the extreme opposite approach, as unfortunately presented in the case of Zimbabwe, illustrates the need for some fundamental certainty in dealing with foreign businesses in order to attract foreign investment and maintain a stable economy. See, e.g., Michael Wines, Around Ruined Zimbabwe, Neighbors Circle Wagons, N.Y. Times, July 6, 2005, at A4 (describing the “fiscal and political collapse” of that country since it began seizing white-owned farms in 2000); David White & John Reed, Showdown over Pariah State Leaves the Commonwealth Divided and Frustrated, Fin. Times, Dec. 9, 2003 (discussing ramifications of Zimbabwe’s withdrawal from the British Commonwealth); Tony Hawkins, Zimbabwe Dollars Cut 80% at Auction, Fin. Times, Jan. 13, 2004 (stating that massive currency devaluation is in line with market expectations for Zimbabwe). 130. Ghana’s gold mines have recently sparked interest from foreign investors, who will spend millions of dollars to upgrade and develop operations following years of neglect and under-maintenance of these operations, because they expect productivity to increase dramatically and produce significant profit as a result. See Mr. Jonah Goes to Jo’burg, Economist, Jan. 17, 2004, at 56 (AngloGold [South Africa] expects to spend between $220 and $500 million to upgrade its newly acquired Ghanaian gold mine: Ashanti Goldfields); Newmont to Go for Ghana Gold, Advertiser (S. Austl.), Dec. 22, 2003, Finance at 50 (Newmont,U.S., plans to spend about $350 million to develop its recently-acquired Ghanaian gold mine, Ahafo). See also Big-Game Hunting, Economist, Aug. 16, 2003, at 57; Gargi Chakrabarty, Newmont OKs Project in Ghana; Gold Producer Invests $350 Million in W. African Mine, Rocky Mountain News, Dec. 19, 2003, at B2; and Gargi Chakrabarty, Latest Global Hot Spot for Gold Mining: Ghana, Rocky Mountain News, Oct. 30, 2003, at B1. 131. South Africa’s Vodacom recently spent $119 million building a cellular network in the Congo, a critically impoverished country that has only recently emerged from devastating civil war. South Africa’s MTN Group spent approximately $1.75 billion building cellular networks in five different Sub-Saharan Africa countries ($900 million in Nigeria

tax treaties for investment and aid to sub-saharan africa: a case study 595

may draw like commitments from future investors.132 However, the components of infrastructure that are not produced by spillover, such as the legal and regulatory framework that protects businesses and creates an environment for growth, generally must be directly supported and funded by the government.133 Despite the infrastructural obstacles present in Ghana, U.S. investment in this country continues to grow, albeit slowly. The following section explores whether and how such a tax treaty between the two countries might affect investment in Ghana. 2. Structure of a tax treaty between Ghana and the United States As discussed in the Introduction, the U.S. Model serves as the template for all new tax treaties negotiated by the Treasury Department, though the OECD Model and other recent treaties are also consulted. Thus, in structure and overall content, a tax treaty between the United States and Ghana would emulate the model treaties, especially the U.S. Model, to a substantial degree. When negotiating with LDCs, however, the Treasury Department also consults the UN Model.134 As a result, these treaties usually contain several standard deviations from the U.S. Model, described in reports and technical explanations as “developing-country concessions.”135 They are called concessions because they typically concede

alone), and experiences a 40% profit margin in these markets – despite having to build power generators to overcome a lack of stable power sources and a transmission network to connect cities and towns across the country—compared to its 30% return at home in South Africa. Itano, supra note 129. 132. See Zachary, supra note 109. 133. Coercion of various forms may induce companies to provide such infrastructure in the absence of voluntary action. For example, in 2003 foreign oil workers were kidnapped in Nigeria in an effort to extract a promise from a foreign company to build a school or a health center. See Nigeria’s Oil-Rich Area Mired in Poverty, Daily Graphic (Ghana), Dec. 3, 2003, at 5. Clearly no government should be encouraged to rely on these kinds of tactics to build adequate infrastructure, but the fact that citizens of a nation are willing to engage in illegal acts to secure public goods illustrates the tensions and pressures facing both international businesses and the governments struggling to attract such businesses. 134. See, e.g., Department of the Treasury Technical Explanation of the Convention Between the Government of the United States of America and the Government of the Democratic Socialist Republic of Sri Lanka for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, Mar. 14, 1985, as Amended by a Protocol Signed at Washington on Sept. 20, 2002, http://www.treas.gov/press/ releases/reports/tesrlanka04.pdf (“Negotiations also took into account the [OECD Model], the [UN Model], and recent tax treaties concluded by both countries”). 135. This designation has been consistently propounded throughout U.S. tax treaty history, and continues virtually unchanged today. Compare, e.g., Staff of Joint Comm. on Taxation, 101st Cong., Explanation of Proposed Income Tax Treaty (and Proposed Protocol) Between the United States of America and the Republic of India 10-11 (Comm. Print 1990) (“The proposed treaty contains a number of developing country concessions . . . providing

596 allison christians

U.S. residence-based taxing jurisdiction in favor of greater source-country taxation.136 An example of a U.S. treaty with an LDC, as compared to the U.S. Model Treaty, demonstrates the operation of these concessions. At the time it was entered into, the U.S. tax treaty with Jamaica was deemed to be the “precedent for negotiations” with other LDCs.137 At 24 years of age, that treaty is substantially out of date, as many tax laws in the United States (and presumably in Jamaica) have changed significantly since it entered into force in 1981.138 However, the principle of enlarging source-country taxation found in the U.S.Jamaica treaty continues to appear in new tax treaties with other LDCs.139 Therefore, the following discussion uses the U.S.–Jamaica treaty to model the terms that might be expected in a U.S.–Ghana tax treaty, should one be concluded. In the U.S. tax treaty with Jamaica, as in most U.S. tax treaties with LDCs, the expectation that nonreciprocal capital flows may negatively impact the LDC is addressed by preserving source-country taxation. This is mainly accomplished through modifications to the articles dealing with the determination of for relatively broad source-basis taxation.”), and Staff of S. Foreign Relations Comm., 101st Cong., Report on the Tax Convention with the Republic of India ([Comm. Print] 1990), with Staff of Joint Comm. on Taxation, 108th Cong., Explanation of Proposed Income Tax Treaty Between the United States and the Democratic Socialist Republic of Sri Lanka 18, 64 (Comm. Print 2004) (hereinafter Explanation of Sri Lanka Treaty) (describing these deviations as substantive, and outlining the major provisions), and Staff of Joint Comm. on Taxation, 97th Cong., Explanation of Proposed Income Tax Treaty Between the United States and the Philippines (Comm. Print 1981). 136. Explanation of Sri Lanka Treaty, supra note 135 at 64. To the extent that sourcebased taxing jurisdiction is theoretically more justifiable, the term concession is something of a misnomer. See discussion in Part II.B. Nevertheless, as much source-based jurisdiction has been ceded in favor of residence-based jurisdiction in the evolution of the model treaties, a reversal of this norm, especially in the case of non-reciprocal capital flows, can in theory shift greater tax revenue collection to the country of source. By so doing, it requires the residence country to revert to the role of relieving double taxation via the generosity of the foreign tax credit, discussed supra, note 43 and accompanying text. Nevertheless, the theory that revenues are conceded under these provisions only holds if the source country actually imposes and collects the tax. This is an assumption which cannot be relied upon in today’s global economy, as discussed infra Part IV.B.2. 137. Staff of Joint Comm. on Taxation, 97th Cong., Explanation of Proposed Income Tax Treaty (and Proposed Protocol) Between the United States and Jamaica 4 (Comm. Print 1981). 138. U.S.–Jamaica Treaty, supra note 4. 139. Evidently, in some cases this is done regardless of the pre-existing legal framework in the LDC. See Explanation of Sri Lanka Treaty, supra note 135, at 62 (stating that “it is not clear that . . . Sri Lankan laws have been fully taken into account” since “[s]everal of the articles of the proposed treaty contain provisions that are less favorable to taxpayers than the corresponding rules of the internal Sri Lankan tax laws”).

tax treaties for investment and aid to sub-saharan africa: a case study 597

thresholds for taxing income from business activities (the permanent establishment provision) and those dealing with the taxation of passive-type income (dividends, interest, and royalties provisions).140 First, under the permanent establishment concept, source-country taxation is enlarged by expanding the definition to allow the LDC to impose taxation on more of the business profits earned by foreign persons in the source country. Thus, in the U.S.–Jamaica treaty, the permanent establishment provision mirrors the structure of the U.S. and OECD Models, but incorporates the UN Model approach, shortening the threshold durational requirement from one year to six months in the case of construction, dredging, drilling, and similar activities.141 It also provides that the furnishing of services can create a permanent establishment if continued for more than ninety days a year.142 Finally, it provides that maintaining substantial equipment or machinery in a country for four months can constitute a permanent establishment.143 One or more of these deviations from the U.S. Model are found in most U.S. tax treaties with LDCs.144 Consequently, similar provisions would likely be suggested, negotiated, and agreed to in a U.S.– Ghana tax treaty. Second, under the passive income provisions, source-country taxation is enlarged by allowing the source country to impose tax rates on these items of

140. See supra Part II.C. 141. U.S.-Jamaica Treaty, supra note 4, art. 5, ¶ 2(i). The activity must continue for “more than 183 days in any twelve-month period,” and at least 30 days in any given taxable year to constitute a permanent establishment. Id. 142. Id. art. 5, ¶ 2(j). The services must continue for “more than 90 days in any twelvemonth period” and at least 30 days in any given taxable year to constitute a permanent establishment. Id. 143. Id. art. 5, ¶ 2(k). The equipment or machinery must be maintained “for a period of more than 120 consecutive days,” and at least 30 days in any given taxable year to constitute a permanent establishment. Id. 144. See, e.g., U.S.–India Treaty, supra note 4, art. 5, ¶ 2(j)-(l) (providing for the same concessions as in the U.S.–Jamaica Treaty, supra note 4). Similar deviations are also in U.S. tax treaties with other developed countries. See, e.g., U.S.-Canada Treaty, supra note 97, art. V, ¶ 4 (providing that the use of a drilling rig or ship for more than three months in any twelve-month period constitutes a permanent establishment). Because Canada is a developed country, the Senate Report does not mention the UN Model as a source of consultation, and the Joint Committee does not identify the deviation as a concession by the U.S., but rather explains that “[t]he shorter period was included in the treaty at the insistence of Canada which felt that a one-year period was unrealistic, given the adverse conditions of drilling in the Canadian offshore and the fact that the drilling season there is very short.” See Staff of S. Comm. on Foreign Relations, 96th Cong., Explanation of Proposed Income Tax Treaty Between the United States and Canada (Comm. Print 1981); Staff of the Joint Comm. on Taxation, 96th Cong., Explanation of Tax Convention with Canada (Comm. Print 1980). Narrow thresholds continue to appear in newly-signed U.S. tax treaties, such as the one with Bangladesh. See supra note 42, art. 5, ¶ 3, 6 (not yet in force).

598 allison christians

income in excess of the maximum rates provided in the U.S. Model. The U.S. Model allows source-country tax rates of no more than 5 and 15% on direct and regular dividends, respectively, and 0% on interest and royalties.145 In contrast, the U.S.–Jamaica treaty provides for source-country tax rates of 10% on direct dividends,146 15% on regular dividends,147 12.5% on interest,148 and 10% on royalties.149 Despite the general trend of higher source-country taxation of passive income items in U.S. tax treaties with LDCs, source-country taxation of certain items of passive income have recently been lowered in a number of U.S. tax treaties, including one with Mexico, arguably an LDC.150 The United States agreed to eliminate source-country taxation on direct dividends paid with respect to stock held by foreign controlling parent companies151 in a recent protocol to the U.S.– Mexico tax treaty.152 A most favored nation provision in the original treaty153 caused the elimination of source-country taxes on these direct dividends when the United States negotiated the same provision in recent treaties and protocols

145. See supra notes 92–93 and accompanying text; U.S. Model, supra note 56, arts. 10-12. The OECD Model differs from the U.S. Model in that it provides for source-country tax rates of 5% in the case of dividends held by 25% or greater corporate shareholders, 15% in the case of all other dividends, 10% in the case of interest, and zero in the case of royalties. OECD Model, supra note 21, arts. 10–12. As discussed in Part II.C, the UN Model leaves the maximum tax rate blank, implying that countries should negotiate a higher rate in the case of treaties between developed and less developed countries. UN Model, supra note 78, arts. 10–12. 146. U.S.–Jamaica Treaty, supra note 138, art. X, ¶ 2(a). 147. Id. art. X, ¶ 2(b). 148. Id. art. XI. 149. Id. art. XII. 150. Second Additional Protocol that Modifies the Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.-Mex., art II Nov. 26, 2002, T.I.A.S No. 108-4 (providing a zero-rate for dividends in the case of certain controlled companies), available at http://www.ustreas.gov/press/ releases/docs/mexico.pdf [hereinafter U.S.-Mex.]. Mexico is not designated as an LDC in the World Factbook, but is included by reference to its OECD membership within the definition of developed countries, even though its per capita GDP of less than $10,000 would align it with other LDCs. See supra note 20. 151. Those owning at least 80% of the foreign subsidiary’s stock. Id. at art. II, § 3(a). 152. See id. at art. II. See also Report of the Senate Foreign Relations Committee on the Additional Protocol Modifying the Income Tax Convention with Mexico, U.S.-Mex., § VI(A), Mar. 13, 2003, S. Exec. Doc. No. 108-4 (2003) (protocols eliminate tax on certain direct dividends). 153. See Protocol Amending the U.S.–Mexico Treaty, supra note 153, ¶ 8(b) (“If the United States agrees in a treaty with another country to impose a lower rate on dividends than the rate specified . . . both Contracting States shall apply that lower rate instead of the rate specified . . .”).

tax treaties for investment and aid to sub-saharan africa: a case study 599

with Australia,154 Japan,155 and Britain.156 According to Treasury officials, the elimination of source-country tax on direct dividends earned by foreign controlling companies reduces tax barriers and increases the economic ties between the partner countries.157 Following the logic of this position, a U.S.–Ghana tax treaty should involve a significant lowering, if not complete elimination, of sourcecountry taxation of dividends. The fact that the U.S. tax treaty with Mexico very recently adopted this position would seem to support the expectation of a similar provision in a tax treaty with Ghana. However, the more likely result is that in a U.S.–Ghana tax treaty, sourcecountry tax rates on dividends would be closer to the rates found in the Jamaica treaty than those found in the Mexico treaty.158 No recent U.S. tax treaty with an LDC has incorporated a zero rate for dividends paid to controlling company shareholders, and all provide for maximum source-country tax rates on passive income items that are higher than those provided in the U.S. Model.159

154. Protocol Amending the Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.-Austl., art. VI, May 2003 (hereinafter Australia Protocol). 155. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Japan, art. XI, Nov. 6, 2003, T.I.A.S. No. 108-14 (2004) (hereinafter U.S.-Japan Treaty). 156. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and on Capital Gains, U.S.–U.K., art. X, July 24, 2001, available at http://www.ustreas.gov/offices/tax-policy/library/uktreaty.pdf [hereinafter U.S.–U.K. Treaty]. 157. See Staff of S. Foreign Relations Comm., 108th Cong., Report on the Convention with Japan (Comm. Print 2003) (noting that many bilateral tax treaties to which the United States is not a party eliminate taxes on direct dividends, that the EU’s Parent-Subsidiary Directive achieves the same result, and that the United States has signed treaty documents with the United Kingdom and Australia that include provisions similar to the one in the Mexico protocol); see also John W. Snow, U.S. Sec’y of the Treasury, Remarks at the U.S.– Japan Income Tax Treaty Signing Ceremony (Nov. 6, 2003), available at http://www.treas. gov/press/releases/js975.htm (stating that the new U.S.–Japan Treaty “will significantly reduce existing tax-related barriers to trade and investment between Japan and the United States” and “will foster still-closer economic ties” between the two countries). 158. The Mexico treaty now provides for a maximum of 5% source-country taxation on direct dividends, 10% on regular dividends, and 0% on direct dividends paid to foreign companies with a controlling interest in the paying company. See U.S.–Mexico Treaty, supra note 97, art. 10. 159. See, e.g., U.S.–Sri Lanka Treaty, supra note 4, arts. X-XII (providing maximum rates of 15% on all dividends and 10% on interest and royalties); U.S.–Bangladesh Treaty, supra note 42 (same rates as in U.S.-Sri Lanka Treaty). Other than the lower rates on dividends, the U.S.-Mexico Treaty is consistent with other tax treaties with LDCs in that it provides for maximum source-country tax rates of 15% on interest and 10% on royalties. See U.S.–Mexico Treaty, supra note 97.

600 allison christians

Thus, as in the case of the permanent establishment provisions, the higher source-country rates that are typical in U.S. tax treaties with LDCs would likely be suggested, negotiated, and agreed to in a U.S.–Ghana tax treaty.160 Using the U.S.–Jamaica treaty and other recent treaties with LDCs as a guide, a U.S.–Ghana tax treaty could be expected to provide maximum source-country tax rates of 10–15% on direct dividends, 10–15% on regular dividends,161 and 10% on interest and royalties. The narrower permanent establishment thresholds and higher source-country tax rates are expected in a U.S.-Ghana tax treaty because they continue to appear in other U.S. tax treaties with LDCs. They appear in these treaties because it is believed that they will provide some benefit to the governments of the LDCs entering into these agreements. Yet, the overriding purpose of these treaties is the same as that for treaties exclusively between developed countries: they are supposed to relieve double taxation and therefore increase cross-border investment between the partner countries. The next part explores the extent to which either of these goals is achieved under this hypothetical tax treaty between Ghana and the United States. C. The effect of a U.S.–Ghana tax treaty on potential U.S. investors Assuming that Ghana is otherwise a viable destination for U.S. investment as described above, a tax treaty between these two countries would theoretically complement U.S. investment interests as well as its trade and aid initiatives. However, this part demonstrates that in today’s global tax climate, a tax treaty that follows the international standards set forth in the model treaties will

160. The U.S. tax treaties with Greece (a developed country), the former countries of the U.S.S.R. (each a transition country), and Trinidad & Tobago (an LDC), each provide for a maximum 30% source-country tax rate for dividends, and those with Israel (a developed country), India, and the Philippines (each an LDC), provide a maximum 25% rate. See Convention for the Avoidance of Double Taxation and Prevention of Fiscal Evasion with Respect to Taxes on Income, Feb. 20, 1950, U.S.–Greece, 5 U.S.T. 47, TIAS 2902; TIAS; Convention on Matters of Taxation, Jun. 20, 1973, U.S.–U.S.S.R., TIAS 8225, 27 U.S.T. 1; U.S.–Trin. & Tobago Treaty, supra note 4; Convention with Respect to Taxes on Income, Nov. 20, 1975, U.S.–Isr., TIAS; U.S.–India Treaty, supra note 4; U.S.–Philippines Treaty, supra note 4. The newest U.S. tax treaty, with Sri Lanka (an LDC), provides for a 15% tax rate on all dividends. See Proposed Tax Treaties with Japan and Sri Lanka: Hearing Before the S. Comm. on Foreign Relations 108th Cong. 6 (2004) (testimony of the Staff of the Joint Comm. on Taxation). The Sri Lanka treaty was considered by the Senate in February, 2004 together with the U.S.–Japan Treaty, which provides for zero taxation on certain dividends paid to controlling shareholders. See id. at 18, 20. 161. 48 of the U.S. tax treaties currently in force provide a rate of 10–15% on regular dividends. Internal Revenue Serv., U.S. Dep’t of the Treasury, Publ’n 901, U.S. Tax Treaties 33–34 tbl. 1 (2004); U.S.–Sri Lanka Treaty at art. 10.

tax treaties for investment and aid to sub-saharan africa: a case study 601

likely be ineffective in achieving its goals as a result of several interrelated phenomena. First, the scope of tax treaties appears to be too narrow in the context of these LDCs. Second, double taxation appears to be disappearing in international transactions involving these LDCs as a result of the widespread reduction in taxation caused by global tax competition and an ever-increasing availability of opportunities to avoid and evade income taxation. Third, there may be little differential between tax treaties and statutory law in the LDCs of Sub-Saharan Africa. Fourth, tax treaties may have little impact on multinational investment behavior in the face of nontax characteristics of LDCs, such as inadequate infrastructure. Finally, tax treaties may offer little more than perception about the commercial and legal climate of a country for foreign investment. Because of the impact of each of these factors on global commercial activity, a tax treaty between Ghana and the United States would yield an insignificant impact on investment and trade between these two countries. 1. Noncomparable taxation The first phenomenon that tends to reduce the potential benefit of a tax treaty between the United States and Ghana is the fact that U.S. multinational companies are likely to face nonincome types of taxation in Ghana.162 Like many LDCs, Ghana relies on a broad range of taxes that are not relieved under tax treaties, including consumption, excise, and trade taxes.163 The reliance on trade and excise taxes is historical, arising out of practices that have since been abandoned in developed countries in favor of personal income taxation and, outside of the United States, consumption taxation, typically in the form of the value added tax (VAT).164 VATs are relatively new to LDCs, having been introduced in the 1970s and 1980s largely as part of tax reforms initiated by international monetary organizations as a condition of lending.165 Prior to the introduction of the VAT, many LDCs, including those in Sub-Saharan Africa, followed the customs of the

162. That is, if they face any taxation at all. See infra Part IV.B. 163. See, e.g., Guttentag, supra note 4, at 452 (“[W]e have noted a trend where developing countries question the desirability of maintaining high source based taxation, but need to find alternative sources of revenue . . . many of them rely to a lesser extent on OECD type tax systems . . . instead, there is a greater reliance on value added taxes and asset taxes”). 164. The shift from trade to income and consumption taxation in the U.S. and other developed countries is discussed in Weisman, supra note 41, at 14, 42, 44; William D. Samson, History of Taxation, in The International Taxation System 33–37 (Andrew Lymer & John Hasseldine eds., 2002); Reuven S. Avi-Yonah, Globalization, Tax Competition, and the Fiscal Crisis of the Welfare State, 113 Harv. L. Rev. 1573, 1576 (2000). 165. From 1950, when the VAT emerged in its modern form, until 1980, many countries shifted from consumption taxes to payroll (social security) taxes, and since 1980, many countries have begun to shift from personal income taxes to VAT. Ken Messere et al., Tax Policy: Theory and Practice in OECD Countries 28 (2003). See also Malcolm Gillis, Tax Reform and the Value-Added Tax: Indonesia, in World Tax Reform: Case Studies of

602 allison christians

developed world that were introduced under colonization and relied heavily on trade taxes for revenue.166 The increased focus on the VAT was part of an overall effort to gradually reduce and, eventually, completely eliminate taxes on international trade.167 In Ghana, the government introduced a 20% VAT in 1995 but quickly repealed it in the face of violent protests.168 After a lengthy educational campaign, the government reinstated the VAT in 1998, this time at 10%.169 Since then, the VAT has not led to a decrease in any other taxes. A decrease in international trade taxes (tariffs) and excise taxes was initially realized soon after introduction of the VAT, but this trend has since reversed itself, and tariffs are currently increasing as a percentage of total revenues collected.170 Moreover, a temporary rise in corporate income taxation that accompanied the introduction of the VAT appears to have leveled off and corporate tax rates are currently decreasing.171 As a result, the introduction of the VAT in Ghana has lead to an overall increase in taxes that are not addressed by treaties.172

Developed and Developing Countries 227, 227 (Michael J. Boskin & Charles E. McClure, Jr., eds., 1990); Stewart, supra note 11, at 169. 166. Vito Tanzi, Taxation in Developing Countries, in Tax Systems in North Africa and European Countries 1, 8-9 (Luigi Bernardi & Jeffrey Owens, eds., 1994) (discussing revenue composition in LDCs). In Sub-Saharan Africa trade taxes averaged about 27% of total revenues from 1994 to 1999. Percentages of revenues collected attributable to trade taxes ranged from 5% in Angola, to 49% in Uganda. Scott Riswold, IMF VAT Policy in Sub-Saharan Africa, WTD, Sep. 1, 2003, at 8. For a discussion of the impact of colonization on tax systems in LDCs, see Stewart, supra note 11, at 145. 167. Such efforts have been encouraged by international monetary organizations such as the International Monetary Fund (IMF) and the World Bank as part of an overall tax reform package introduced in various forms as a condition to ongoing lending arrangements. Stewart, supra note 11, at 170. 168. Ghana, Despite Its Successes, Is Swept by Anti-Tax Protests, N.Y. Times, May 23, 1995, at A6 (describing VAT-related riot that led to five deaths). 169. Miranda Stewart & Sunita Jogarajan, The International Monetary Fund and Tax Reform, 2 Brit. Tax Rev. 146, 155 (2004). 170. The remainder of Ghana’s tax revenue derives from excise taxes, mainly on petroleum. Institute of Statistical, Social and Economic Research, The State of the Ghanaian Economy in 2002, at 26, 28-29 tbl.2.4 (2003) (hereinafter State of the Ghanaian Economy). 171. Id. 172. See Reuven S. Avi-Yonah, From Income to Consumption Tax: Some International Implications, 33 San Diego L. Rev. 1329, 1350 (1996) (theorizing the obsolescence of the U.S. tax treaty network in the event the U.S. adopts a consumption tax and repeals the income tax, because “[f]undamentally, income tax conventions apply to taxes on ‘income and capital’”). There are some tax treaties that address consumption taxes, specifically VAT. In most countries, however, the VAT employed is destination-based, meaning that exports are exempt from VAT and imports are subject to VAT. As a result, double VAT is avoided to a certain extent without need for international agreement (some double

tax treaties for investment and aid to sub-saharan africa: a case study 603

Finally, investors are likely to encounter noncomparable taxation in Ghana as a result of government stake-holding in many formerly state-owned enterprises. For example, cocoa produced in Ghana is not subject to income taxation,173 but is subject to levy by the Ghana Cocoa Board, a monopsony for the international sale of Ghanaian cocoa products.174 Similarly, the government extracts mining profits by owning shares in all mining operations and requiring the payment of dividends on such shares.175 Thus, a focus on the VAT, income, international trade, and excise taxes in Ghana provides only an incomplete picture of the full burden of taxation imposed in this country. As treaties focus only on income taxation, they address taxation in LDCs to a very limited degree. 2. Decreasing global tax burdens As noncomparable taxation increases, income taxation is decreasing throughout the world. As a result, multinationals investing in LDCs may face little or no income taxation on their foreign earnings. First, taxation may be reduced or eliminated by residence countries pursuant to rules that provide assets in offshore companies an indefinite suspension (deferral) of residence-based taxation. Second, taxation may be reduced or eliminated by source countries pursuant to tax incentives that eliminate taxation for specified durations or perpetually. Third, taxation by both countries may be reduced or eliminated through strategies of tax avoidance and evasion. Finally, taxation by both countries may be reduced or eliminated pursuant to express efforts to do so by both taxing jurisdictions, usually through a tax treaty. The combination of reduction or elimination of taxation in both countries, whether express or not, leads to complete nontaxation176 of multinational activities. As discussed more fully below, the resulting lack of taxation obviates the need to pursue tax relief under treaty.

taxation will continue to occur to the extent there are varying definitions of exempted and included items). The inconsistency occurs to various degrees in every country that employs a VAT. Developed countries, however, continue to rely more heavily than LDCs on income taxation, which is relieved by, and therefore necessitates the continued existence of, tax treaties. 173. G.I.R.A., supra note 83, § 11(7) (“income from cocoa of a cocoa farmer is exempt from tax”). 174. Acting as the intermediary between farmers and the global market, the Ghana Cocoa Board has the “sole responsibility for the sale and export of Ghana cocoa beans,” and delivers only a fraction of realized proceeds to farmers, thus imposing a gross basis tax that currently approximates some 33%. See Ghana Cocoa Beans Production, Export And Prices, available at http://www.cocobod.gh/corp_div.cfm?BrandsID=13. See also State of the Ghanaian Economy, supra note 170, at 26. 175. See Thomas C. Wexler, Introduction to Mining in Ghana, The Mining Journal, Nov. 10, 1995, at 353 (discussing strict government controls over and rights in Ghanaian mining operations). 176. Sometimes called double non-taxation to indicate the coordinative effort that produces it.

604 allison christians

a. Reduced taxation through deferral. As discussed above, most developed countries impose taxation on a worldwide basis, yet most protect this right only with respect to certain items of income, allowing suspension of taxation on other items to continue indefinitely at the will of the shareholders.177 Thus, despite the support for the primacy of residence-based taxation that originally served as a major reason for entering into tax treaties,178 much residence-based taxation is undermined by the persistent allowance of deferral. Deferral is antithetical to residence-based taxation. By allowing it, nominally residence-based jurisdictions like the United States mirror source-based (or territorial) systems by effectively providing tax exemptions for foreign income.179 Deferral is defended on grounds of neutrality: it is argued that companies from residence-based countries like the United States face heavier global tax burdens than companies from territorial countries, when both operate in third countries that impose little or no source-based taxation. For example, it is suggested that U.S.-based multinational companies operating abroad may be subject to little source-based taxation as foreign countries compete to attract their investment by offering low tax burdens, but because of the U.S. system of worldwide taxation, the U.S.-based company is still subject to the higher U.S. domestic tax rates. In contrast, it is supposed that multinationals from territorial systems will have a tax advantage in the minimally-taxing foreign country because these companies can combine low taxation abroad with exemption at home.180 Based on this argument, deferral continues to be vigorously defended under principles of capital import neutrality,181 as requisite to allow U.S. companies to compete in low-tax countries against the multinational companies of territorial jurisdictions.182 That few multinational companies are actually residents of 177. See supra text accompanying note 35. 178. See supra text accompanying note 68. 179. See Peroni, supra note 65, at 987. Passive income items such as dividends, interest, and royalties, are generally not eligible for deferral and are therefore subject to current tax in the U.S. 180. See Roin, supra note 63, at 114 (citing deferral proponents who argue that “[a]ny businesses that Americans can successfully operate in low tax jurisdictions . . . foreign investors can carry on equally well [and that if deferral was ended] foreign investors would use their now unique tax advantage to overwhelm their American competitors, wherever located”). 181. See discussion of neutrality supra note 61 and accompanying text. 182. See, e.g., Mark Warren, Democrats Would Increase Taxes on Companies’ Income Earned Abroad Repealing the Deferral Rule: The Wrong Answer to U.S. Job Losses, 2004 WTD 88-16 (May 3, 2004) (arguing that some countries exempt the foreign earnings of their multinationals, U.S. companies would face a higher overall tax burden when operating in low-tax jurisdictions in the absence of deferral, and that U.S. companies “cannot be expected to succeed if they are handicapped by a 35% corporate-tax rate on their worldwide income”); National Foreign Trade Council, Inc., The NFTC Foreign Income Project: International Tax Policy for the 21st Century, 1999 WTD 58-37 (Mar. 25, 1999); Impact of

tax treaties for investment and aid to sub-saharan africa: a case study 605

purely territorial systems,183 and that deferral provides the equivalent of exemption for much of the foreign income earned by U.S. multinationals while simultaneously providing them with a competitive advantage over their domestic counterparts,184 appears to have little effect on the efforts of U.S. multinationals to preserve the deferral privilege.185 The effect of deferral is to increase the sensitivity of U.S. taxpayers to foreign tax rates, thus forcing source countries to continually lower their internal tax burdens so as to attract the ever more demanding foreign capital. Deferral thus causes tax competition, as any income taxation imposed by a source country, such as Ghana, subjects a potential foreign investor to a burden it could otherwise avoid.186 Elimination of competition and tax sensitivity could be achieved if

U.S. Tax Rules on International Competitiveness: Hearing Before the H. Comm. on Ways and Means, 106th Cong. 64 (1999) (statement of Fred F. Murray, Vice President for Tax Policy National Foreign Trade Council, Inc.), available at http://frwebgate.access.gpo.gov/cgibin/getdoc.cgi?dbname=106_house_hearings&docid=f:66775.pdf (arguing that “[i]f the local tax rate in the company of operation is less than the U.S. rate, . . . competitors will be more lightly taxed than their U.S.-based competition,” whether they are locally based or foreign, “unless their home countries impose a regime that is as broad as subpart F, and none have to date done so”). The argument is perhaps as old as U.S. taxation itself. In the newly independent United States, import duties were favored over export duties or other forms of taxation, because the imposition of either export duties or property taxes on farmers would equally increase the price of goods destined for export, thus serving to “enable others to undersell us abroad.” See United States in Cong. Assembled, Reply to the Rhode Island Objections, Touching Import Duties (1782), reprinted in 1 The Debates in the Several State Conventions on the Adoption of the Federal Constitution 100, 105 (Jonathon Elliot, ed. 1996). 183. For example, of the top 100 multinationals, eighteen are from generally territorial systems (one from Hong Kong, three from Switzerland, one from Malaysia, and thirteen from France). Since France imposes a form of worldwide taxation on low-taxed earnings of controlled foreign companies, even this number is an exaggeration. Other countries may impose worldwide income generally, but exempt the foreign income of their multinationals under treaty. See UNCTAD World 100 Non-Financial TNCs, supra note 82. 184. Domestic companies are subject to worldwide taxation and cannot generally opt to suspend the taxation of their profits. See generally Clifton Fleming Jr. et al., An Alternative View of Deferral: Considering a Proposal to Curtail, Not Expand, Deferral, 2000 WTD 20-15 (Jan. 31, 2000) (arguing that deferral is a subsidy for operating business abroad and that proponents of deferral “have not candidly acknowledged the broad nature of the scope of the existing deferral privilege”). 185. See supra text accompanying note 182. 186. Deferral removes the existing (residual) tax burden, thereby ensuring that any tax imposed by a foreign country is a tax wedge. In the absence of deferral, the tax wedge is created by the home country and, outside of limitations on foreign tax credits, taxes imposed by the source country do not increase the wedge. For a discussion of the interaction of deferral and the subsequent efforts of source countries to eliminate tax wedges, see Dagan, supra note 40, at 952–56.

606 allison christians

all countries adhered to principles of capital export neutrality. This would, however, require international coordination and cooperation to a degree that appears overwhelmingly unattainable.187 The consequence is that U.S. multinationals may generally avoid U.S. taxation on their foreign income by operating through subsidiary companies in source countries,188 which they generally do.189 As suspension and effective elimination of taxation on foreign income becomes the norm in the developed world, LDCs respond accordingly, by increasingly offering corresponding tax relief in the form of tax incentives. These incentives have become a standard tool for capturing a share of the global flow of foreign investment.190 b. Reduced taxation through tax incentives. Most countries use various forms of tax incentives to encourage particular behavior in taxpayers, and neither the United States nor Ghana is an exception. The U.S. employs numerous tax incentives to attract foreign investment and encourage domestic investment. These provisions are generally embedded in the tax base, rather than being reflected in 187. See Victor Thuronyi, International Tax Cooperation and a Multilateral Treaty, 26 Brook. J. Int’l L. 1641, 1642 n.8 (2001) (an internationally harmonized system is “too utopian to merit discussion”); Charles E. McLure, Jr., Tax Policies for the XXIst Century, in Visions of the Tax Systems of the XXIst Century 9, 28-29 (1997). But see Yariv Brauner, An International Tax Regime in Crystallization, 56 Tax L. Rev. 259, 260 (2003) (arguing that there has been a “modelization” of the international tax rules that could be built upon to achieve some measure of rule harmonization). Recent developments in the EU indicate that less, rather than more, cooperation is likely. See Joann M. Weiner, EU Governments Fear Increased Tax Competition in Wake of Accession, 2004 WTD 81-1 (Apr. 6, 2004); Joe Kirwin, International Taxes European Commission Rejects Effort For Harmonized Corporate Tax Rates, Daily Tax Rep., June 1, 2004, at G-8. 188. Stephen E. Shay, Exploring Alternatives to Subpart F, 82 Taxes 3-29, 31 (2004) (multinationals are free to choose to operate through a branch or subsidiary, and they will generally choose subsidiary form unless the foreign effective tax rate is greater than the U.S. rate or if they benefit from pooling high- and low-taxed earnings). 189. For example, several of the largest foreign investments in Ghana are U.S. controlled foreign corporations, including the Valco, Regimanuel Gray, and Equatorial Bottlers, discussed supra note 108. Operating through a domestic subsidiary is also more advantageous from a Ghanaian perspective, since foreign companies are subject to strict scrutiny from the taxing and regulatory authorities to an extent exceeding that paid to domestic companies. The differential treatment is especially acute in the case of mining and other extractive operations, which are strictly regulated and limited as to foreign ownership by the Government of Ghana. Interview with Bernard Ahafor, Attorney, in Ghana (Dec. 2, 2003). See also Shay, supra note 188, at 31. 190. The evidence is perhaps most obvious in regards to the number of countries offering tax holidays—over one hundred in 1998 and increasing—and the share of foreign investment directed at tax havens that are decried by the OECD for their harmful tax practices. While these countries command a fraction of the world’s population and its GDP, they attract a disproportionately large amount of U.S. foreign investment capital. See Avi-Yonah, supra note 164, at 1577, 1589, 1643.

tax treaties for investment and aid to sub-saharan africa: a case study 607

the tax rates.191 For example, along with the privilege of deferral, tax credits for research and development (R&D) and accelerated depreciation deductions are among the major tax incentives the United States offers.192 Ghana also offers accelerated depreciation deductions and R&D credits similar to—but perhaps not as generous as—those of the United States.193 However, most

191. Since the 1960s, an awareness of the danger of the hidden costs of such incentives has led to expenditure budgeting, which quantifies the cost of embedded provisions. For an example, see Analytical Perspectives, supra note 65, at 285 (explaining the concept of expenditures and providing a selected list). Incentives currently provided in the U.S. tax base include accelerated depreciation and exclusions from taxation for certain forms of income such as tax-exempt interest. Tax incentives include any exclusions or exemptions that reduce or defer the tax base. See generally Alex Easson & Eric M. Zolt, Tax Incentives, 2002 World Bank Inst. 3 (“[t]ax incentives can take the form of tax holidays for a limited duration, current deductibility for certain types of expenditures, or reduced import tariffs or customs”). Ireland and Belgium, which offer low rates for foreign investors, are exceptions (and a source of consternation to their OECD counterparts) to the general rule of tax base rather than tax rate concessions in developed countries. See, e.g., Avi-Yonah, supra note 164, at 1601. 192. Congress first provided a deduction for research and experimental expenditures in 1981, because it saw a decline in research activities it attributed to inadequacies in the I.R.C. § 174 deduction, which at that time only applied to investment in machinery and equipment employed in research or experimental activities. Congress concluded that “[i]n order to reverse this decline in research spending . . . a substantial tax credit for incremental research and experimental expenditures was needed.” Staff of the J. Comm. on Tax’n, 97th Cong., General Explanation of the Economic Recovery Tax Act of 1981, reprinted in Internal Revenue Acts, 1980–1981, at 1369, 1494 (1982). In the same act, Congress provided for accelerated depreciation deduction allowances because the existing depreciation deduction allowances “did not provide the investment stimulus that was felt to be essential for economic expansion.” Id. at 1449. Enhanced bonus depreciation provisions were enacted in 2001 under the theory that “allowing additional first-year depreciation will accelerate purchases of equipment, promote capital investment, modernization, and growth, and will help to spur an economic recovery.” H.R. Rep. No. 107-251, pt. 2, at 20 (2001). Bonus depreciation was expanded in 2003 for the same reason. H.R. Rep. No. 108-94, pt. 2, at 23 (2003) (“increasing and extending the additional first-year depreciation will accelerate purchases of equipment, promote capital investment, modernization, and growth, . . . help to spur an economic recovery, . . . [and] increase employment opportunities in the years ahead). See also Richard E. Andersen, IRS Relaxes Rules for Research Credit; Opportunities for R&D-Intensive Multinationals?, 4 J. Tax’n. Global Transactions 17 (CCH) (Spring, 2004) (discussing structures with which foreign and domestic multinationals can use R&D credits to generate tax-free profits in the U.S., and citing a 2003 study by Bain & Co., entitled Addressing the Innovation Divide, in which it was found that in the past decade, European drug makers placed their R&D in the United States versus in local expansion by a two-to-one margin). 193. G.I.R.A., supra note 83, Third Schedule (depreciation allowance), § 19 (deductions for research and development expenditures).

608 allison christians

LDCs, including Ghana, also offer significantly more generous incentives in the form of low corporate tax rates and myriad tax exemptions.194 Ghana imposes only an 8% tax on income from the export of most goods, rates ranging from 16 to 25% for certain industries and businesses conducted in certain geographic areas, and complete exemption from taxation (tax holidays) for periods ranging from three to ten years for new activities conducted in certain industries or geographic areas.195 Many LDCs, including Ghana, have also set aside geographic areas as havens from the normal tax and regulatory regimes (free zones), specifically to host manufacturing and processing plants. In Ghana’s free zone, established in 1995, companies enjoy a ten year tax holiday followed by tax rates never to exceed 8%.196 International organizations such as the World Bank and the IMF currently decry the harm that tax holidays cause in depriving LDCs of much-needed revenue.197 The elimination of income taxation on corporate taxpayers, coupled with the pressure to reduce taxes on international trade, has created critical revenue

194. For example, by 1998, over 100 countries had tax holidays. Avi-Yonah, supra note 164, at 1577. See, e.g., Zmarak Shalizi, Lessons of Tax Reform 23 (1991). 195. G.I.R.A., supra note 83, §11 (Industry Concessions) & First Schedule, Part II (Rates of Income Tax Upon Companies). Although tax holidays are limited in duration, insufficient enforcement prevents the IRS from curbing instances in which companies facing expiring tax holidays simply dissolve and reincorporate under a different name to restart the clock. Interview with Kweku Ackaah-Boafo, Esq. (Feb. 6, 2004) (Discussing Canadian Bogosu Resources, a mining company operating in Ghana, which reincorporated as Billington Bogusu Gold Limited and again five years later as Bogusu Gold Limited, in order to avail itself of tax benefits that otherwise would have expired). 196. G.I.R.A., supra note 83, First Schedule. 197. See, e.g., Janet Stotsky, Summary of IMF Tax Policy Advice, in Tax Policy Handbook 279, 282 (Parthasarathi Shome, ed., International Monetary Fund 1995) (stating that tax incentives “have proved to be largely ineffective, while causing serious distortions and inequities in corporate taxation.”); Shalizi, supra note 194, at 60 (“The use of so-called tax expenditures—tax preferences and exemptions to promote specific economic and social objectives—should, in general, be deemphasized”). This is a reversal of position for the World Bank, which at one point encouraged LDCs to offer tax incentives to attract foreign investment and was concerned with the effect elimination of tax incentives might have on its assistance projects. Stewart, supra note 11 at 169; Shalizi, supra note 194, at 68–69. The World Bank has since “recommended the removal or tightening of incentives in Argentina (1989), Bangladesh (1989), Brazil (1989), Ghana (1989), and Turkey (1987), among others.” Shalizi, supra note 195, at 69. Tax incentives are also contrary to WTO rules prohibiting tax subsidies. See WTO Agreement on Subsidies and Countervailing Measures, Apr. 14, 1994, Annex 1A, Art. 1, ¶ 1.1. These provisions, however, are rarely enforced with respect to LDCs. See Reuven S. Avi-Yonah & Martin B. Tittle, Foreign Direct Investment in Latin America: Overview and Current Status 26–28 (2002), available at http:// www.iadb.org/INT/Trade/1_english/2_WhatWeDo/Documents/d_TaxDocs/20022003/a_Foreign%20Direct%20Investment%20in%20Latin%20America.pdf.

tax treaties for investment and aid to sub-saharan africa: a case study 609

shortfalls in many countries.198 Nevertheless, new tax incentives continue to be introduced in both developed and less developed countries around the world,199 often in response to private sector lobbying.200 Some recent examples include the introduction of a free zone in the United Arab Emirates,201 a five-year exemption period for audit, accounting, and law firms in Singapore,202 and a ten-year corporate tax holiday for income from investments of at least 150 million in Turkey.203 As a result of these kinds of initiatives, U.S. multinationals may face little or no income taxation on income derived in LDCs. The impact of tax treaties on activities giving rise to such income is therefore minimized, as double taxation, and even single taxation, is avoided through unilateral tax rules. However, even if home or source countries nominally impose taxation on multinationals, widespread tax avoidance and evasion neutralizes these taxes. Tax treaties appear to have little effect in these circumstances. c. Reduced taxation through tax avoidance and evasion. In the event that deferral or tax incentives are not available, multinational companies manage their worldwide tax exposure by using tax planning techniques to shift income to lowor no-tax jurisdictions through earnings stripping, transfer pricing, thin capitalization, and similar means of tax avoidance and, in the extreme, tax evasion.204

198. Cordia Scott & Sirena J. Scales, Tax Competition Harms Developing Countries, IMF Official Says, 2003 WTD 238-9 (Dec. 10, 2003). 199. See, e.g., Kwang-Yeol Yoo, Corporate Taxation of Foreign Direct Investment Income 1991-2001 (Econ. Dep’t, Working Paper No. 365, 2003), available at http://www.olis.oecd. org/olis/2003doc.nsf/43bb6130e5e86e5fc12569fa005d004c/48ae491b8e2db4a9c1256d 8e003b567f/$FILE/JT00148239.PDF. 200. See, e.g., David Roberto R. Soares da Silva, Tech Companies in Brazil Seek Tax Incentives to Promote R&D, 2004 WTD 138-6 (Jul. 19, 2004) (domestic and multinational technology companies are currently lobbying for a three-year exemption from federal taxes for income from sales of “all new products that contain significant technological innovation”). 201. Under this new initiative, free-zone companies in Dubai will be exempt from income tax. See Cordia Scott, Dubai Woos Europe With Tax-Free Outsourcing Zone, 2004 WTD 118-12 (June 17, 2004). 202. Lisa J. Bender, Singapore Launches Tax Incentives for Audit, Accounting, Law Firms, 2004 WTD 66-5 (Apr. 5, 2004). 203. Mustafa Çamlica, Turkey Plans Tax Holidays for Large Investments, 2004 WTD 82-8 (Apr. 28, 2004). 204. See Reuven S. Avi-Yonah, The Rise and Fall of Arm’s Length: A Study in the Evolution of U.S. International Taxation, 15 Va. Tax Rev. 89, 95 (1995) (“Transfer pricing manipulation is one of the simplest ways to avoid taxation.”); David Harris, Randall Morck, Joel Slemrod & Bernard Yeung, Income Shifting in U.S. Multinational Corporations, in Studies in International Taxation 277, 301 (Giovannini et al. eds., 1993); James R. Hines, Jr., Tax Policy and the Activities of Multinational Corporations, in Fiscal Policy: Lessons from Economic Research 401, 414–15 (Alan J. Auerbach ed., 1997). The line between tax avoidance and tax

610 allison christians

For example, U.S. multinationals typically use over- and under-invoicing to assign foreign profits to subsidiaries in tax havens.205 As a result, firms can increasingly make physical location decisions that are largely independent of tax-related business decisions, shifting profits to the most advantageous tax destination.206 Efforts by governments to curb such practices are abundant207 but largely ineffective208 in the face of efforts by taxpayers to engage in them.209

evasion is murky. Tax avoidance generally refers to lawful attempts to minimize taxation, as Judge Learned Hand famously noted in Comm’r v. Newman, 159 F.2d 848, 850–51 (2d Cir. 1947) (Hand, J., dissenting) (“Over and over again courts have said that there is nothing sinister in so arranging one’s affairs as to keep taxes as low as possible. Everybody does so, rich or poor; and all do right, for nobody owes any public duty to pay more than the law demands . . . ”). Tax evasion generally encompasses the unlawful and fraudulent avoidance of tax accomplished by hiding taxable income and assets from taxing authorities. 205. See Council of the European Union, Final Draft Report of the Ad hoc Working Party on Tax Fraud 16–17 (Brussels, April 27, 2000). Direct tax fraud is typically committed through false invoicing, under- and over-invoicing, non-declaration of income earned in foreign jurisdictions, and “use by taxpayers of a fictitious tax domicile, with the purpose to evade fulfilling their tax obligations in their country of domicile for tax purposes.” Id. at 4–5. See also Martin A. Sullivan, U.S. Multinationals Move More Profits to Tax Havens, 2004 WTD 31-4 (Feb. 9, 2004) (although they comprise just 13% of productive capacity and 9% of employment, subsidiaries of U.S. multinationals located in the top eleven tax havens were assigned 46.3% of foreign profits in 2001); Hooke, supra note 124, at 86–87 (suggesting that to control costs, it is “sound operating procedure” for a foreign investor of an export platform in an LDC to interpose an offshore bank, and overcharge the foreign company for imported supplies and management fees to reduce income in the source country). 206. See Christoph Spengel & Anne Schäfer, International Tax Planning in the Age of ICT, ZEW Discussion Paper No. 04–27 (2004), available at http://ssrn.com/abstract= 552061 (arguing that information and commercial technology makes geographic distance “less relevant” and allows companies to choose location and form of investment on the basis of international tax differentials). 207. Transfer pricing rules are a common feature in the tax systems of most countries, as are rules denying deductions for interest and royalties in certain cases and rules requiring a certain combination of debt and equity (thin capitalization rules). See Hugh J. Ault and Brian J. Arnold, Comparative Income Taxation: A Structural Analysis 420–28 (1997). 208. In the U.S., the transfer pricing rules are long and complicated and constantly evolving, but still considered inadequate in preventing profit-shifting, as are U.S. earnings- and interest-stripping rules. See, e.g., I.R.C. § 163(j) (2005). These rules are essentially thin capitalization rules, each of which are similarly limited in their success in curbing avoidance of U.S. taxation. For an overview of U.S. efforts to control transfer pricing, see generally Avi-Yonah, supra note 204. For a recent example of the failure of interest stripping rules, consider the growing use of Canadian Income Funds to avoid the application of I.R.C. § 163(j) (2000). See, e.g., Jack Bernstein & Barbara Worndl, Canadian-U.S. Cross-Border Income Trusts: New Variations, 34 Tax Notes Int’l 281, 283 (April 19, 2004). 209. See, e.g., Shay, Exploring Alternatives, supra note 188, at 36 (“The drive on the part of taxpayers, multinational and others, to push down effective tax rates has accelerated in recent years”).

tax treaties for investment and aid to sub-saharan africa: a case study 611

In LDCs such as Ghana, where enforcement of the tax law has been relatively less of a focus than reform of the tax law, tax authorities are all but helpless against these practices.210 It is popularly said that Ghanaian companies keep three sets of books: one for the banks, showing large profits so as to secure financing; one for the Ghanaian Internal Revenue Service (IRS), showing large losses so as to avoid paying taxes; and one set, very closely-guarded by the owners, which contains the most accurate information.211 There is no official data available regarding whether, and to what extent, U.S. multinationals take advantage of enforcement weaknesses.212 Ghana recently introduced a Large Taxpayers Unit to curtail tax evasion, but the Ghanaian IRS relies on the good faith of company officials and their independent auditors because the resources are lacking to perform audits on all but a few companies.213 Given that the overall tax

210. See Stewart, supra note 11, at 181. 211. Interview with Margaret K. Insaidoo, Justice, High Court of Ghana, in Ghana (Dec. 9, 2003) (on file with author); Interview with Bernard Ahafor, Attorney, Private Practice, in Ghana (Dec. 2, 2003) (on file with author); Interview with Sefah Ayebeng, Chief Inspector of Taxes, Internal Revenue Service, in Ghana (Dec. 11, 2003) (on file with author). The implication is that firms keep separate books in an attempt to defraud the government, rather than in the ordinary course of keeping separate tax and cost accounting books, for which there is generally no statutory proscription. See, e.g., Charles E. Hyde & Chongwoo Choe, Keeping Two Sets of Books: The Relationship Between Tax & Incentive Transfer Prices, http://ssrn.com/abstract=522623 (arguing that keeping two sets of books with respect to transfer pricing is “not only legal but also typically desirable” for many MNEs). 212. Anecdotal evidence that multinationals are thought to evade taxation where possible is not lacking, however. See, e.g., Sirena J. Scales, Venezuela Temporarily Closes McDonald’s Nationwide, 2005 WTD 26-11 (Feb. 9, 2005) (“Venezuela’s Tax Agency (SENIAT) has temporarily closed all 80 McDonald’s restaurants in the nation, citing failure to comply with tax rules . . . ”). 213. Seth E. Terkper, Ghana Establishes Long-Awaited Large Taxpayer Unit, 2004 WTD 64-10 (Apr. 2, 2004). A mid-size taxpayers unit is also in the planning stages. Ayebeng, supra note 211. A more effective audit process may not be sufficient to induce increased compliance, however. A recent empirical study about Australian investors that were accused of engaging in abusive tax transactions argues that taxpayers’ level of trust regarding the fairness, neutrality, and respect accorded to them by the revenue authorities was correlated to their level of voluntary compliance, and that although trust alone should not be relied upon in enforcing a tax system, “a regulatory strategy that combines a preference for trust with an ability to switch to a policy of distrust is therefore likely to be the most effective.” Kristina Murphy, The Role of Trust in Nurturing Compliance: A Study of Accused Tax Avoiders, 28 Law and Human Behavior 2, 187, 203 (2004). In an interesting twist, South Korea recently announced that domestic and foreign companies meeting target job creation goals will be free from audits in 2004 and 2005 under a new tax incentive program. James Lim, South Korea Offering Companies that Create Jobs Shield from Audits, 34 Daily Tax Report (Feb. 23, 2004), at G-3.

612 allison christians

compliance rate is estimated to be less than 20% in Ghana, good faith appears to be rather elusive.214 As a consequence of tax avoidance and evasion strategies, income is often exempt from taxation even if tax nominally applies in the residence country, the source country, or both. In such a taxing environment there is little taxation, let alone double taxation, to be relieved by treaty. Governments are not unaware of the problem. Tax avoidance and evasion has typically been addressed in treaties through information sharing provisions, in which the respective taxing jurisdictions agree to assist each other in collecting revenues.215 These provisions have a (perhaps unintended) consequence, however. Introduction of a tax treaty may decrease investment, as investors seek to avoid the implementation of the information sharing provisions that have become standard in tax agreements.216 The intersection of the taxation of portfolio interest and U.S. interest reporting rules illustrates this tension. The United States is a potential tax haven for foreign investors because of its zero tax on portfolio interest and rules under which banks are generally not required to report interest payments made to nonresident aliens.217 Efforts to require interest payment reporting have consistently 214. The compliance rate is an estimate of Ghanaian IRS officials and not an official government statistic. Ayebeng, supra note 211 (estimating compliance at less than 20%); Interview with Fred Ajyarkwa, Official, Internal Revenue Service, in Ghana (Dec. 11, 2003) (estimating it at 17%). 215. The U.S. Model requires contracting states to exchange all relevant information to carry out the provisions of the tax treaty or the domestic laws of the states concerning taxes covered by the treaty, including assessment, collection, enforcement, and prosecution regarding taxes covered by the convention. See U.S. Model, supra note 56, at art. 26, ¶ 1. It also calls for treaty override of domestic bank secrecy or privacy laws. The OECD Model does not include the assessment/collection language but extends the scope of taxes to “every kind and description imposed on behalf of the contracting states.” See OECD Model, supra note 21, art. 26, ¶ 1. It does not include an equivalent to the U.S. Model’s secrecy law override. The UN Model limits assistance to taxes covered by the Convention as in the U.S. Model, and explicitly adds that information exchange is intended to prevent fraud or evasion of taxes. See UN Model, supra note 78, art. 26, ¶ 1. 216. Bruce A. Blonigen & Ronald B. Davies, The Effects of Bilateral Tax Treaties on U.S. FDI Activity (Nat’l Bureau of Econ. Research, Working Paper No. 8834, 2002), available at http://www.nber.org/papers/w8834 (showing a decrease in foreign investment upon the introduction of a tax treaty and suggesting that such decrease may be the result of the dampening effect tax treaties may have on tax evasion due to information sharing provisions); Ronald B. Davies, Tax Treaties, Renegotiations, and Foreign Direct Investment (University of Oregon Economics, Working Paper No. 2003-14, 2003), available at http:// papers.ssrn.com/sol3/papers.cfm?abstract_id=436502 (“[t]reaties have either a zero or even a negative effect on FDI” because they dampen the ability of businesses to engage in tax evasion activities, especially through transfer pricing). 217. I.R.C. §§ 871(h), 882(a), (c) (2005); Treas. Reg. § 1.6049-5 (1983). Canadian residents are a current exception to the interest reporting rules, Treas. Reg. § 1.6049-8(a) (1996), and proposed regulations would extend the reporting requirements to include all

tax treaties for investment and aid to sub-saharan africa: a case study 613

met strong resistance by the private sector, which argues that such rules would “hinder tax competition between nations” and “undermine [the] global shift to lower tax rates and international tax reform.”218 Several members of Congress echo these sentiments, arguing that expanded reporting rules “would likely result in the flight of hundreds of billions of dollars from U.S. financial institutions” and could cause “serious, irreparable harm to the U.S. economy.”219 The implication is that while the United States does not condone tax evasion, there has emerged no political will strong enough to counter the private interests benefiting from the rules as they currently exist.220 Similar sentiments may exist in the context of tax treaties, especially when the partner country, as in the case of Ghana, has a very limited ability to enforce the tax laws prior to the introduction of a treaty. If foreign investors are able to avoid taxation in Ghana, for instance through aggressive tax planning, a tax treaty that requires or permits Ghana to provide tax information to the U.S. taxing authority may not be welcome.221 d. Reduced taxation through tax sparing. The proliferation of tax incentives and tax holidays in LDCs, coupled with deferral in the United States and tax avoidance opportunities in both countries, limits the need for tax treaties to relieve double taxation. Since the 1950s, tax sparing has been promoted as a way to use tax treaties to increase investment to targeted LDCs, even in the absence of double taxation.222 Tax sparing prevents residence-country taxation of income

interest over $10 paid to any non-resident alien individual. Prop. Treas. Reg. § 1.60498(a), 67 Fed. Reg. 50389 (Aug. 2, 2002). 218. Katherine M. Stimmel, Free Market Interest Groups Urge Treasury to Withdraw Alien Interest Reporting Rules, 16 Daily Tax Rep. (BNA), at G-2 (Jan. 27, 2004). 219. Alison Bennett, House Lawmakers Ask Bush to Withdraw IRS Interest Reporting Rules for Aliens, 69 Daily Tax Rep. (BNA), at G-8 (Apr. 10, 2002). See also Sen. Gordon Smith (R-Ore.), Letter on Proposed Nonresident Alien Interest Reporting Rules (REG133254-02) to Treasury Secretary John Snow, TaxCore (BNA) (Feb. 20, 2003) (urging Treasury not to move forward with interest reporting rules because it “would drive the savings of foreigners out of bank accounts in the United States and into bank accounts in other nations,” and expressing the Senator’s failure to understand “why we put the enforcement of other nations’ tax laws as a priority at Treasury”). 220. Perhaps recent efforts to create a multinational task force to combat abusive taxavoidance can provide the pressure needed to reform this long standing impasse. See Sirena J. Scales, Multination Task Force Created to Combat Abusive Tax Avoidance, 2004 TNT 81-4 (Apr. 26, 2004). 221. Moreover, to the extent that a U.S. tax treaty coordinates transfer pricing rules, a treaty might increase the taxation of a multinational that could otherwise benefit from conflicting domestic standards. 222. See generally OECD, Tax Sparing: A Reconsideration (1998) (hereinafter Tax Sparing). Recent literature on tax sparing includes Brown, supra note 7; Damian Laurey, Reexamining U.S. Tax Sparing Policy with Developing Countries: The Merits of Falling in Line with International Norms, 20 Va. Tax Rev. 467, 483 (2000) (arguing that LDCs “need tax

614 allison christians

exempted from tax by source countries,223 by providing that if a source country refrains from taxing income derived in its jurisdiction (usually pursuant to a tax holiday), the residence country nevertheless grants a tax credit for the nominally imposed tax.224 Thus, under tax sparing, two taxing jurisdictions cooperate to exempt multinational companies from income taxation in both countries. Although similar effects could be accomplished unilaterally by residence countries,225 tax sparing is generally seen as a mechanism that should be offered in the context of a tax treaty, as a measure to encourage foreign investment to selected LDCs.226 Tax sparing has particularly been promoted as a vehicle for investment and aid to the nations of Sub-Saharan Africa.227

holidays to attract foreign investment,” and therefore tax sparing is requisite to counter the effect of residual home country taxation under tax treaties). Tax sparing is also defended as justifiable on grounds of capital import neutrality, on the basis that it allows American multinationals to compete with companies from other exemption-providing countries in the global marketplace. See discussion infra notes 225–26 and accompanying text. However, tax sparing violates the concept of capital export neutrality, and has been consistently rejected by the Treasury Department on the grounds that tax treaties are supposed to relieve double taxation, not eliminate taxation altogether, and that tax treaties are not meant to provide benefits to U.S. persons. 223. Tax sparing was first introduced in the United Kingdom by the British Royal Commission, which prepared a report in 1953 recommending tax sparing as a means of “aiding British investment abroad.” Tax Sparing, supra note 222, at 15. Rejected by the United Kingdom in 1957 after several years of debate, tax sparing was enabled in U.K. tax treaties as a result of legislative action in 1961. Id. The purpose of the legislation was “enabling the UK to give relief to developing countries for taxes spared under foreign incentive programmes.” Id. 224. Many examples and explanations of tax sparing exist. For an overview of tax sparing, see generally Richard D. Kuhn, United States Tax Policy with Respect to Less Developed Countries, 32 Geo. Wash. L. Rev. 261 (1963). 225. For example, the U.S. could expand the definition of a creditable tax to include certain nominally-imposed taxes. See, e.g., Paul R. McDaniel, The U.S. Tax Treatment of Foreign Source Income Earned in Developing Countries: A Policy Analysis, 35 Geo. Wash. Int’l L. Rev. 265, 268–69 (2003). 226. See, e.g., Brown, supra note 7; Laurey, supra note 222 (suggesting proposals regarding the use of tax treaties to implement foreign aid initiatives by encouraging foreign investment through tax sparing). See also J. Clifton Fleming, Jr., Robert J. Peroni & Stephen E. Shay, Fairness in International Taxation: The Ability-to-Pay Case for Taxing Worldwide Income, 5 Fla. Tax Rev. 299, 347 (2001) (suggesting that limiting tax sparing to its use in tax treaties “would allow appropriate distinctions to be made among nations and would assist the United States in negotiating appropriate reciprocal tax concessions for its residents”). 227. Brown, supra note 7, at 83 (arguing for tax sparing in tax treaties specifically with Sub-Saharan Africa).

tax treaties for investment and aid to sub-saharan africa: a case study 615

There is little evidence, however, that tax sparing increases foreign investment.228 On the contrary, tax sparing could potentially decrease investment in LDCs, since it enables foreign investors to repatriate earnings that they would otherwise leave abroad under the protection of deferral.229 As such, tax sparing appears fundamentally inconsistent with the goal of using tax treaties to increase investment flows from developed to less developed countries. Moreover, tax sparing increases tax competition by creating an additional disadvantage for countries that do not have tax holidays, while leaving countries that have a tax holiday in effect in the same or worse position as they were when only deferral was available.230 The OECD has initiated efforts to combat what it terms “harmful tax practices”—in essence, any tax regime that undermines residence-based taxation by providing tax breaks and refusing to cooperate in information sharing.231 Persisting in the allowance of deferral and tax holidays while promoting tax sparing seems equally inconsistent with the treaty-related goal of protecting residence-based tax bases. Foreseeing that the ratification of any treaty with tax sparing would prompt a surge of lobbying by U.S. multinationals seeking the expansion of such provisions to other countries, the United States has been unequivocal in its rejection

228. See McDaniel, supra note 225, at 284 (providing an overview of the conflicting economic literature regarding the interaction of tax sparing and FDI). 229. See, e.g., Peroni, supra note 67, at 469 (deferral encourages “[r]etention and reinvestment of earnings by [foreign companies]”); see also Laurey, supra note 222, at 484–85 (tax sparing would “allow U.S. multinationals to repatriate earnings based on business needs instead of on adverse tax consequences”). In a 2003 study of the annual filings of the companies in the S&P 500, it was found that such companies had accumulated over $500 billion in un-repatriated foreign earnings. Anne Swope, Bruce Kasman & Robert Mellman, Bringing It All Back Home: Repatriation Legislation’s Final Lap (2004), http:// www.morganmarkets.com. This figure represents a trend of ever-increasing “trapped” foreign profits. Conversely, by acting as an incentive to repatriate capital, tax sparing may be advantageous to the U.S. economy even though it has long been rejected for policy reasons. For example, in the context of the repeal of I.R.C. § 114 (a tax exemption for certain foreign earnings that was found to be in violation of WTO standards), legislators enacted a temporary reduction in the rate of tax imposed on repatriated profits, citing in support the need to direct capital back to the U.S. in the quest to create jobs and boost the economy. See American Jobs Creation Act of 2004, P.L. 108–357, H.R. 4250, Sec. 101(a) (repealing I.R.C. § 114) and Sec. 965 (enacting temporary dividends-received deduction). See also Staff of the Cmtee on Ways and Means, 108th Cong., Short Summary of Conference Report 108-755 2 (October 7, 2004), available at http://waysandmeans.house.gov/legis. asp?formmode=read&id=2117 (last visited Nov., 2005) (Sec. 956 “Encourages companies to reinvest foreign earnings in the United States”). 230. See, e.g., Margalioth, supra note 82, at 198. 231. See generally Staff of the OECD Fiscal Affairs Comm., The OECD’s Project on Harmful Tax Practices: The 2004 Progress Report (2004).

616 allison christians

of these provisions.232 While the potentially negative impact on investment in LDCs is one valid reason why tax sparing should continue to be rejected, the primary position of the United States has been that tax sparing inappropriately allows the reduction of U.S. taxation of U.S. persons, a result specifically precluded by all U.S. treaties currently in force.233 Some LDCs, notably those in Latin America, have terminated tax treaty negotiations with the United States over the issue of tax sparing.234 Still, the U.S. position on tax sparing is only “one of several obstacles in the way of U.S.-developing country tax treaties.”235 In fact, tax sparing is largely unnecessary in the quest for complete nontaxation. As discussed above, tax holidays granted by LDCs to investors from deferral-granting countries, such as the United States, are effective in providing double nontaxation so long as capital is reinvested rather than repatriated.236 3. Domestic tax rates equal to or better than treaty rates In treaties between developed countries, domestic tax regimes are often significantly different than treaty-based tax regimes.237 This is especially the case with respect to tax rates on passive income paid to foreign persons, which are typically much higher under domestic statutes than under tax treaties.238 LDCs, however, increasingly impose tax rates that are much closer to, and in some cases are less than, the typical rates provided in treaties. For example, dividends paid to foreign shareholders would normally be subject to a 10% tax in Ghana, unless the company paying the dividend operates in a free

232. Tax sparing was contemplated but ultimately rejected in tax treaties with Egypt, India, and Israel, largely due to the efforts of Stanley Surrey, who argued vigorously against the provision. See Laurey, supra note 222, at 475–76. Tax sparing was also introduced in a tax treaty with Pakistan, but a subsequent change in Pakistan law rendered the provision obsolete and the treaty entered into force without it. Staff of S. Foreign Relations Comm., 85th Cong., Report on Double Tax Conventions, S. Exec. Rpt. No. 1, 85-2, ¶ 3 (1958). 233. This rule is enforced under the “saving clause” found in all U.S. tax treaties. See, e.g., U.S. Model, supra note 56, art. 1, ¶ 4. 234. Laurey, supra note 222, at 471, 493 (many LDCs have “refused to sign U.S. tax treaties that do not contain tax sparing clauses,” especially those in Latin America because this region “resents the U.S. [residence-based] tax policy”). 235. McDaniel, supra note 223, at 292. 236. Tillinghast, supra note 62, at 477. 237. Some countries have incorporated treaty concepts into their domestic laws. For example, permanent establishment thresholds that parallel or closely follow the OECD model treaty definition have long been the domestic rule in Japan, the Netherlands, Sweden, Germany, and France. Ault & Arnold, supra note 207 at 432–34 (1997). 238. OECD Model rates do not exceed 15% for dividends, 10% for interest, and 0% for royalties. OECD Model, supra note 21, arts. 10–12. In contrast, maximum statutory tax rates in OECD countries average 18%, 14%, and 16% on dividends, interest, and royalties, respectively. See generally Ernst & Young, supra note 35.

tax treaties for investment and aid to sub-saharan africa: a case study 617

zone, in which case the tax rate may be zero.239 Thus, Ghana’s statutory tax rate is the same as or less than what would be expected under the hypothetical U.S.Ghana treaty outlined above.240 In addition, Ghana’s internal rate is lower than the 15% maximum provided in the U.S. Model for regular dividends.241 Nevertheless, it is higher with respect to direct dividends than the maximum 5% provided in the U.S. Model and the zero rate for dividends paid to foreign controlling company shareholders found in new treaties. Because most dividends paid out of Ghana would likely constitute direct dividends, many of which would be paid to controlling shareholders,242 a treaty rate that followed the U.S. Model or recent U.S. treaty practice would reduce taxation on U.S. investors in Ghana from the internal rate of 10% (or zero)243 to 5% or zero. However, as discussed above, if U.S. tax treaty precedent is followed, it is unlikely that a U.S.-Ghana tax treaty would provide for these lower rates. In fact, if, pursuant to the U.S. Model, a U.S.-Ghana tax treaty provided for regular dividend taxation lower than 10%, direct dividend taxation at 5%, and no source country taxation of interest and royalties, it would be the first and only U.S. tax treaty to do so with any country, developed or less developed.244 If, as a “concession” to Ghana, the United States provided that instead of a maximum 5% rate for direct dividends, the maximum source-country rate would be 10%, the only result would be that Ghana’s statutory 10% rate would be maintained.245 No benefit in the form of reduced taxation would be realized under this agreement. In fact, if the recently concluded U.S.-Sri Lanka treaty

239. Ghana currently imposes a 10% tax on most dividends paid to nonresidents, but provides tax incentives, including exemptions of taxation on passive income paid by domestic companies to foreign investors, as described above. See G.I.R.A., supra note 83, § 2, Schedule I, Part V (2000); see also supra, text accompanying notes 195–97. 240. See supra Part III.B. 241. U.S. Model, supra note 56, art. 10. 242. See supra, text at note 189. 243. The rate depends on whether the payment derives from sources protected by a free zone or tax holiday regime. 244. The closest rates to these are found in the treaty with Russia, which provides for source country tax rates of 10% on regular dividends, 5% on direct dividends, and 0% on interest and royalties. Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital, U.S.-Russ., arts. 10–12, June 17, 1992, K.A.V. 3315 (hereinafter U.S.-Russia Treaty). See Internal Revenue Service, U.S. Dep’t of the Treasury, Publ’n No. 901, U.S. Tax Treaties 33–34 (Rev. May, 2004) (hereinafter U.S. Tax Treaties) (providing rate information in other treaties). Note that although the IRS published this document in May, 2004, it has no information regarding the U.S. tax treaty with Sri Lanka, which was signed on March 14, 1985, because it did not enter into force until June 13, 2004. See generally id. See also U.S.-Sri Lanka Treaty, supra note 4. 245. The treaty with Ghana would be one of six U.S. treaties with a top 10% rate for dividends. See U.S. Tax Treaties, supra note 244, at 33–34 (providing 10% as the maximum tax rate on dividends in U.S. tax treaties with China, Japan, Mexico, Romania, and Russia).

618 allison christians

serves as a model, a U.S.-Ghana treaty could even provide for maximum rates that are higher than Ghana’s internal rates, though again this could hardly benefit current or potential investors.246 Similarly, Ghana’s statutory rates of 5 to 10% on interest and 15% on rents and royalties247 comport with the average respective rates offered under other U.S. treaties, although the U.S. Model contemplates zero source taxation of both.248 Just as in the case of direct dividends, preserving a higher rate of tax would be likely under UN Model standards, but would generally be a neutral factor for investors. Concessions that allow for higher source country taxation of passive income items reflect the concerns addressed by the UN Model regarding the worldwide allocation of tax revenues. These concessions are meant to protect the taxing jurisdiction of capital importing nations like Ghana against the effects of the U.S. and OECD Model treaties, which allocate income away from source and towards residence countries.249 As the case of Ghana illustrates, however, preserving higher source-country taxation is a neutral measure at best. It is also contradictory to the notion otherwise promoted by U.S. policy makers that reducing tax rates will reduce tax barriers to direct investment and thereby increase capital flows between countries. To date there is no consensus regarding the appropriate balance of attracting investment through lower tax rates and preserving the allocation of revenue to

246. In the U.S. treaty with Sri Lanka, the Joint Committee queries whether this result is intended, and supposes that Sri Lanka could raise its rates up to the maximum 15% provided, thereby increasing its revenues from foreign investment. See Explanation of Sri Lanka Treaty, supra note 135, at 62–63. Yet in the same document, the Committee proclaims that the treaty will be good for the U.S. because it reduces Sri Lankan tax on U.S. investors and provides a clearer framework. Id. These two positions are difficult to reconcile, as the Joint Committee appears to recognize. 247. Ghana currently imposes a 10% tax on most interest payments, and a 15% tax on rents and royalties, with alternate rates ranging from 5–15% for certain payments, depending on the residence of the recipient and the payor. G.I.R.A., supra note 83, ch. I, Part I, §§ 2, 84; First Schedule, Part IV–VIII. 248. With respect to interest, see U.S. Model, supra note 56, art. 11. Thirty-one existing U.S. treaties, including several of the most recently signed treaties and protocols, reflect the goal of zero source-based taxation of interest, rents, and royalties. See, e.g., U.S.–Japan Treaty, supra note 155, art. 11; Australia Protocol, supra note 154, art. 7; U.S.–U.K. Treaty, supra note 156, art. 11. Interest tax rates range from 5–30% in the remaining treaties. U.S. Tax Treaties, supra note 244. With respect to royalties, see U.S. Model, supra note 56, art. 12. Twenty–six existing U.S. treaties, including several of the most recently signed treaties and protocols, provide zero source country tax on most royalties. See, e.g., U.S.–Japan Treaty, supra note 155, art. 12; U.S.-U.K. Treaty, supra note 156, art. 12. As in the case of interest, royalty tax rates range from 5–30% in the remaining treaties. U.S. Tax Treaties, supra note 244, at 33–34. 249. See supra notes 135–36 and accompanying text.

tax treaties for investment and aid to sub-saharan africa: a case study 619

source countries.250 Preserving source-country revenues has been prioritized on the grounds that low taxation has a deleterious effect on infrastructure. In LDCs, providing adequate infrastructure to attract multinationals has been a continuous challenge that is further complicated by tax competition, a phenomenon that is not alleviated by tax treaties. 4. Inadequate infrastructure and nontax barriers U.S. investors may be significantly influenced in their investment location decisions by broad infrastructure-related criteria such as the rule of law and the protection of property, as well as the immediate need for a suitable workforce and adequate physical infrastructure.251 The need for a suitable workforce in turn necessitates basic infrastructure including institutions such as schools and health care systems. In direct tension with these needs is the diminishing ability of LDCs to finance infrastructural development as they decrease taxes on business profits. Many countries, including Ghana, offer tax incentives such as tax holidays and tax-free zones because attracting investment to sustain economic development is deemed of greater importance than protecting tax revenues.252 Nevertheless, there is little consensus regarding the effectiveness of tax incentives and tax holidays in actually attracting foreign investment. Anecdotal evidence from various countries suggests that providing tax incentives to attract foreign investment has failed to deliver the promised benefits.253 Despite a plethora of tax holidays and other tax incentives, few permanent employment opportunities have been created, and exports have failed to increase in the many free zones located throughout West Africa, including Ghana.254 According to John Atta-Mills, former Commissioner of the Ghanaian IRS, “experience shows that tax holidays and tax reductions are ranked very low in the priority of investors in

250. See, e.g., UN Model, supra note 78, art. X–XII (illustrating the lack of consensus through the omission of standard rates). 251. Hooke, supra note 124, at 47–49. For example, a stable macroeconomic environment and a well-educated workforce are two factors that correlate with greater foreign investment flowing into LDCs. UNCTAD, supra note 2, at 23. 252. Brian J. Arnold & Patrick Dibout, General Report, 55 Cahiers De Droit Fiscal International 25, 28 (2001) (“Certain countries . . . are more concerned with attracting activity and investment of the multinationals in order to sustain their economic development”). 253. See, e.g., Tamas Revesz, EU, Companies Urge Reform of Hungary’s Local Industry Tax, 2004 WTD 97-10 (May 14, 2004) (“Although Slovakia offered big investment subsidies and tax relief for foreign investors, its budget is in ruins, and the resulting forced cuts in government spending (especially transfers to households) have triggered serious hunger riots among the most seriously hit Roma population”). 254. Papa Demba Thiam, Market Access and Trade Development: Key Actors, in Towards a Better Regional Approach to Development in West Africa 97, 101 (John Igue & Sunhilt Schumacher eds., 1999). See also supra note 117 (stating that trade data indicates imports from Ghana to the U.S. are currently declining).

620 allison christians

their choice of location for their business,” and that product demand, a skilled workforce, and infrastructure are more important to businesses.255 Economic evidence regarding the connection between taxation and foreign investment provides little additional certainty. A number of economic studies indicate that multinationals are very sensitive to tax considerations, and therefore corporate location decisions may be heavily influenced by tax regimes in source countries.256 Conflicting studies indicate, however, that taxation is not a significant factor in the location decisions of U.S. multinationals.257 Instead, these studies argue that “market size, labor cost, infrastructure quality . . . and stable international relations,” among other considerations, are the most important factors for location decisions.258 Studies focused particularly on foreign investment in Sub-Saharan Africa come to the same conclusion.259 In contrast, recent literature suggests that past studies present an incomplete picture of the role of taxation because they have focused on source-country corporate income taxes, the burdens of which are relatively insignificant as compared to the burdens of non-income taxation in source countries.260 As a result, these past studies may have obscured the more significant influence 255. Seth E. Terkper, Tax Measures in Ghana’s 2004 Budget Inadequate, Opposition Party’s Presidential Candidate Says, 2004 WTD 63–12 (Apr. 1, 2004). 256. For an overview of this economic literature, see Avi-Yonah (2000), supra note 164, at 1590–92; James R. Hines, Jr., Tax Policy and the Activities of Multinational Corporations, in Fiscal Policy: Lessons from Economic Research 401, 401-45 (Alan J. Auerbach ed., 1997); James R. Hines, Jr., Lessons from Behavioral Responses to International Taxation, 52 Nat’l Tax J. 305, 305–22 (1999). 257. See McDaniel, supra note 223, at 280 (providing an overview of some of this literature); see also G. Peter Wilson, The Role of Taxes in Location and Sourcing Decisions, in Studies in International Taxation, supra note 204, at 196–97, 229 (arguing that taxes are more influential in location decisions for administrative and distribution centers, but they “rarely dominate the decision process” in the case of manufacturing locations). 258. McDaniel, supra note 225, at 280. 259. See, e.g., Elizabeth Asiedu, On the Determinants of Foreign Direct Investment to Developing Countries: Is Africa Different? 1, 6 (2001), available at http://ssrn.com/ abstract=280062 (arguing that location-specific factors such as natural resource availability may make infrastructure and stability of particular importance in the context of investment to Sub-Saharan Africa); World Bank, World Business Environment Survey 2000, available at http://www.ifc.org/ifcext/economics.nsf/Content/ic-wbes (finding as a result of a survey of business including Ghana and fifteen other Sub-Saharan African countries that firms investing in these regions indicate less sensitivity to taxation than to corruption, infrastructure, crime, inflation, financing, and political stability). 260. Mihir A. Desai, C. Fritz Foley & James R. Hines Jr., Foreign direct investment in a world of multiple taxes, 88 J. Pub. Econ. 2727, 2728 (2004) (“Foreign indirect tax obligations of American multinational firms are more than one and a half times their direct tax obligations”). In previous studies, James Hines found a “small but significant” link between lower source country taxes and foreign investment levels. See McDaniel, supra note 225, at 281; see also Avi-Yonah, supra note 164, at 1644.

tax treaties for investment and aid to sub-saharan africa: a case study 621

of non-income taxation on foreign investment decisions.261 Because foreign nonincome tax burdens significantly exceed income tax burdens, these taxes may strongly influence the behavior of U.S. multinationals.262 The main explanation given for this influence is that non-income taxation cannot be credited against U.S. residual taxation.263 The findings of this literature are consistent with earlier studies that suggest the relative importance of taxation in a particular country may be increasing with the availability of opportunities for avoiding taxation elsewhere.264 These findings, however, conflict with other studies demonstrating that multinationals can use debt financing and transfer pricing manipulation to achieve tax neutrality in investment location decisions,265 and that despite earlier studies showing a connection between tax and foreign direct investment, nontax factors dominate the location decisions of multinational firms.266 Given the possibility that taxation may not be an overriding factor in foreign investment location decisions, the influence of infrastructure cannot be ignored. To the extent infrastructure is important to potential investors, efforts to reduce taxation to attract foreign investment may be counterproductive, since raising sufficient revenues is integral to the level of infrastructure a country can offer.267

261. See Desai, supra note 260, at 2728. 262. See id. at 2729. 263. Id. at 2728 (“The role of non-income taxes may be particularly important for FDI, since governments of many countries (including the United States) permit multinational firms to claim foreign tax credits for corporate income taxes paid to foreign governments but do not extend this privilege to taxes other than income taxes. As a result, taxes for which firms are ineligible to claim credits may well have greater impact on decisionmaking than do creditable income taxes”). For an argument that the definition of creditable taxes should be broadened to encompass many non-income taxes, see generally Glenn E. Coven, International Comity and the Foreign Tax Credit: Crediting Nonconforming Taxes, 4 Fla. Tax Rev. 83 (1999). 264. See Grubert, supra note 67, at 22, 28 (suggesting that tax rates are extremely important to U.S. multinationals in allocating their foreign direct investment, especially in the case of manufacturing, and that the relative importance of taxes may be increasing). 265. See Avi-Yonah, supra note 25, at 1315; Gary Clyde Hufbauer, U.S. Taxation of International Income: Blueprint for Reform 134 (1992). 266. See Haroldene Wunder, The Effect of International Tax Policy on Business Location Decisions, Tax Notes Int’l, Dec. 24, 2001, at 1331–48. 267. See Nicholas Kaldor, Will Underdeveloped Countries Learn to Tax?, 41 Foreign Aff. 410, 410 (1963) (stating that “[t]he importance of public revenue to the underdeveloped countries can hardly be exaggerated if they are to achieve their hopes of accelerated economic progress”). See also H. David Rosenbloom, Response to: “U.S. Tax Treatment of Foreign Source Income Earned in Developing Countries: Administration and Tax Treaty Issues,” 35 Geo. Wash. Int’l. L. Rev. 401, 406 (2003) (stating that “taxes are, by definition, involuntary exactions”). Thailand has recently taken a slightly different approach. In June, 2004, the Prime Minister, the Ministry of Education, and the Social and Human Development

622 allison christians

As tax competition ensures less taxation of multinationals, the ability of LDCs to fund sufficient infrastructure to attract and sustain foreign investment relies more heavily on the ability to tax resident individuals, whether directly or indirectly. Historically, this has been a great challenge for LDCs.268 Compliance rates for income and non-income taxation are typically very low in Ghana. It is estimated that 80% of business is conducted on the informal market—that is, not subject to regulation or taxation because it is conducted in the form of cash or barter.269 Thus, only those who work for the government or for the few companies that comply with wage withholding obligations pay their income taxes.270 An appearance that the laws are not applied uniformly may in turn lead to increased tax avoidance and evasion.271 The situation is exacerbated in an environment in which corruption or mismanagement of public funds also exists. While Ghana’s corruption factor is relatively modest in comparison to many of its neighbors in Sub-Saharan Africa,272 the notion persists that wealth

Services Ministry unveiled tax incentives for individuals and companies that make charitable donations to social development programs including education, museums, libraries, art galleries, recreational facilities, children’s playgrounds, public parks, and sports arenas. The government hopes that “[these incentives] will raise funds from the private sector to alleviate the poverty crisis in Thailand.” Sirena J. Scales, Thai Government Announces Tax Incentives for Charitable Contributions, 2004 WTD 129-10 (July 6, 2004). 268. Kaldor, supra note 267, at 410. 269. The agricultural industry contributes significantly to this number, since over 60% of Ghana’s population is employed in agriculture (a slightly lower percentage than the average of approximately 75% for LDCs in Sub-Saharan Africa). These percentages were compiled by averaging the stated percentage for each LDC in Sub-Saharan Africa from the World Factbook. (Spreadsheet containing data on file with author.) World Factbook, supra note 1, available at http://www.cia.gov/cia/publications/factbook/ fields/2048.html. The informal economy also includes most professionals such as doctors and lawyers, other service providers, and shopkeepers and sellers of goods in local markets. Interview with Justice Insaidoo, supra note 211; Interview with Sheila Minta, Solicitor/Barrister, Addae & Twum Company, in Accra, Ghana (Dec. 9, 2003) (on file with author). 270. Interview with Justice Insaidoo, supra note 211. 271. Murphy, supra note 213, at 201 (“perceptions of unfair treatment” appear to affect trust, and “taxpayer resistance could be sufficiently predicated by decreased levels of trust”). 272. See Transparency International, Global Corruption Report 2003, 215, 220, 225, 264, available at http://www.globalcorruptionreport.org/gcr2003.html (suggesting that although the government faces much criticism in failing to address corruption within the civil service, prompting President Kufuor to promise an increase in accountability, Ghana’s perceived corruption is much lower than that of many of its neighbors in SubSaharan Africa). In extreme comparison stand countries like the Congo, where corruption and bribery at all levels are openly acknowledged as requisite for survival. See Davan Maharaj, When the Push for Survival is a Full-Time Job, L.A. Times, July 11, 2004, at A1 (hereinafter Maharaj, Push) (explaining that while government employees are not paid a salary, they still show up for work every day to collect bribes ranging from “about $5 for

tax treaties for investment and aid to sub-saharan africa: a case study 623

can be acquired by becoming a government official.273 These perceptions plague the revenue collection efforts of tax agencies in LDCs such as Ghana.274 The ability of LDCs to collect sufficient revenue to fund infrastructure is also challenged by international pressure to open markets and reduce trade barriers.275 To the extent that Ghana continues to rely heavily on trade taxes for its revenue, recent developments in tariff reduction at the WTO may cause additional revenue shortfalls in the future. Ghana also faces difficulty in finding consistent resources to fund infrastructure because success in collecting revenues from excise taxes, royalties, dividends, and similar payments may depend on fluctuating global market prices for exported commodities.

a birth certificate to about $100 for an import license”). In Benin, a close neighbor to Ghana, bribes collected from traders trying to import illegal goods into Nigeria provide some 15% of the nation’s revenues. Davan Maharaj, For Sale—Cheap: ‘Dead White Men’s Clothing,’ L.A. Times, July 14, 2004, at A1. 273. The phenomenon appears to exist throughout Sub-Saharan Africa. See Transparency International, supra note 272, at 215. In the Congo, people say that “[t]he only ones who have ever gotten rich are the leaders and those with connections.” Maharaj, Push, supra note 272. 274. The perceptions of a few individuals cannot represent national sentiment, nor can such sentiment, even if widespread, indicate the accuracy of the charge. Nevertheless, a perception of unfairness and corruption may undermine the efficacy of a tax regime. A study to quantify the effect of corruption on tax compliance is underway in Tanzania, but more research is needed in this area. A further issue that may be significant to the tax collection efforts of LDCs is the perceived misuse of funds by the government, whether as a result of corruption or the ineptitude of officials. Informally, this author heard many expressions of dissatisfaction with the ability of the government to provide necessary services to the citizenry. Since that is a common complaint in developed countries as well, I do not deal with it here, but only note its existence as an additional potential difficulty in raising sufficient revenues from individuals. Finally, the extent to which local conditions and attitudes regarding taxation affect the behavior of multinationals is not conclusively established. It may be that multinationals generally conduct their business operations fundamentally in compliance with the laws in force, regardless of the degree to which their compliance is monitored or enforced, simply because their global operations may be subject to scrutiny by other governments or the public. Evidence proving (or disproving) this theory, however, appears to be lacking in the economic literature. 275. The transition of the U.S. from an agrarian society “rich in resources but lacking in capital investment” to an industrial one is credited in part to tariffs, without which the transition would have been much slower. See Weisman, supra note 41, at 14; see also William A. Lovett, Alfred E. Eckes Jr. & Richard L. Brinkman, U.S. Trade Policy: History, Theory, and the WTO 45 (2004) (finding the current association of free trade with rapid economic growth “incompatible with American economic history,” which shows that “[t]he most rapid growth occurred during periods of high protectionism”).

624 allison christians

Finally, Ghana’s ability to fund infrastructure is subject to uncertainty due to its reliance on assistance from foreign donors.276 In 2002, Ghana received large amounts of foreign aid, much of which was connected to the peaceful transition of power through the democratic process. But the amount of aid has fallen recently, and it is expected to continue to decline as finances are directed to other countries or fall off as a result of donor fatigue.277 The consequence is consistent budget shortfalls in Ghana.278 An increase in the overall level of funding by donor countries might alleviate the shortfall.279 Nevertheless, a subsequent change of policies by the aid donor countries could cripple expectant aid recipients like Ghana, as foreign aid typically substitutes for—rather than supplements— domestic revenue raising efforts.280 Multinational companies may be expected to increase the government’s ability to collect revenues by creating a larger wage base for personal income tax. Wages in LDCs such as Ghana, however, average $1 per day, producing little for the government to share.281 If wages are raised through regulatory action, many

276. In Ghana, 17% of total revenue derives from non-tax sources. State of the Ghanaian Economy, supra note 170, at 26–31. Of this amount 86% (or just under 15% of total revenues) derives from grants. The other 14% derives from receipts from various fees charged by the government for particular services, and from amounts received in divestiture of state-owned enterprises. In this respect, Ghana is somewhat better off than many of the other LDCs in Sub-Saharan Africa, which rely heavily on foreign aid to subsidize their expenditures. For example, 53% of Uganda’s budget comes from external loans and grants. See Gumisai Mutume, A New Anti-poverty Remedy for Africa?, 16 Africa Recovery 12 (2003), available at http://www.un.org/ecosocdev/geninfo/afrec/vol16no4/164povty.htm. 277. See State of the Ghanaian Economy, supra note 170, at 30, 34. 278. Ghana’s 2002 budget provided for an expected budget deficit of 4.4%. Yaw OsafoMaafo, The Budget Statement and Economic Policy of the Government Of Ghana, 2002 at 11. Citing “substantial shortfalls in expected foreign inflows,” the 2003 budget nevertheless projected a smaller deficit of 3.1% of GDP. Yaw Osafo-Maafo, The Budget Statement and Economic Policy of the Government Of Ghana, 2003 at 8, 16, 24, 96. The 2004 budget, acknowledged that actual receipts were significantly lower than projections in 2003 due to shortfalls in grants and other non-tax revenues, which resulted in an actual deficit in 2003 of 3.4% rather than the anticipated 3.1%, projected a budget surplus for 2004 of 1.67%. Yaw Osafo-Maafo, The Budget Statement and Economic Policy of the Government Of Ghana, 2004 at 16, 19, 42. 279. For example, if developed countries follow through on their recent pledges to relieve existing debt and double aid efforts in Africa. See, e.g., A First Step on African Aid, N.Y. Times, June 14, 2005, at A22 (describing Bush’s pledge to “ease the burden of debt in Africa”); Celia W. Dugger, U.S. Challenged to Increase Aid to Africa, N.Y. Times, June 5, 2005, at A10 (describing building consensus for doubling of aid to Africa); Paul Blustein, After G--8 Aid Pledges, Doubts on ‘Doing It,’ Wash. Post, July 10, 2005, at A14 (describing pledges of the G-8 countries to double their aid to Africa, but noting that “[t]he amounts actually spent have a history of falling far short of the amounts pledged”). 280. Kaldor, supra note 267, at 410. 281. Nearly 45% of the population of Ghana lives on less than $1 per day. United Nations Development Programme, Human Development Report, 2005, at 228, available

tax treaties for investment and aid to sub-saharan africa: a case study 625

multinationals may disengage to seek low wages elsewhere, since the low cost of labor is often a primary reason multinationals set up in LDCs.282 Although workers may benefit individually from employment created by foreign investment, even if wages are only minimally higher than that offered by other local employment, they are not necessarily placed in a better position with respect to paying taxes.283 Investment protection or insurance—whether made available through private or public institutions—may promote foreign investment despite a country’s infrastructural deficiencies. In the U.S., investment protection is provided to U.S. investors through the United States Export-Import Bank (“Ex-Im Bank”), an independent federal government agency that “assume[s] credit and country risks the private sector is unable or unwilling to accept,” through export credit insurance, loan guarantees, and direct loans to U.S. businesses investing in foreign countries.284 For example, Ex-Im Bank insurance covers the risk of foreign buyers not paying bills owed to U.S. investors, the risk that a foreign government might restrict the U.S. company from converting foreign currency to U.S. currency, and even the risk of loss due to war.285 In effect, this kind of investment protection substitutes U.S. infrastructure for that existing in LDCs.286 at http://hdr.undp.org/reports/global/2005. In all of Sub-Saharan Africa, the figure is close to 46%. See Patricia Kowsmann, World Bank Finds Global Poverty Down By Half Since 1981, UN Wire, April 23, 2004, available at http://www.un.org/special-rep/ohrlls/ News_flash2004/23%20Apr%20World%20Bank%20Finds%20Global.htm. Globally, it is estimated that about half of the earth’s population lives on under $2 per day, a fact that has been central to the most recent efforts of the U.S. to combat poverty with new foreign aid strategies aimed at economic growth. See, e.g., Colin L. Powell, Give Our Foreign Aid to Enterprising Nations, Newsday (New York), June 11, 2003, at A34 (discussing the role of the Millennium Challenge Account in a new strategy of directing foreign aid to “support for sustainable development” in the face of the ongoing challenge of widespread global poverty). 282. Hooke, supra note 124, at 18–19. 283. See Nicholas D. Kristof, Inviting All Democrats, N.Y. Times, Jan. 14, 2004, at A19 (arguing that “the fundamental problem in the poor countries of Africa and Asia is not that sweatshops exploit too many workers; it’s that they don’t exploit enough,” as illustrated by the example of a young Cambodian woman who averages seventy-five cents a day from picking through a garbage dump and for whom “the idea of being exploited in a garment factory—working only six days a week, inside instead of in the broiling sun, for up to $2 a day—is a dream”). 284. See Mission of Export-Import Bank of the United States, http://www.exim.gov/ about/mission.html (last visited Jan. 13, 2006). 285. See, e.g., Multi-Buyer Export Credit Insurance, http://www.exim.gov/products/ insurance/multi_buyer.html (last visited Jan. 13, 2006). 286. The subsidy is not without controversy. See, e.g., Heather Bennett, House OKs Measure to Block Loans to Companies Relocating in Tax Havens, 2004 WTD 139-4 (reporting that as part of a foreign aid bill, U.S. Export-Import Bank loans would no longer be made to corporate entities chartered in one of several listed tax havens because, according to Representative Sanders, who offered the bill, “[c]ompanies that dodge U.S. taxes should

626 allison christians

The Ex-Im Bank has a Sub-Saharan Africa Advisory Committee devoted specifically to supporting U.S. investment activity in this region.287 With investment protection available as a substitute for prohibitive infrastructural shortcomings, investment in LDCs like Ghana may not be economically prohibitive. Yet, the persistently low level of foreign investment in Ghana and Sub-Saharan Africa as a whole suggests that investment protection is not enough to overcome the barriers perceived by potential investors. 5. Entrenched investor perception. As tax treaties with LDCs may provide little commercial benefit to investors when little or no income tax is imposed in these countries, it is perhaps not surprising that they are correspondingly low on the list of U.S. treaty priorities.288 Nevertheless, tax treaties continue to be promoted for their ability to increase investment between developed and less-developed countries. One theory for their promotion is that increased investment can be expected due to the signaling effects of tax treaties.289 For example, it has been suggested that tax treaties may signal a stable investment and business climate in which treaty partners express their dedication to protecting and fostering foreign investment.290 not be rewarded with taxpayer handouts,” but should “go to the government” of the applicable tax haven for such privileges). 287. See Export-Import Bank of the United States, Sub-Saharan Africa Advisory Committee, http://www.exim.gov/about/leadership/africa.htm (last visited Dec. 6, 2005). 288. See Statement of Barbara Angus, supra note 5, at 10, stating that the United States generally does not: conclude tax treaties with jurisdictions that do not impose significant income taxes, because there is little possibility of the double taxation of income in the cross-border context that tax treaties are designed to address: with such jurisdictions, an agreement focused on the exchange of tax information can be very valuable in furthering the goal of reducing U.S. tax evasion. 289. Mutén, supra note 79, at 5. 290. The Secretary of the Treasury proclaimed the importance of signing a tax treaty with Honduras in 1956, stating that as the first treaty with any Latin American country, [t]he agreement may . . . have a value far beyond its immediate impact on the economic relations between the United States and Honduras. It may generate among smaller countries an increased awareness of the need to create an economic atmosphere that will lend itself to increased private American investment and trade. Dulles, supra note 5, at 1444. Similar sentiment has been expressed in the context of many treaties, especially those with LDCs, over the years. See, e.g., Staff of S. Foreign Relations Comm., 108th Cong., Taxation Convention and Protocol With The Government of Sri Lanka 4 (S. Exec. Rpt. 108-11, Mar. 18, 2004) (“in countries where an unstable political climate may result in rapid and unforeseen changes in economic and fiscal policy, a tax treaty can be especially valuable to U.S. companies, as the tax treaty may restrain the government from taking actions that would adversely impact U.S. firms”); Staff of the Joint Comm. on Taxation, 106th Cong., Explanation of Proposed Income Tax Treaty and Proposed Protocol Between The United States and The Republic of Venezuela 61 (Comm. Print 1999) (“the proposed treaty would provide benefits (as

tax treaties for investment and aid to sub-saharan africa: a case study 627

Proponents of this argument suggest that in the process of negotiating a tax treaty, governments of LDCs may subject their operations to increased transparency and accountability, thus providing additional benefits to potential investors (as well as domestic taxpayers) in the form of assurances regarding the proper management of public goods.291 Thus, bilateral tax treaties may serve largely to “signal that a country is willing to adopt the international norms” regarding trade and investment, and hence, that the country is a safe place to invest, especially “in light of the historical antipathy that many developing and transition countries have in the past exhibited to inward investment.”292 Signaling is a slippery concept because it is difficult to measure whether signaling is occurring and, if so, whether and to what extent it is impacting investors. The potential for signaling a stable investment climate through tax treaties with LDCs in Sub-Saharan Africa is especially hampered by the persistence of negative perceptions about this region’s investment climate.293 Foreign investors in LDCs often take a regional, rather than national, approach to investment, attributing the negative aspects of one country to others in the vicinity.294 Since few Sub-Saharan African countries have tax treaties, and many countries in the region suffer from civil unrest and economic failure, Ghana’s ability to

well as certainty) to taxpayers”). These concepts are also reflected in commentary. See, e.g., Andersen & Blessing, supra note 10, at ¶ 1.02[3][b] (“[in the context of LDCs,] tax treaties provide foreign investors enhanced certainty about the taxation of the income from their investments.”); see also Davies, supra note 216, at 3 (“even a treaty that merely codifies the current practice reduces uncertainty for investors by lowering the likelihood that a government will unilaterally change its tax policy”). 291. See, e.g., UN Dep’t of Econ. & Soc. Affairs, The Significance of Bilateral Tax Treaties Between Developed and Developing Countries, ¶ 10, UN Doc. ST/SG/2001/L.9 (Apr. 2, 2001) (prepared by Mayer Gabay), available at unpan1.un.org/intradoc/groups/public/ documents/un/unpan000550.pdf (suggesting that the first advantage to a LDC of entering into a bilateral tax treaty is the negotiation process itself, because that process creates a degree of transparency, which in turn promotes “greater rationality in decision making[,]” which “can be of great economic benefit to the developing country.”) 292. Stewart, supra note 11, at 148 (citing Richard J. Vann, International Aspects of Income Tax, in 2 Tax Law Design and Drafting 726 (Victor Thuronyi ed., 2000)). 293. See UNCTAD, supra note 2, at 1. 294. UNCTAD, supra note 2, at iv (“[L]ittle attempt is often made to differentiate between the individual situations of more than 50 countries of the continent.”); Laura Hildebrandt, Senegal Attracts Investors, But Slowly, 17 Africa Recovery 2-15 (2003), available at http://www.un.org/ecosocdev/geninfo/afrec/vol17no2/172inv3.htm (“[F]oreign investors tend to lump countries together in regions, without making much distinction among individual countries,” which might explain Senegal’s limited success in attracting foreign investment despite “relatively good infrastructure . . . a history of political stability and secular democracy, with decidedly pro-market leanings”).

628 allison christians

demonstrate stability and certainty may garner little individual attention from foreign investors unfamiliar with its particular situation.295 In addition, the signaling effect is tied to a country’s reputation in upholding its international compacts. Short of terminating a treaty, there is no formal enforcement mechanism should a country proceed to ignore its treaty obligations.296 For example, it is difficult to imagine that a tax treaty could independently create a sense of stability in a country that would otherwise be unattractive due to historical failure to protect property rights. Finally, treaty proponents point to the certainty achieved in establishing rules consistent with international norms so that investors will know what to expect regarding the taxation of their investments in foreign countries. Nevertheless, signaling certainty and stability is achieved more directly through agreements designed to provide these specific benefits. For example, delivering certainty and stability is the primary purpose of investment protection provisions in global and regional trade agreements.297 These goals are also encompassed in a global network of over 2,100 bilateral investment protection treaties (BITs).298 Ghana has seventeen such treaties currently in force.299 The United States has 295. See, e.g., Hildebrandt, supra note 294, at 15 (“Senegal’s reputation for stability may be offset by conflicts elsewhere in the region, such as Côte d’Ivoire.”); Thabo Mbeki: A Man of Two Faces, The Economist, Jan. 22, 2005, at 27 (“[A]ny . . . plan for Africa’s redemption, will work only if functioning states with reasonably good leaders (South Africa, Botswana, Senegal, Ghana, Mozambique) can be set apart from the awful ones . . .”). 296. In the case of a treaty violation, a taxpayer would request the Competent Authority of its home country to negotiate with the Competent Authority in the treaty partner country. For this reason, investors may desire a tax treaty to be in place so that assistance in negotiating disputes with a foreign country could be sought from the U.S. Competent Authority. Treaties provide little recourse, however, in the event the Competent Authorities fail to reach a resolution. See U.S. Model, supra note 56, art. 25. 297. See Stephen S. Golub, Measures of Restrictions on Inward Foreign Direct Investment for OECD Countries 6 (OECD Econ. Dep’t, Working Paper No. 357, 2003). Most of the LDCs in Sub-Saharan Africa, including Ghana, have signed multilateral agreements dealing with the protection of foreign investment, such as the Convention establishing the Multilateral Investment Guarantee Agency and the Convention on the Settlement of Investment Disputes between States and Nationals of Other States. See UNCTAD, supra note 2, at 7–8. 298. WIR 2003, supra note 118, at 89–91 (stating that BITs signal a country’s attitude towards and climate for foreign investment, and that investors “appear to regard BITs as part of a good investment framework”). Worldwide, there are 2,181 BITs currently in force, encompassing 176 countries. Id. at 89. As in the case of tax treaties, significantly more BITs would be required to achieve global coverage. See supra note 71 and accompanying text. 299. These treaties include agreements with Benin, Bulgaria, Burkina Faso, China, Cuba, Denmark, Egypt, France, Germany, Guinea, India, Malaysia, Mauritius, Netherlands, Romania, Switzerland, and the United Kingdom. See UNCTAD Bilateral Investment Treaty Database, http://www.unctadxi.org/templates/DocSearch.aspx?id=779 (last visited Jan. 14, 2006).

tax treaties for investment and aid to sub-saharan africa: a case study 629

forty-seven in force and relies on these agreements to protect investment in source countries.300 Investment protection provisions and treaties outline the applicable legal structure and regulatory framework of the signatory countries and provide settlement provisions in the event of disputes between investors and source-country governments. Common features include guarantees of compensation in the event of expropriation, guarantees of free transfers of funds and repatriations of capital and profits, and dispute settlement provisions.301 The goal of these agreements is to promote transparency, stability, and predictability for regulatory frameworks in source countries, and therefore to reduce obstacles to the flow of foreign investment.302 BITs are further bolstered through subsidized loans, loan guarantees, and other financial assistance made available to foreign investors.303 Even if the stability and certainty signaled by tax treaties could make a source country that has such agreements more attractive than one that does not, U.S. investors are unlikely to lobby for tax treaties if they do not have a direct financial interest at stake, namely, an exposure to taxation that could be alleviated by treaty.304 The foreign investment patterns of U.S. businesses also imply that tax treaties may be an insufficient signal to investors.305 First, U.S. investors will pursue investments in a country with a sufficiently attractive business environment, even in the absence of a tax treaty.306 For example, although the United States has

300. Hearing Before the S. Foreign Relations Comm. on Economic Treaties, 108th Cong. 9-10 (2004) (statement of Shaun Donnelly, Principal Deputy Assistant Secretary, Bureau of Economic and Business Affairs, State Dep’t) (“BITs have afforded important protections to U.S. investors”). The U.S. currently has four BITs with LDCs in Sub-Saharan Africa: Cameroon, Mozambique, Senegal, and the Democratic Republic of the Congo. For a list of U.S. BITs currently in force, see UNCTAD Bilateral Investment Treaty Database, supra note 299. 301. WIR 2003, supra note 118, at 89. 302. Id. at 91. 303. See supra note 284 and accompanying text (discussing the role of the Ex-Im Bank in subsidizing U.S. investors to LDCs). 304. The lobbying efforts of U.S. businesses may not be the most appropriate means of establishing a list of priorities for new treaties, however, it is one of the primary factors considered by the office of International Tax Counsel in making such decisions. See Testimony of Barbara Angus, supra note 5, at 10. 305. See Mutén, supra note 79, at 4. 306. See, e.g., Jones, supra note 87, at 4, 5 (arguing that tax treaties “make less difference to domestic taxpayers investing abroad,” especially if taxes are low in source countries).

630 allison christians

no treaty with Brazil,307 U.S. foreign direct investment in Brazil is significant.308 Second, the mere presence of a tax treaty will generally overcome neither an otherwise poor business climate, nor one that deteriorates after a treaty is in place. For example, the United States entered into a treaty with Venezuela in 1999, but the amount of U.S. capital flowing to Venezuela subsequently dropped sharply due to “concerns over regulations and political instability in the country.”309 Finally, some investors may not necessarily want a tax treaty because such agreements usually include measures that prevent tax evasion, as discussed above.310 Thus, while tax treaties may send positive signals to investors, they may as likely send negative signals to the extent they lead the way to stronger enforcement of tax laws. Supporting tax evasion is clearly indefensible as a policy for encouraging investment in LDCs, but the benefits of such opportunities to existing investors, and the cost of eliminating such opportunities, cannot be ignored.311 Nevertheless, easing enforcement and administration of the tax laws of potential LDC treaty partners may be an alternative reason to continue expanding the U.S. tax treaty network.312 For example, the information-exchange provisions 307. Brazil is one of the South American countries that refuses to negotiate with the U.S. due to the tax sparing controversy. See Laurey, supra note 222, at 491 n.155 (noting that due to the tax sparing controversy, Mexico was the first Latin American nation to sign a tax treaty with the U.S.); Mitchell, supra note 11, at 213; Guttentag, supra note 4, at 451, 452. The latest U.S. discussions with Brazil were held in 1992. See Venuti et al., supra note 14. As Brazil continued to insist on tax sparing and the U.S. refuses to negotiate with countries that insist on including such a provision, no further meetings are planned. See id. 308. As valued at historical cost (book value of U.S. direct investors’ equity in and net outstanding loans to Brazilian affiliates), U.S. foreign direct investment in Brazil is currently valued at almost $30 billion. Borga & Yorgason, supra note 101, at 49. At 1.7% of total U.S. foreign direct investment, Brazil’s market for U.S. foreign direct investment is not far behind that of some developed countries, including Spain (with 2.1% of U.S. foreign direct investment) and Australia (with 2.3%). Id. at 42. 309. UNCTAD, FDI in Brief: Venezuela (2004), http://www.unctad.org/sections/ dite_fdistat/docs/wid_ib_ve_en.pdf. 310. In the past, tax treaties may have contributed to tax evasion by creating opportunities for “treaty shopping” through the use of multicountry tiered structures such as the one shut down in Aiken Indus, Inc. v. Comm’r, 56 TC 925 (1971). In that case, the U.S.– Honduras treaty then in force was used to channel interest payments free of tax from the U.S. to the Bahamas. Id. at 929–31. The U.S.–Honduras treaty was terminated in 1966, before the case was decided but in connection with these kinds of structures, deemed to be void of any “economic or business purpose” by the Tax Court. Id. at 929, 934. Treaty shopping has since been curtailed in newer treaties and protocols by means of stronger limitation of benefits provisions. See Arnold & Dibout, supra note 252, at 73–74. 311. Just as in the cases of deferral and bank secrecy, the private sector can be expected to protect tax advantages regardless of whether they comport with a coherent tax policy. 312. Obtaining cooperation in tax enforcement through information sharing provisions is a major factor in the completion of treaties from the perspective of the U.S. For

tax treaties for investment and aid to sub-saharan africa: a case study 631

might enable Ghana to extend its current, basically territorial, regime to a worldwide regime.313 The benefit of such a regime would depend on the amount of savings shifted to the United States by Ghanaian persons before and after the treaty. This is presumably a relatively tiny amount by global standards,314 but it might be significant to the overall revenue picture in Ghana. Nevertheless, Ghana’s limited tax treaty network significantly restricts its ability to enlarge its taxing jurisdiction, since Ghanaians could simply choose a location other than the United States for their offshore activities, avoiding Ghanaian tax even under a worldwide system.315 Moreover, as in the case of investment protection, the benefits of information exchange are as readily—and more broadly—achieved through agreements specifically addressing this issue. Information exchange is comprehensively addressed in Tax Information Exchange Agreements (TIEAs), which are generally bilateral, and through multilateral agreements such as the OECD TIEA.316 Under U.S. TIEAs, assistance in tax enforcement and collection is extended not only to income taxes but to other taxes as well, making such agreements

example, the newly-ratified tax treaty with Sri Lanka was originally negotiated almost twenty years ago but only entered into force this year. U.S.-Sri Lanka Treaty, supra note 4; Treasury Press Rel. JS-1809, supra note 4. Ten years of the delay were due to Sri Lanka’s reluctance in accepting U.S. requests regarding information exchange. See Letter of Submittal from Colin L. Powell, U.S. Department of State, to the President (Aug. 26, 2003), reprinted in Protocol Amending Tax Convention with Sri Lanka, U.S–Sri Lanka, Sept. 20, 2002, S. Treaty Doc. No. 108-9 (2003). The fact that, as in the case of Ghana, Sri Lanka’s statutory rates and tax incentive regimes indicate that the domestic tax regime is as or more favorable than that provided under the treaty, suggests that prevention of double taxation plays a much less significant role than prevention of tax evasion. 313. See U.S. Model, supra note 56, art. 26. For example, when Venezuela entered into a tax treaty with the United States, its tax regime was territorial: Venezuela imposed no tax on the foreign income of its residents. Its tax treaty with the U.S. included the typical exchange-of-information provision, which would theoretically allow Venezuela to pursue its residents who engaged in activities outside of the country, and Venezuela subsequently expanded its jurisdiction to encompass residence-based taxes. U.S.–Venezuela Treaty, supra note 4, art. 27. 314. The U.S. Bureau of Economic Affairs compiles data regarding direct investment in the U.S., but Ghana is included only collectively with the rest of Africa, excluding South Africa. Borga & Yorgason, supra note 101, at 51. Inbound direct investment from this region is valued at $1.8 billion, which represents less than 0.2% of that from Europe. Id. 315. Ghana would not generally benefit from the larger U.S. tax treaty network since the exchange of information is only limited to that which is relevant to the two contracting states. U.S. Model, supra note 56, art. 26. 316. The OECD Agreement has been signed by the U.S. and Canada, among others. OECD Model, supra note 21, at 2.

632 allison christians

potentially more effective than tax treaties in fulfilling the goal of improved tax administration and enforcement.317 Absent reduction of double taxation, the non-commercial benefits of tax treaties appear incapable of independently exerting a significant influence on U.S. foreign investment, and some of the aspects of tax treaties may tend to discourage such investment. Ultimately, the value of continued expansion of the U.S. tax treaty network to LDCs may therefore be extremely limited in the context of a global tax climate that reflects the circumstances illustrated in this case study.

conclusion The investment and aid goals of tax treaties are undermined by competing tax regimes, including domestic U.S. rules that provide relief of current U.S. tax burdens on foreign income earned by multinational companies. To the extent multinationals can escape U.S. taxation simply by investing abroad, the U.S. fosters tax competition throughout the world as foreign countries compete to attract the U.S. capital fleeing taxation at home. As a result of this international tax competition and a corresponding divergence in tax mix between developed and less developed countries, much of the tax ostensibly relieved under tax treaties no longer exists to a significant extent with respect to investment in many LDCs. As a result, traditional tax treaties with these countries may offer few commercial benefits to investors. Tax treaties may provide non-commercial benefits to partner countries and investors by signaling stability or suitability or by providing certainty. These incidental benefits, however, are likely insufficient to significantly impact investment in many LDCs, particularly those in Sub-Saharan Africa. Thus, as this case study of Ghana demonstrates, much of the conventional wisdom about the impact of tax treaties on the global flow of investment does not apply in the context of many of the LDCs most in need of realizing the benefits attributed to these agreements.

317. The U.S. entered into tax treaties with many countries in Sub-Saharan Africa and the Caribbean simultaneously by territorial extension with their various colonial powers (from 1957–1958) and terminated most of these treaties simultaneously three decades later (in 1983–1984). See supra note 6. The U.S. subsequently entered into TIEAs only with the Caribbean nations. See, e.g., Bruce Zagaris, OECD Report on Harmful Tax Competition: Strategic Implications for Caribbean Offshore Jurisdictions, 17 Tax Notes Int’l 1507, 1510 (1998). The U.S. has trade agreements with countries in both the Caribbean and Sub-Saharan Africa, sends foreign aid to both regions, and has expressed a desire to increase investment, trade, and aid to both regions. See supra notes 2–3. Yet, there is no agreement on tax matters with respect to the LDCs in Sub-Saharan Africa. See discussion supra Part I.

tax treaties for investment and aid to sub-saharan africa: a case study 633

Tax treaties represent a significant opportunity cost for LDCs, diverting attention and resources away from the exploration of more direct ways to increase cross-border investment. Thus, every potential tax treaty relationship with LDCs should be approached critically. If a tax treaty cannot be expected to provide sufficient benefits to investors, it should not be pursued simply to include the target country in the network of treaty countries in a myopic adherence to traditional notions about the international tax and business community. After decades of faithfulness to the promise of tax treaties, their inability to deliver in situations involving LDCs must be acknowledged. If the United States is truly committed to increasing trade and investment to the LDCs of Sub-Saharan Africa, it must pursue alternative means of achieving these goals.

This page intentionally left blank

22. it’s all in the timing: assessing the impact of bilateral tax treaties on u.s. fdi activity∗ daniel l. millimet and abdullah kumas introduction Foreign affiliate sales (FAS) grew by 11% in the 1990s, roughly double the growth rate for exports and quadruple the growth rate for worldwide gross domestic product (GDP) (Markusen 2002). In 2004, total FAS represented nearly 51% of world GDP, with world exports representing roughly half this amount (Ramondo 2007). In light of the important and expanding role of foreign direct investment (FDI) in today’s global economy, a vast literature has emerged attempting to uncover the salient factors determining the spatial and temporal pattern of FDI activity. In this chapter, we investigate one potential factor: bilateral tax treaties. Perhaps surprisingly, the empirical impact of bilateral tax treaties on FDI is very murky. This ambiguity exists amid fairly pervasive empirical evidence that cross-country variation in taxation does influence the distribution of FDI activity (e.g., Chakrabarti 2001; Gresik 2001; Gordon and Hines 2002; De Mooij and Ederveen 2003; Mutti and Grubert 2004; Blonigen 2005), as well as the fact that tax treaties are costly to negotiate and implement, yet nonetheless cover much of today’s bilateral FDI activity. Notably, the number of tax treaties in force has increased from one hundred in the 1960s to over 2,500 more recently (Egger et al. 2006). The U.S. presently belongs to roughly sixty such treaties, covering approximately 78% of total U.S. outbound FDI and 96% of total U.S. inbound FDI, with over one-third being implemented since 1990 (Blonigen and Davies 2004). While the theoretical literature on bilateral tax treaties is more developed, empirical studies are relatively sparse. Blonigen and Davies (2004, 2005) have strong positive effects of old tax treaties on FDI, but negative effects of new tax treaties, using 1980–1999 U.S. and 1983–1992 OECD data, respectively, particularly when modeling FDI in levels (as opposed to logs).1 Egger et al. (2006) also obtain a significant negative effect of new tax treaties on OECD outbound FDI

∗ The authors wish to thank Bruce Blonigen and Ron Davies for generously providing the data, as well as Bruce Blonigen, Ron Davies, Rusty Tchernis, and participants at the Texas Camp Econometrics XII, February 2007, for helpful comments. 1. Results from log specifications with country fixed effects yield positive, but insignificant results (Blonigen and Davies 2004). Mutti and Grubert (2004) also advocate the use of log-linear models when analyzing FDI.

636 daniel l. millimet and abdullah kumas

from 1985–2000 using a difference-in-difference propensity score matching estimator. Davies et al. (2007) use Swedish data on multinational firms from 1965–1998 and find no impact of tax treaties on total affiliate sales, but do find that a treaty increases the likelihood of an investment occurring. On the other hand, di Giovanni (2005) analyzes cross-border capital flows for mergers and acquisitions from 1990–1999 and finds positive effects of tax treaties. Similarly, Stein and Daude (2007) obtain a positive and statistically significant effect using data on OECD outbound FDI stocks from 1997–1999; Neumayer (2007) obtains positive effects using data on U.S. outbound FDI from 1970–2001 and inbound FDI to developing countries over the same time period. Finally, Davies (2003a) finds no effect of revisions of existing bilateral tax treaties on FDI, and Hartman (1985) and Sinn (1993) find that the expansion of activities of multinational enterprises (MNEs) is essentially independent of withholding taxes. Davies (2004) provides an excellent review.2 The mixed, and perhaps counter-intuitive, empirical results could be artifacts of the restriction implicit in the empirical approach that overshadows nearly all of these previous studies: treaties are allowed to have only a one-time, discrete effect on FDI. Given the complexity of the treaties, the political environment in which they are negotiated, and the fixed costs associated with FDI activity, it is possible that firms are aware of at least the possibility of a treaty prior to its actual formation, leading to effects of the treaty that pre-date official implementation (referred to as anticipatory effects). On the other hand, it is possible that firms may react slowly to a newly formed treaty, leading to effects of a treaty several periods after official implementation (referred to as lagged effects). To assess the timing issue empirically, we utilize data on U.S. inbound and outbound FDI stocks, flows, and FAS over the period 1980–1999. The data are from Blonigen and Davies (2004), thus enabling us to compare our findings to the existing literature. Our results are striking, indicating the importance of allowing for a more complex timing of responses to tax treaties in empirical research, particularly when analyzing U.S. inbound FDI.3 Nevertheless, allowing for this flexibility does not resolve some of the empirical puzzles present in the data. The remainder of the chapter is organized as follows: Section A describes the empirical methodology. Section B discusses the data. Section C presents the results, while Section D concludes.

2. A related literature assessing the impact of bilateral investment treaties (BITs) also fails to produce a consensus. For example, while Hallward-Driemeier (2003) fail to find a positive effect of such treaties on FDI, Egger and Pfaffermayr (2004) and Stein and Daude (2007) do. 3. Egger et al. (2006) similarly undertake some exploratory analysis of possible anticipatory and lagged effects of bilateral tax treaties involving OECD countries. The authors find little evidence of either; anticipatory effects seem nonexistent in their data, and treatment effects appear constant over a five-year window following the implementation of a tax treaty.

it’s all in the timing 637

A. Empirical methodology 1. A model with flexible timing To contrast the assumptions required for various estimators to identify the causal impact of bilateral tax treaties, we utilize the potential outcomes framework often adopted in the program evaluation literature. However, in a more general model that allows for both anticipatory and lagged effects of the treatment being analyzed—in this case, the presence of bilateral tax treaty—the framework is a bit more complex. To begin, let yijt (Dij1, Dij2,........, Dijk), t = 1,....,T

(1)

denote a measure of FDI activity involving countries i and j at time t, which in principal may depend on the tax treaty status of the two countries in all periods. Equation (1) defines potential outcomes since it reflects FDI activity in period t under any combination of historical, current, and future tax treaty status. Specifically, Dijt is an indicator variable, taking on the value one if i and j have a bilateral tax treaty in place in year t, and zero otherwise. For compactness, we denote potential outcomes as yijt (D), where D represents the complete vector of tax treaty indicators. In contrast to the existing literature, the setup in (1) allows FDI activity to depend on more than just whether a bilateral tax treaty is currently in place. For example, suppose we are considering three periods (T = 3). Thus, there are eight (2³ = 8) potential outcomes in time period t, given by yijt(0,0,0), yijt(1,0,0), yijt(0,1,0), yijt(0,0,1), yijt(1,1,0), yijt(1,0,1), yijt(0,1,1), yijt(1,1,1), t = 1,2,3.

Given this flexibility in the timing of the treaty effect, defining the treatment effect of a tax treaty between contracting countries i and j on FDI activity at time t requires defining a ‘treated’ state, say D1 and a ‘control’ state, say D0. This entails specifying a complete pattern of tax treaty status. The observation-specific treatment effect is then given by τijt(D1,D0) ≡ yijt(D1) − yijt(D0) which depends on how the treatment and control states are defined. For instance, in our previous three-period example, we may wish to compare FDI activity if countries have a tax tre aty in effect in periods 2 and 3 (but not in period 1) to FDI activity under the control state of no tax treaty in any period. Thus, D1 = [0, 1, 1] and D0 = [0, 0, 0], and the observation-specific treatment effect is given by τijt(D1,D0) ≡ yijt(0,1,1) − yijt(0,0,0).

Although relatively straightforward, the fundamental problem of causal inference is that, at most, only one state of the world (and, hence, the potential outcome) is observed for a given pair of countries at a particular point in time (Holland 1986). For example, if one of the two states – D1 or D0 – actually occurs, then one observes yijt = D1 yijt (D1) + D0yijt(D0). Thus, the effect of the treatment D1 relative to the control D0 cannot be computed; instead, it must be estimated.

638 daniel l. millimet and abdullah kumas

This estimation entails estimation of the missing counterfactual under some set of identifying assumptions. More generally, one observes (2)

where Ω denotes the set of all possible combinations of D and I[•] is an indicator function. To circumvent the missing counterfactual problem, we specify a structural relationship for the potential outcome associated with each possible pattern of tax treaty status. Define

where is a vector of observable attributes of country pair ij (including an intercept). uijt(d) captures the impact of unobservable attributes on FDI activity when the vector of treatments is given by d, d ∈ Ω. Following in the spirit of Heckman et al. (1999), if one assumes that md (xijt)=xij b (d) and b (d) = b (d′ ), d ≠ d′ except for a certain pattern of nonequivalent intercept terms, then one can obtain the following regression model

(3)

where D = 0 represents a pair of countries with no bilateral tax treaty in any period, τs+ captures the lagged effects of tax treaties (i.e., the effect of a tax treaty being in place s periods in the past), τs− captures the anticipatory effects of tax treaties (i.e., the effect of a tax treaty being in place s periods in the future), and τ0 is the contemporaneous effect of a tax treaty. We obtain (3) by assuming that effect of a tax treaty in some period s on FDI activity in some period t (s,t = 1,. . .,T) depends only on t−s (and not t or s individually), where t−s > 0 captures lagged effects and t−s < 0 captures anticipatory effects. Referring to our three-period example once more, the model in (3) contains five treatment parameters: the impact on current FDI activity of a tax treaty two periods ahead (τ2−), one period ahead (τ1−), in the current period (τ0), one period in the past (τ1+), and two periods in the past (τ2+). OLS estimation of (3) yields a consistent estimate of the τ parameters if, conditional on x, the presence or absence of a tax treaty in any time period is independent of contemporaneous (i) unobservables affecting FDI activity without a tax treaty in each time period, and (ii) unobserved, country-pair-specific gains in each time period from deviating from no tax treaty in all time periods (D = 0) to the observed pattern (D = d). In contrast, a consistent estimate of the τ parameters may be obtained under an alternative set of assumptions. Given the presence of panel data, where at least some country-pairs are observed both with and without an effective tax treaty during the sample period, then (3) may be estimated using either fixed

it’s all in the timing 639

effects (FE) or first-differences (FD). Either method allows one to decompose unobservable determinants of FDI activity as (4)

where αij represents time invariant, country-pair-specific unobservables affecting FDI activity and u˜ijt(d) represents idiosyncratic shocks to FDI between countries i and j (given D = d). Substituting (4) into (3) yields

(5)

As is well known (see, e.g., Wooldridge 2001, Chapter 10), FE estimation of (5) transforms the model by mean-differencing (i.e., subtracting yij. from both sides of the equation, where yij. represents the average value of yijt computed over all time periods). FD estimation of (5), on the other hand, transforms the model by subtracting yijt−1 from both sides of the equation. Both methods eliminate αij from the error term of the transformed model, thus allowing for consistent estimation of the remaining parameters even if x or the tax treaty variables are not independent of αij. Nevertheless, FE and FD still place restrictions on the correlations between x and the tax treaty variables and the idiosyncratic error term, u˜ijt(d). Strict exogeneity of x and the tax treaty variables are sufficient to guarantee FE and FD will both yield consistent estimates. To make the model in (5) a bit more tractable and the coefficient estimates a bit easier to interpret, the estimating equation actually employed in the analysis is given by

(6) where

t−s and s = −1,0,1,2,3+. To be clear, D˜ −i1jt equals one for country pair ij in year t if the countries will have a tax treaty become effective in the next period (zero otherwise); D˜ 0ijt equals one in the period when the tax treaty becomes effective (zero otherwise); D˜ 1ijt equals one in the period immediately after the tax treaty becomes effective (zero otherwise); D˜ ijt2 equals one in the period two periods after the tax treaty becomes effective (zero otherwise); and, D˜ 3+ ijt equals one in any period three periods or more after the tax treaty becomes effective (zero otherwise). Thus, at most only one of the tax status indicators is non-zero for any ijt combination, and τ1− captures any (short-run) anticipatory effects, τ0 captures the instantaneous response, and τs+, s = 1,2,3+, capture any lagged effects. The omitted category contains periods two or more years prior to a tax treaty becoming effective.

640 daniel l. millimet and abdullah kumas

2. A model with restricted timing While existing empirical studies of bilateral tax treaties have utilized FE estimation to remove time invariant, country-pair-specific unobservables, they do not permit the sort of flexible timing of response to a bilateral tax treaty as in (5) Instead, tax treaties are restricted to have only a one-time, discrete change in FDI activity, as in (7)

where (8)

and u ijt(0) = u∼ijt(0). Thus, (7) reduces to (5) only in the event of no anticipatory or lagged effects of tax treaties. Given the discussion earlier about the several possible dates that one could use to define the tax treaty variable (discussed in Section 3), combined with the fact that some of the effects of a tax treaty may precede the treaty (anticipatory effects) or operate with a lag (lagged effects), the restricted model in (7) seems likely to be mis-specified in the current context. A formal specification test for the restricted model in (7) is provided in Laporte and Windmeijer (2005). The intuition behind their test is quite straightforward: if (7) is correctly specified (including the assumption that x and D are strictly exogenous), then FE and FD will each yield a consistent estimate of τ0. However, in the presence of anticipatory or lagged effects, FE and FD will provide statistically different estimates of τ0. Thus, a test of equality of τ0FE and τ0FD constitutes a test of the specification in (7). The test involves estimation of the stacked model ≈

(9)

by OLS and testing H0: ϕ = 0 via a standard t-test using standard errors robust to heteroskedasticity and (serial) correlation. B. Data The data come from Blonigen and Davies (2004); thus, we provide only limited details.4 The data include information on U.S. inbound FDI from 91 countries, as well as U.S. outbound FDI to 44 countries, over the period 1980–1999. Thus, we analyze U.S. inbound and outbound FDI separately, thereby allowing the

4. The data are found at http://www.uoregon.edu/~bruceb/. We are very grateful to the authors for making the data available.

it’s all in the timing 641

effects of treaties to differ depending on the direction of investment. Three measures of FDI are utilized: (i) FDI stock, (ii) FDI flows, and (iii) FAS. In addition, we conduct the analysis using each measure in both levels and logs. As noted earlier, log specifications tend to preferred in the FDI literature given the skewness of the data, although this may not matter for the distributional approach. Nonetheless, it does provide some interesting insight into the pattern of regression results observed when using logs versus levels. The FDI data come from the U.S. Bureau of Economic Analysis (BEA) website, and are converted into millions of real 1996 U.S. dollars using the U.S. chain-type price index for gross domestic investment calculated by the Economic Report of the President. Information on U.S. bilateral tax treaties is taken from the Worldwide Tax Treaties database at Tax.com (2002). Following Blonigen and Davies (2004), we compare FDI activity with and without a tax treaty in effect, as opposed to a tax treaty being signed or in force. Given the different possibilities concerning measurement of ‘treatment’ date, allowing for flexibility in timing of responses by firms—as in the model discussed in Section A.1—seems crucial. As noted in the previous sections, each estimator relies on some assumption concerning conditional independence between the presence of an effective tax treaty and unobservable determinants of FDI activity. The United States, however, may negotiate tax treaties with countries on the basis of such unobservables (e.g., countries having historical ties to U.S., or countries for whom the gains from such a treaty are increasing over time). While some troublesome unobservables may be time invariant, non-random selection on the basis of timevarying unobservables remains a possibility. To at least partially circumvent this issue, we follow Blonigen and Davies (2004) and assess the impact of “new” treaties, where “new” treaties are those negotiated after 1979. As shown in Blonigen and Davies (2004), the relative rank of a country in terms of the stock of U.S. outbound FDI appears unrelated to the decision by the U.S. to enter into a new tax treaty with that country. The controls included in x follow the specification developed in Carr et al. (2001) and Markusen and Maskus (2001), combined with the skill adjustment applied in Blonigen et al. (2003). The specification is based on the knowledgecapital model of MNE activity. In the models analyzing both U.S. inbound and outbound FDI activity, the vector of covariates includes the sum of real GDPs, the GDP difference between the U.S. and foreign country squared, the absolute value of skill difference between the U.S. and foreign country, the distance between U.S. and foreign country, a trade cost measure for home and host country, an investment barrier measure for the host country, the interaction between the skill difference and GDP difference, and the interaction between the host trade cost and the squared skill difference. We also include a dummy variable for old treaty countries. Finally, to improve the likelihood that the tax treaty variables are strictly exogenous, in some specifications we augment the covariate set to

Inbound

FDI Stock Flow FAS Controls Sum of GDPs GDP difference squared Skill difference (inbound) Skill difference (outbound) Skill difference∗ GDP difference Distance Investment cost Trade costs (host) Trade costs (home) Trade costs (host)∗ Skill difference squared

Outbound

Mean

SD

N

Mean

SD

N

5.897.798 983.490 21.946.257

20.662.218 5.050.522 60.490.618

1470 1468 806

11.100.921 1.110.685 35.454.022

22.311.850 3.029.439 58.470.896

871 862 652

6.947.968 42.413.787.434 5.837 – 36.910.721

1.436.883 17.311.073.759 2.609 – 17.937.410

1556 1556 1556 – 1556

7.142.787 39.999.271.187 -4.992 4.992 30.289.531

1.515.579 16.894.367.031 2.412 2.412 15.328.371

881 881 881 881 881

4.988.012 29.236 80.709 33.869 3.322.917

2.391.464 2.561 4.888 44.927 2.503.452

1556 1556 1556 1556 1556

5.128.033 47.268 42.438 80.710 1.504.743

2.248.001 12.472 43.033 4.884 2.206.793

881 881 881 881 881

Notes: Data cover 1980-1999. Inbound refers to U.S. inbound FDI from 91 countries; outbound refers to U.S. outbound FDI to 44 countries. FDI variables are in millions of 1996 dollars. *GDP variables in trillions of real dollars.

642 daniel l. millimet and abdullah kumas

table 1. summary statistics

it’s all in the timing 643

include interactions between a dummy variable for “rich” countries and each of the aforementioned variables.5 Table 1 displays summary statistics. C. Results 1. Inbound FDI To begin, Table 2 presents estimates of τ0 from the restricted model given in (7).6 We utilize six dependent variables, each of the three measures of FDI activity, measured in levels and logs. In addition, we estimate each model twice, once omitting the rich country interactions and once including the interaction terms. Finally, we estimate four versions of each specification: using pooled OLS, treating αij as random effects (RE), transforming the model using FD, and transforming the model using FE. Turning to the results when FDI is measured in levels (Panels I-III), we obtain a statistically and economically significant, negative impact of an effective tax treaty on all three measures of FDI in the POLS specifications without rich country interactions. The impact remains negative and statistically significant, but of more reasonable magnitude, for FDI stocks and FAS (Panels I and III) after we include rich country interactions; the coefficient is positive and statistically insignificant for FDI flows (Panel II). However, when we switch from POLS to RE, we obtain positive and statistically insignificant tax treaty effects for each measure of FDI when we include rich country interactions; results omitting the rich country interactions remain negative, statistically significant, and quite large in magnitude. The FD estimates with rich country interactions switch signs relative to the RE estimates in Panels II and III (now becoming negative), but remain statistically insignificant at conventional levels in all three cases; results omitting the rich country interactions remain negative, but are considerably smaller in magnitude and are only statistically significant in two of three cases (excluding FDI flows (Panel II)). Lastly, the FE estimates return to being negative, statistically significant, and implausibly large when omitting the rich country interactions, but positive and statistically significant for FDI stocks and flows (Panels I and II) when including such interactions. Viewing the results from the level specifications, two observations stand out: First, as argued in Blonigen and Davies (2004), inclusion of the rich country interactions matters, and thus the estimates including such interactions are preferable. Second, modeling assumptions matter even when rich country interactions are included. Specifically, while POLS yields negative and statistically significant results for two of the three FDI measures, the other models do not. Moreover, even among the models that remove time invariant, country-level heterogeneity, the results are sensitive to modeling choice as the FE estimator 5. The set of rich countries includes EU countries, Australia, Austria, Canada, Finland, Hong Kong (China), Japan, New Zealand, Norway, Sweden, and Switzerland. 6. The full set of regression estimates is available at http://faculty.smu.edu/millimet, Tables A1–A4.

No rich country interactions Pooled OLS

RE

Pooled OLS

RE

FD

I. FDI = inbound stock (levels) New Treaty −9,879.008∗∗∗ −8,402.201∗∗∗ −741.404∗∗∗ −8,569.491∗∗ (1.123.737) (3.051.727) (270.829) (3.357.197) N 1470 1470 1335 1470 Spec test p = 0.021

−413.884∗∗∗ (146.169) 1470

164.034 (211.520) 1470

104.868 (64.130) 1335

397.283∗ (211.395) 1470 p = 0.078

II. FDI = inbound flow (levels) New Treaty −1,626.135∗∗∗ −1,508.543∗∗ (293.048) (590.087) N 1468 1468 Spec test

11.715 (40.999) 1468

43.087 (53.558) 1468

−54.535 (88.063) 1326

110.631∗ (57.698) 1468 p = 0.074

−782.477∗∗∗ (283.697) 806

307.142 (717.502) 806

641.969 −63.279 (242.780) (798.905) 686 806 p = 0.430

III. FDI = inbound FAS (levels) New Treaty −27,581.878∗∗∗ −15,899.797∗∗ (3.776.081) (6.889.170) N 806 806 Spec test

FD

Rich country interactions

−177.784 (120.586) 1326

FE

−1,576.250∗∗ (728.756) 1468 p = 0.057

−1,850.999∗∗ −15,388.288∗∗ (804.066) (7.075.517) 686 806 p = 0.059

FE

644 daniel l. millimet and abdullah kumas

table 2. regression estimates: inbound fdi

IV. FDI = inbound stock (logs) New Treaty −0.121 0.306 (0.273) (0.427) N 1470 1470 Spec test

−0.085 (0.060) 1335

0.299 (0.445) 1470 p = 0.404

−0.605∗∗ (0.299) 1470

0.281 (0.435) 1470

−0.048 (0.057) 1335

0.347 (0.450) 1470 p = 0.393

V. FDI = inbound flow (logs) New Treaty −0.236 (0.254) N 1468 Spec Test

0.255 (0.316) 1468

−0.973 (01.20) 1326

0.481 (0.381) 1468 p = 0.212

−0.27 (0.268) 1468

0.416 (0.324) 1468

−0.924 (01.192) 1326

0.551 (0.407) 1468 p = 0.199

VI. FDI = inbound FAS (logs) New Treaty −1.169∗∗∗ (0.316) N 806 Spec test

1.250∗∗∗ (0.399) 806

−0.008 (0.164) 686

1.353∗∗∗ (0.422) 806 p = 0.003

−1.542∗∗∗ (0.333) 806

1.283∗∗∗ (0.413) 806

−0.001 (0.167) 686

1.422∗∗∗ (0.452) 806 p = 0.004 it’s all in the timing 645

Notes: FAS = foreign affiliate sales. Standard errors in parentheses are robust in pooled OLS models and clustered in random effects (RE), first-differenced (FD), and fixed effects (FE) models. Specification test is from Laporte and Windmeijer (2005) and tests equality between the coefficient on New Treaty in the FD and FE models. ∗, ∗∗, ∗∗∗ denotes statistical significant at the 10%, 5%, and 1% level, respectively. See text for further details, as well as list of covariates included.

646 daniel l. millimet and abdullah kumas

yields a positive and statistically significant of an effective tax treaty on FDI stocks and flows, whereas FD does not (in fact, two of the three point estimates are negative). The differences between the FD and FE estimates suggest the possibility of mis-specification. To formally test the equivalence of the FD and FE estimates, we utilize the specification test proposed in Laporte and Windmeijer (2005). The results are shown in the row labeled Specification Test. We reject equality of the FD and FE estimates at the p < 0.10 level for all three FDI measures when excluding rich country interactions, and two of the three measures (Panels I and II) when including the interactions. Thus, the assumption of no anticipatory and lagged effects of tax treaties is not supported by the data. To relax this assumption, then, we estimate the model given in (6) via FD and FE. The estimates of the τ parameters for the level specifications are given in Panel I of Table 3.7 Concentrating on the models including rich country interactions, three findings emerge: First, the FD and FE estimates are much more closely aligned; in only one situation (the coefficient on New Treaty two years prior for FDI flows) are the coefficients of the opposite sign, but then each is statistically insignificant. Second, for both FDI stocks and flows, we find positive and statistically significant lagged effects of tax treaties, particularly in the FE models. In other words, the FDI-inducing impact of an effective tax treaty is not realized until a couple of years after the treaty becomes effective. This may suggest that the most important FDI-inducing component of bilateral tax treaties is the reduction in uncertainty in the foreign tax environment, as such diminished uncertainty may be realized after a lag. It is also consonant with the gradualism argument in Chisik and Davies (2004b), where declines in tax rates may be gradual since tax treaties need to be self-enforcing. Nevertheless, these effects are very modest in economic terms, representing roughly 0.03–0.04 standard deviations. Finally, consonant with the results in Table 2, we find no impact of an effective tax treaty on FAS once rich country interactions are included. Thus, we find some evidence of positive effects of a new tax treaty on the level of U.S. inbound FDI stocks and flows, but not FAS, although the effects tend to be small and operate with a lag. As noted previously, Blonigen and Davies (2004), Mutti and Grubert (2004), and others advocate the estimation of models of FDI activity in logs given the skewed nature of the data. Panels IV–VI in Table 2 present the estimates for the restricted timing model when we log all (non-binary) variables.8 In the interest of brevity, we focus on the major findings: First, for FDI stocks and flows, we fail

7. The full set of results is available at http://faculty.smu.edu/millimet, Tables A5–A6. 8. Note, Blonigen and Davies (2004) exclude the two interaction terms — between the skill difference and GDP difference and between the host trade cost and the squared skill difference — in their log specifications since the log of the interaction is collinear with the other variables entered in the model. However, rather than taking the log of the

it’s all in the timing 647

to find any statistically significant impact of an effective tax treaty in the RE, FD, and FE models, either with or without rich country interactions. Moreover, in all four cases, we fail to reject equality between the FD and FE estimates at conventional levels of statistical significance. Second, we find positive, statistically significant, and (unreasonably) large effects on FAS using the RE and FE estimators either with or without rich country interactions; negative, statistically significant, and (unreasonably) large effects using POLS. Finally, we reject equality of the FD and FE estimates using the Laporte and Windmeijer (2005) specification test for FAS either with or without rich country interactions. Turning to the more flexible model for FAS in Table 3, we again find much greater alignment between the FD and FE estimates.9 In particular, in the models including rich country interactions, we find positive and statistically significant lagged effects of new tax treaties according to both the FD and FE estimates, with the magnitudes remaining quite large. Moreover, the FE estimates also indicate fairly sizeable, statistically significant anticipatory effects of tax treaties. Perhaps surprising, not one of the coefficients reflecting the contemporaneous effect of tax treaties is statistically significant in either Panel I or II of Table 3 once rich country interactions are included. In sum, then, the regression analysis yields a positive, statistically significant, and relatively robust effect of an effective bilateral tax treaty when one allows for a more flexible timing of the impact on U.S. inbound FDI stocks and flows (in levels) and FAS (in logs) and includes rich country interactions. We now turn to the regression results for U.S. outbound FDI. 2. Outbound FDI Table 4 presents estimates of τ0 from the restricted model given in (7). Results from the flexible timing models are presented in Table 5.10 In the interest of brevity, we focus on the major findings. First, when estimating the models in levels, we only obtain one statistically significant effect of a new tax treaty when using an estimation method other than POLS: FDI stocks (Panel I) when using RE without rich country interactions. For the remainder of the non-POLS models in levels, the estimated impact of an effective tax treaty is statistically insignificant. Moreover, we fail to reject equality between the FD and FE estimates at conventional levels in all cases. The point estimates in the FE models with rich country interactions are positive, however, for all three FDI measures. Second, when estimating the models in logs, we obtain a positive and statistically significant impact of an effective tax treaty on FAS using POLS and FD both with and without rich country interactions. For log FDI stocks and flows, all estimation methods yield a statistically insignificant coefficient on an interactions, we include the interactions of the logs so that these variables remain in the model. The impact on the results is minor. 9. The full set of results is provided in the Appendix, Tables A7–A8. 10. The full set of regression estimates is available at http://faculty.smu.edu/millimet, Tables A9–A16.

FDI stock

FDI flow

FE

FD

Foreign affiliate sales

FD

FE

FD

FE

FD

FE

FD

−792.776∗∗

-3,390.553∗∗

119.244

150.615

−118.272 −697.575

207.093∗∗ 101.114

(345.011)

(1.602.914)

(97.306) (174.131)

(129.734) (442.708)

(84.192)

(68.785) (724.899)

−1,442.359∗∗

−2.362.26

141.853

236.397

−324.234 −882.836∗

140.997

44.560

(559.676)

(2.095.459)

(85.378)

(146.978) (205.848) (489.214)

−1,944.282∗∗

−2.853.03

291.125∗ 468.995∗∗ −269.959 −335.563

327.284∗

135.512

(751.596)

(2.275.323)

(169.045) (224.750) (329.431) (494.453)

(187.813)

(102.458) (1.734.346)

198.173

−9.177

FE

FD

FE

I. Levels (New treaty)−1

(New treaty)0

(New treaty)+1

(New treaty)+2

−2,741.414∗∗∗ −4.195.202

173.614

(963.051)

(130.712) (177.948) (421.250) (618.067)

(2.850.485)

(New treaty)+3+ −3,380.269∗∗∗ −10,601.061∗∗ 208.984 (1.198.871)

406.184∗∗ −655.376 −929.017

−1,659.765∗∗ −1.889.331

392.881

468.737

(452.051)

(643.487)

285.034

408.232

(3.275.731)

(746.170)

(852.161)

−4,156.651∗∗ −2.560.045

1.048.683

899.173

−3,509.657∗∗ −4.804.728

(118.601) (51.008) (1.395.519)

(177.392) (37.541)

647.114∗∗ −836.643 −2,100.313∗ 319.297∗

(2.551.893)

(3.903.684)

(1.019.70) (1.265.968)

−6,410.244∗∗ −7,375.583∗

1.361.361

(2.599.012)

(1.091.721) (1.328.889)

(4.308.350)

146.212∗∗ -8,010.765∗∗ −22,414.127∗∗ 1.708.137

1.192.506

667.951

(4.774.681)

(128.502) (274.90)

(566.099) (1.103.691) (190.937) (59.406) (3.179.010)

(9.711.917)

(1.190.428) (1.545.290)

Rich interactions No

No

Yes

Yes

No

No

Yes

Yes

No

No

Yes

Yes

N

1002

908

1002

898

990

898

990

402

462

402

462

908

648 daniel l. millimet and abdullah kumas

table 3. regression estimates: inbound fdi

II. Logs (New treaty)−1

(New treaty)0

(New treaty)+1

(New treaty)+2

0.774

0.164

0.845

0.292

0.745

0.853

0.817

0.941

0.029

0.350

0.061

0.462∗

(0.558)

(0.699)

(0.567)

(0.649)

(0.843)

(0.988)

(0.853)

(0.984)

(0.104)

(0.218)

(0.116)

(0.236)

0.661

0.173

0.785

0.334

−0.453

−0.147

−0.385

0.038

0.113

0.381

0.171

0.524

(0.577)

(0.573)

(0.588)

(0.520)

(1.567)

(0.988)

(1.601)

(1.016)

(0.227)

(0.372)

(0.234)

(0.380)

-0.270

−0.072

−0.173

0.052

0.072

0.512

0.066

0.676

0.171

0.495

0.242

0.650∗

(0.628)

(0.755)

(0.623)

(0.770)

(1.550)

(0.895)

(1.631)

(0.931)

(0.246)

(0.348)

(0.269)

(0.376)

−0.331

−0.049

−0.261

0.094

−0.682

−0.153

−0.744

0.031

1.070∗

1.250∗∗∗

1.165∗∗

1.415∗∗∗

(0.398)

(0.634)

(0.450)

(0.630)

(1.752)

(1.108)

(1.851)

(1.127)

(0.534)

(0.455)

(0.569)

(0.513)

(New treaty)+3+ 0.286

0.584

0.277

0.715

0.344

0.707

0.267

0.832

1.219∗∗

1.619∗∗∗

1.311∗∗

1.852∗∗∗

(0.309)

(0.488)

(0.350)

(0.488)

(1.715)

(0.520)

(1.794)

(0.531)

(0.557)

(0.511)

(0.604)

(0.592)

908

1002

908

1002

898

990

898

990

402

462

402

462

Rich interactions N

Notes: (New treaty)t is an indicator for t years prior to (after) a new tax treaty becoming effective if t is negative (positive). See Table 2 for further details.

it’s all in the timing 649

No rich country interactions Pooled OLS

RE

FD

−5,585.681∗ (3.026.619) 871

Rich country interactions FE

Pooled OLS

RE

FD

FE

−381.71 −5.205.76 (338.974) (3.133.447) 820 871 p = 0.123

−1,794.917∗∗∗ (648.569) 871

−166.38 (1.462.261) 871

−3.348 (212.923) 820

291.158

−748.083 (487.060) 862

−70.349 (185.637) 808

−664.012 (503.507) 862 p = 0.252

−205.719∗ (106.355) 862

−7.516 (270.804) 862

−40.227 (189.943) 808

76.157 (301.767) 862 p = 0.777

−8.796.70 (6.160.690) 652

−928.791 (817.635) 597

−7.585.57 (6.464.825) 652 p = 0.305

1.958.82 (652.0) 652

4.147.60 (652.0) 652

527.043 (597.0) 597

4.239.89 (652.0) 652 p = 0.377

I. FDI = outbound stock (levels) New treaty N Spec test

−7,798.073∗∗∗ (1.236.821) 871

(1.519.070) 871 p = 0.858

II. FDI = outbound flow (levels) New Treaty N Spec test

−844.018∗∗∗ (183.239) 862

III. FDI = outbound FAS (levels) New Treaty N Spec Test

−22,339.600∗∗∗ (3.537.605) 652

650 daniel l. millimet and abdullah kumas

table 4. regression estimates: outbound fdi

IV. FDI = outbound stock (logs) New Treaty N Spec test

−0.329 (0.237) 871

0.166 (0.278) 871

−0.006 (0.128) 820

0.194 (0.255) 871 p = 0.536

−0.163 (0.260) 871

0.082 (0.350) 871

−0.005 (0.124) 820

0.116 (0.352) 871 p = 0.801

−0.436 (0.592) 862

−1.908 (01.552) 808

−0.378 (0.704) 862 p = 0.381

−0.601 (0.376) 862

−0.648 (0.596) 862

−1.895 (01.509) 808

−0.717 (0.784) 862 p = 0.516

0.525 (0.345) 652

0.120∗ (0.066) 597

0.537 (0.350) 652 p = 0.243

0.622∗∗ (0.268) 652

0.446 (0.321) 652

0.089∗∗ (0.042) 597

0.441 (0.323) 652 p = 0.281

V. FDI= outbound flow (logs) New treaty N Spec test

−0.429 (0.359) 862

VI. FDI = outbound FAS (logs) New treaty N Spec test

0.441∗ (0.248) 652

it’s all in the timing 651

Notes: See Table 2.

FDI stock

FDI flow

FD

FE

FD

FE

FD

FE

−452.497

−1.584.768

30.011

159.576

−260.365 −323.995

(347.861)

(1.981.188)

(181.527) (1.026.708)(186.662) (324.705)

−954.353

−2.586.307

−1.681

(705.039)

(2.527.082)

(396.670) (1.332.280) (199.556) (446.427)

Foreign affiliate sales

FD

FE

FD

FE

FD

FE

−125.750

−104.477 40.166

−572.825

1.535.672

3.783.134

(170.618) (180.551) (1.851.091)

(4.640.540)

(1.646.462) (3.113.790)

−74.593

−54.814

−640.356

2.420.225 5.354.599

(177.323)

(292.249) (2.749.457)

(5.892.914)

(2.117.716) (3.943.343)

I. Levels (New treaty)−1

(New treaty)0

(New treaty)+1

(New treaty)+2

164.011

−254.080 −420.701

−1.417.585

v3.122.311

−28.086 95.479

(969.428)

(2.941.577)

(514.950) (1.579.331) (303.492) (622.719)

−135.342 −374.473

151.251

24.757

−1.576.492

−1.169.706

3.650.820 7.198.811

(247.90)

(450.961) (3.852.599)

(8.136.705)

(2.961.939) (5.731.671)

181.931

127.756

−1.794.608

−3.532.768

−115.002 452.418

−3.954.420

−2.557.494

2.045.257

(1.186.178)

(3.160.344)

(594.669) (1.639.714) (317.158) (565.538)

(239.425) (318.958) (3.617.902)

(7.481.116)

(2.474.179) (5.346.933)

−6.533.865

−484.834 727.494

−407.356 −89.392

−4.869.929

−6.238.731

1.917.582

(New treaty)+3+ −2,260.770∗ (1.341.265)

−129.483 −427.650

−944.645

−681.646∗ −1.104.22

6.257.995

8.473.180

(4.433.825)

(716.803) (2.474.381) (362.792) (836.134)

(257.070) (555.484) (4.693.189)

(9.580.181)

(3.680.316) (7.477.991)

Rich interactions No

No

Yes

Yes

No

No

Yes

Yes

No

No

Yes

Yes

N

670

621

670

601

652

601

652

403

449

403

449

621

652 daniel l. millimet and abdullah kumas

table 5. regression estimates: outbound fdi

II. Logs (New treaty)−1

(New treaty)0

(New treaty)+1

(New treaty)+2

0.033

0.287∗∗

0.048

0.394∗∗

−0.187

0.165

−0.035

−0.085

−0.001

0.253

0.009

0.259

(0.052)

(0.140)

(0.049)

(0.159)

(0.299)

(0.795)

(0.388)

−0.071

0.165

−0.043

0.289

−1.027

−0.924

−0.962

(0.763)

(0.080)

(0.189)

(0.080)

(0.188)

−1.188

0.038

0.223

0.044

0.203

(0.081)

(0.164)

(0.093)

(0.188)

(1.539)

(1.328)

(1.525)

(1.343)

(0.120)

(0.239)

(0.121)

(0.246)

−0.061

0.181

−0.028

0.298

0.024

0.029

0.332

−0.194

0.063

0.144

0.067

0.080

(0.087)

(0.162)

(0.102)

(0.182)

(1.395)

(1.299)

(1.331)

(1.384)

(0.145)

(0.268)

(0.149)

(0.284)

−0.027

0.253

−0.003

0.378∗∗

−0.051

0.568

−0.032

0.277

0.070

0.140

0.075

0.121

(0.089)

(0.162)

(0.107)

(0.187)

(1.408)

(0.868)

(1.504)

(0.987)

(0.165)

(0.30)

(0.174)

(0.301)

0.048

−0.090

0.176

−3.129∗

−1.458

−3.506∗

−1.957

0.199

0.359

0.161

0.323

(0.107)

(0.241)

(0.123)

(0.290)

(1.856)

(1.363)

(1.868)

(1.560)

(0.226)

(0.408)

(0.244)

(0.405)

621

670

621

670

601

652

601

652

403

449

403

449

(New treaty)+3+ −0.094

Rich interactions N

Notes: See Table 4.

it’s all in the timing 653

654 daniel l. millimet and abdullah kumas

effective tax treaty. Moreover, for all three FDI measures, we continue to fail to reject equality between the FD and FE estimates. Finally, the results from the more flexible specifications with rich country interactions indicate a statistically insignificant impact of an effective tax treaty on all three FDI measures in levels (Panel I), as well as FAS in logs. The more flexible FE model, however, indicates some positive and statistically significant (anticipatory and lagged) effects on log FDI stocks; the more flexible FD model indicates a negative and statistically significant effect on log FDI flows three or more years after a bilateral tax treaty becomes effective. In sum, then, the regression analysis for U.S. outbound FDI yields a much more muted impact of effective tax treaties relative to U.S. inbound FDI. Specifically, there is no statistically meaningful evidence of a non-zero effect of an effective tax treaty, even in the more flexible specifications, when analyzing FDI in levels. There is, however, some evidence of a positive and statistically significant impact of an effective tax treaty on FDI stocks and FAS in the log models, the former (latter) occurring with a lag (instantaneously), and a negative lagged effect on log FDI flows.

conclusion Economists have been a bit puzzled by bilateral tax treaties because of the divergence of the empirical and theoretical results in the literature, the fragility of existing empirical estimates, as well as the extreme magnitudes obtained in some specifications. Whereas the theoretical literature suggests that such treaties can be FDI-inducing, the empirical (and legal) literature disputes these claims in practice. In this chapter, we have re-examined the panel data set from Blonigen and Davies (2004) on U.S. inbound and outbound FDI, which spans the period 1980–1999, in light of recent methodological advances in the program evaluation literature that emphasize the role of timing of policy effects. In doing so, we reach two primary conclusions: First, regression estimates of bilateral tax treaty effects are indeed quite fragile. Not only do statistical modeling assumptions matter, but assumptions concerning the timing of the effect of tax treaties are important. In our preferred specifications based on the more flexible Laporte and Windmeijer (2005) model with rich country interactions, we find some evidence of a positive, lagged response to tax treaties becoming effective. Although contrary to Egger et al. (2006), the coefficients are of a reasonable magnitude. Moreover, the positive, lagged response is consistent with the gradualism argument in Chisik and Davies (2004b), who assert that tax rates can only be reduced gradually under tax treaties due to the self-enforcing nature of such agreements. Our findings are also consistent with the tax certainty role of tax treaties, in which such certainty is only revealed over time, again perhaps due to lack of formal enforcement of such treaties. However, these positive,

it’s all in the timing 655

lagged effects are statistically significant only for U.S. inbound FDI stocks and flows in levels, and inbound FAS and outbound stocks in logs. Second, the regression estimates indicate some asymmetric impacts of effective tax treaties on U.S. inbound and outbound FDI. As previously stated, in our preferable specifications for U.S. inbound FDI, we obtain positive effects of an effective bilateral tax treaty on FDI stocks and flows (in levels) and FAS (in logs) several years after the tax treaty becomes effective. For U.S. outbound FDI, we obtain some less robust evidence of positive effects of an effective bilateral tax treaty on FDI stocks and FAS (in logs), the former occurring with a lag. However, we also obtain some evidence of negative effects of an effective bilateral tax treaty on FDI flows (in logs) several years after the tax treaty becomes effective. Finally, while relaxing the assumption of no anticipatory or lagged effects of a bilateral tax treaty resolves some of the puzzle with respect to the empirical effects of such treaties, some issues remain unresolved. Most notably, the decision to model FDI activity in levels or logs remains crucial, as does the decision to focus on inbound or outbound FDI.

references Blonigen, B.A. (2005). “A review of the empirical literature on FDI determinants,” Atlantic Economic Journal, 33, pp. 383–403. Blonigen, B.A., R.B. Davies and K. Head (2003). “Estimating the knowledge-capital model of the multinational enterprise: comment,” American Economic Review, 93, pp. 980–994. Blonigen B.A. and R.B. Davies (2004).”The effects of bilateral tax treaties on US FDI activity,” International Tax and Public Finance, 11, pp. 601–622. —— (2005). “Do bilateral tax treaties promote foreign direct investment?” in J. Hartigan, ed., Handbook of International Trade: Economic and Legal Analysis of Laws and Institutions (Oxford: Blackwell Publishing). Carr, D., J.R. Markusen, and K. E. Maskus (2001). “Estimating the knowledge-capital model of the multinational enterprise,” American Economic Review, 91, pp. 693–708. Chakrabarti, A. (2001). “The determinants of foreign direct investment: sensitivity analyses of cross-country regressions,” Kyklos, 54, pp. 89–114. Chisik, R. and R.B. Davies (2004a). “Asymmetric FDI and tax-treaty bargaining: theory and evidence,” Journal of Public Economics, 88, pp. 1119–1148. —— (2004b). “Gradualism in tax treaties with irreversible foreign direct investment,” International Economic Review, 45, pp. 113–139. Cole, M.A., R.J.R. Elliot and P.G. Fredriksson (2006). “Endogenous pollution havens: does FDI influence environmental regulations?” Scandinavian Journal of Economics, 108, pp. 157–178. Davies, R.B. (2003a). “Tax treaties, renegotiations, and foreign direct investment,” Economic Analysis and Policy, 32, pp. 251–273. —— (2003b). “The OECD model tax treaty: tax competition and two-way capital flows,” International Economic Review, 44, pp. 725–753.

656 daniel l. millimet and abdullah kumas —— (2004). “Tax treaties and foreign direct investment: potential versus performance,” International Tax and Public Finance, 11, pp. 775–802. Davies, R.B. and T.A. Gresik (2003). “Tax competition and foreign capital,” International Tax and Public Finance, 10, pp. 127–145. Davies, R.B., P.J. Norbäck and A. Tekin-Koru (2007). “The effect of tax treaties on multinational firms: new evidence from Microdata,” Oxford University Centre for Business Taxation, Working Paper No. 07/21. De Mooij, R.A. and S. Ederveen (2003).”Taxation and foreign direct investment: a synthesis of empirical research,” International and Tax Public Finance, 10, pp. 673–693. di Giovanni, J. (2005), “What drives capital flows? The case of cross-border M&A activity and financial deepening,” Journal of International Economics, 65, pp. 127–149. Egger, P. and M. Pfaffermayr (2004). “The impact of bilateral investment treaties on foreign direct investment,” Journal of Comparative Economics, 32, pp. 788–804. Egger, P., M. Larch, M. Pfaffermayr, and H. Winner (2006). “The impact of endogenous tax treaties on foreign direct investment: theory and evidence,” Canadian Journal of Economics, 39, pp. 901–931. Gordon, R. and J.R. Hines, Jr. (2002). “International taxation,” in A.J. Auerbach and M. Feldstein, eds., Handbook of Public Economics, 4 (Amsterdam: Elsevier). Gravelle, P. (1988). “Tax treaties: concepts, objectives and types,” International Bureau of Fiscal Documentation Bulletin, 522, pp. 522–526. Gresik, T.A. (2001). “The taxing task of taxing transnationals,” Journal of Economic Literature, 39, pp. 800–838. Gubert, H. (2003). “The tax burden on cross-border investment: company strategies and country responses,” CESifo, Working Paper No. 964. Hallward-Driemeier, M. (2003). “Do bilateral investment treaties attract foreign direct investment? A bit . . . and they could bite,” World Bank, Working Paper No. 3121, (Washington, D.C.: World Bank). Heckman, J.J., R.J. LaLonde and J.A. Smith (1999). “The economics and econometrics of active labor market programs,” in O. Ashenfelter and D. Card, eds., Handbook of Labor Economics, 3, (Amsterdam: Elsevier). Holland, P. (1986). “Statistics and causal influence,” Journal of the American Statistical Association, 81, pp. 945–960. Janeba, E. (1995). “Corporate income tax competition, double taxation treaties, and foreign direct investment,” Journal of Public Economics, 56, pp. 311–325. Laporte, A. and F. Windmeijer (2005). “Estimation of panel data models with binary indicators when treatment effects are not constant over time,” Economics Letters, 88, pp. 389–396. Markusen, J. (2002). Multinational Firms and the Theory of International Trade (Cambridge: MIT Press). Markusen, J.R. and K.E. Maskus (2001). “Multinational firms: reconciling theory and evidence,” in M. Blomstrom and L.S. Goldberg, eds., Topics in Empirical International Economics: A Festschrift in Honor of Robert E. Lipsey (Chicago, IL: University of Chicago Press for National Bureau of Economic Research), pp. 71–97. Mutti, J. and H. Grubert (2004). “Empirical asymmetries in foreign direct investment,” Journal of International Economics, 62, pp. 337–358. Neumayer, E. (2007). “Do double taxation treaties increase foreign direct investment to developing countries?” Journal of Development Studies, 43, pp. 1501–1519.

it’s all in the timing 657

Radaelli, C.M. (1997). The Politics of Corporate Taxation in the European Union (London: Routledge Research in European Public Policy). Ramondo, N. (2005). “Size, geography, and foreign direct investment,” unpublished manuscript, (Texas: University of Texas). Sinn, H.W. (1993). “Taxation and the birth of foreign subsidiaries,” in H. Heber and N. van Long, eds., Trade, Welfare, and Economic Policies: Essays in Honor of Murray C. Kemp (Ann Arbor, MI: University of Michigan Press). Stein, E. and C. Daude (2007). “Longitude matters: time zones and the location of foreign direct investment,” Journal of International Economics, 71, pp. 96–112. Wooldridge, J.M. (2001). Econometric Analysis of Cross Section and Panel Data (Cambridge, MA: MIT Press).

This page intentionally left blank

23. do double taxation treaties increase foreign direct investment to developing countries?∗ eric neumayer introduction Developing countries1 sign double taxation treaties (DTTs) in order to attract more foreign direct investment (FDI). They succumb to restrictions on their ability to tax corporate income from foreign investors, which can only pay off if more FDI is the reward.2 But do DTTs attract more FDI to developing countries? This is the topic of the analysis provided here. There are reasons to presume that DTTs can increase FDI. Double taxation occurs if a multinational company (MNC) pays tax on the same corporate income earned from economic activity in a foreign country twice: once to the tax authorities of the foreign country, who are hosts to the economic activity, and once to the tax authorities of the home country, in which the company is domiciled. By burdening economic activity in a foreign country twice, double taxation can represent an obstacle or barrier to foreign investment, thus distorting the efficient allocation of scarce financial resources across the countries of the world. Yet DTTs can also dampen FDI in as much as they reduce tax avoidance, tax evasion and other more or less legal tax-saving strategies such as transfer pricing by ∗ This chapter was reprinted with permission from the Journal of Development Studies. The chapter was originally published as “Do double taxation treaties increase foreign direct investment to developing countries?” 43 Journal of Development Studies 1501 (2007). 1. For the purpose of this chapter, the category of developing countries refers to all countries other than the United States and Canada, Western Europe, Japan, Australia, and New Zealand. 2. For the purpose of this chapter, we presume that a higher FDI inflow is beneficial to the host nation. This presumption can of course be contested (see I. De Soysa and J.R. Oneal (1999), “Boon or bane? Reassessing the productivity of foreign direct investment,” American Sociological Review, 64, pp. 766–782.). It can also be contingent on the existence of other political or institutional factors. For example, see Hermes and Lensink (2003),“Foreign direct investment, financial development and economic growth,” Journal of Development Studies, 40, 1, pp. 142–163, who find that a well-developed financial system is an important precondition for FDI to have a stimulating effect on economic growth. High FDI inflows can also have effects on, for example, regional inequality that might be undesirable (for evidence on China, see Zhang and Zhang (2003), “How does globalisation affect regional inequality within a developing country? Evidence from China,” Journal of Development Studies, 39, 4, pp. 47–67).

660 eric neumayer

multinational companies.3 The 2003 Revision to the Commentary to the treaty model of the Organisation for Economic Co-operation and Development (OECD) explicitly mentions prevention of tax avoidance as an objective of DTTs; however, this has to be seen in the context of “increased opportunities for tax avoidance” made possible by the growing and increasingly complex web of DTTs among countries in the first place.4 Despite the large and increasing number of DTTs concluded, there exists little evidence on their ability to attract FDI. This is surprising given that this issue is of great importance to developing countries. They invest time and other scarce resources to negotiate, conclude, sign, and ratify DTTs. Also, such treaties typically imply a non-trivial restriction on their authority to tax corporate income from foreign investors. If no increase in FDI can be expected, then the effort spent concluding DTTs would be wasted and the costs imposed would fail to be recovered. This chapter is structured as follows: The next section describes the importance of foreign investment to developing countries and illustrates the growth of DTTs. We then review the existing empirical studies, present our research design, and report results. The final section concludes with the finding that DTTs are effective in attracting FDI, but only in the group of middle-, not low-, income developing countries. A. DTTs and FDI FDI has dramatically increased in the past several decades to become a major force in the worldwide allocation of funds and technology. Prior to 1970, world trade generally grew at a greater pace than that of FDI, but in the decades since that time, the flow of FDI has grown at more than twice the pace of the growth of worldwide exports. By the early 1990s, the sales of worldwide exports were eclipsed by the sales of foreign affiliates of multinational companies.5 Not only has the flow of FDI increased worldwide, but the importance of FDI as a source of funds to developing countries in particular has also significantly increased. Private international flows of financial resources have become increasingly important to developing countries. In the 1980s, tight budgets, the debt crisis,

3. Blonigen and Davies (2002), “Do bilateral tax treaties promote foreign direct investment?” Working Paper, No. 8834 (Boston, MA: National Bureau of Economic Research); Egger et al. (2004), “The impact of endogenous tax treaties on foreign direct investment: theory and empirical evidence,” Working Paper (University of Innsbruck, Department of Economics and Statistics). 4. B. Arnold (2004), “Tax treaties and tax avoidance: the 2003 revisions to the commentary to the OECD Model,” Bulletin for International Fiscal Documentation, 58, p. 244. 5. J.H. Dunning (1998), “The changing geography of foreign direct investment,” in N. Kumar, ed., Globalization, Foreign Direct Investment and Technology Transfers: Impacts on and Prospects for Developing Countries (New York: Routledge).

do double taxation treaties increase foreign direct investment 661

and an overall decreased interest in providing traditional development aid led to a decline in official development assistance from the developed world. When capital flows to developing nations began to rise again in the latter part of that decade, the flows were increasingly composed of FDI.6 Only very recently have aid flows slightly increased again in the wake of the so-called Monterrey Consensus; however, in 2003, FDI was the largest component of the net resource flows to developing countries and this is bound to remain the case for some time to come.7 Although the developed countries remain both the dominating source and the major recipient of FDI, their dominance has decreased over time: in 2003, developing countries received almost 31% of FDI as opposed to only about 20% in the 1980s.8 Indeed, FDI inflows per unit of GDP are much higher in many developing countries than in developed ones.9 It was during this same period that DTTs between developed and developing countries proliferated, and in light of the importance of FDI, particularly to developing nations, the extent to which these two phenomena are causally related warrants careful scrutiny. In their aim to increase FDI inflows, developing countries have resorted to bilateral treaties to signal their commitment to stable, correct, and often favorable treatment of foreign investors. By signing DTTs, developing countries provide foreign investors with security and stability as regards the issue of taxation in addition to the relief from double taxation. By signing bilateral investment treaties (BITs), developing countries commit to granting certain relative standards, such as national treatment (foreign investors may not be treated any worse than national investors, but may be treated better and, in fact, often are) and mostfavored nation treatment (privileges granted to one foreign investor must be granted to all foreign investors). They also agree to guarantee certain absolute standards of treatment, such as fair and equitable treatment for foreign investors in accordance with international standards after the investment has taken place. BITs typically ban discriminatory treatment against foreign investors and include guarantees of compensation for expropriated property or funds, and free transfer and repatriation of capital and profits. Further, the BIT parties agree to submit to binding dispute settlement should a dispute concerning these provisions arise.10

6. H. Zebregs (1998), “Can the neoclassical model explain the distribution of foreign direct investment across developing countries?” IMF Working Paper, WP/98/139 (Washington, D,C,: International Monetary Fund). 7.UNCTAD (2003 - WIR03), World Investment Report 2003: FDI Policies for Development National and International Perspectives (New York and Geneva: United Nations). 8. UNCTAD (2004 - WIR04), World Investment Report 2004: The Shift Toward Services (New York and Geneva: United Nations). 9. Ibid. 10. UNCTAD (1998), Bilateral Investment Treaties in the Mid-1990s (New York and Geneva: United Nations).

662 eric neumayer

Efforts aimed at avoiding double taxation have a long history: the first DTTs were concluded well before the first BITs were signed. According to Easson, the treaty between Austria-Hungary and Prussia from 1899 represents the first modern DTT, whereas the first BIT was signed between Germany and Pakistan in 1959.11 Multilateral organizations such as the League of Nations (and later the United Nations) and the Organisation for European Economic Co-operation (later known as the Organisation for Economic Co-operation and Development) also promoted DTTs from an early stage. Until the late 1960s, DTTs were mainly concluded among developed countries, but since then an increasing number of treaties has been concluded between developed and developing countries (and, to a smaller extent, among developing countries).12 This resembles the spread and diffusion of BITs around the world.13 By the end of the 1960s, there were 322 treaties, which rose to 674 by the end of the 1970s and to 1,143 by the end of the 1980s. The number of DTTs worldwide grew rapidly in the 1990s and by 2002 there were 2,255 DTTs worldwide.14 In 2002, China topped the list of developing countries, having concluded 21 DTTs with Organisation for Economic Co-operation and Development (OECD) countries, followed by the Czech and Slovak Republics, India, Poland, and South Korea with 20 treaties each, Hungary and Romania (19), Russia (18), Bulgaria, Indonesia, Malaysia, Mexico, Philippines, South Africa, and Thailand (17), Argentina, Latvia, and Pakistan (16), Brazil, Estonia, Lithuania, Morocco, and Tunisia (15). Most of these are major hosts of FDI; however, in the middle range are countries like Zambia (12), Bangladesh (10), Barbados, Côte d’Ivoire, and Zimbabwe (8) that are not particularly known as major recipients of FDI. At the bottom end are a great many countries that have concluded either zero or very few DTTs. A list of the number of DTTs with OECD countries in 2002 for all the countries in the sample appears in Appendix 1. For reasons to be explained later, countries are grouped as low- and middleincome developing countries. There are two model treaties for DTTs available, which are regularly updated and on which treaty partners can base their treaty if they wish to do so: one from the OECD, the other one from the United Nations. Not surprisingly, the OECD model treaty clearly favors residence taxation which benefits developed countries since it is mainly developed country investors who invest in developing countries, not the other way around, and residence taxation favors countries with net

11. A. Easson (2000), “Do we still need tax treaties?” Bulletin for International Fiscal Documentation, 54, p. 619. 12. Easson (2000). 13. V. Fitzgerald (2002), “International tax co-operation and capital mobility,” Oxford Development Studies, 30 (3), pp. 251–266; E. Neumayer and L. Spess (2005), “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33 (10), pp. 1567–1585. 14. UNCTAD (2003).

do double taxation treaties increase foreign direct investment 663

positive foreign asset positions. The UN model treaty, on the other hand, provides more room for source-based taxation, which is more beneficial to developing countries for the same reason. Critics argue, however, that the UN model treaty is not sufficiently different from the OECD model treaty and is still biased against developing country interests.15 Also, the vast majority of DTTs are based more on the OECD model.16 There are, of course, substantial differences in the way different developed countries tax their multinational companies abroad, a detailed discussion of which is beyond the scope of this chapter. For example, Collins and Shackelford compare and contrast the tax burden imposed by Canada, Japan, Germany, the United Kingdom, and the United States on their multinational companies.17 There are also differences among developed countries in the DTTs they typically conclude. For example, most developed countries include “tax sparing” arrangements, to be explained in the next section, with poor developing countries in their DTTs, whereas the United States does not.18 Vogel provides a detailed analysis of German and American DTTs and how they compare to the OECD and UN model conventions.19 Few would argue that double taxation represents the major impediment to foreign direct investment in developing countries. And yet, all other things being equal, the avoidance of double taxation can make a country more attractive to foreign investors who often have a choice among multiple locations. As Egger et al. point out, “One of the most visible obstacles to cross border investment is the double taxation of foreign-earned income.”20 Investors like stability and the legal and fiscal certainty that comes with a DTT can reassure foreign investors 15. A.H. Figueroa (1992), “Comprehensive tax treaties,” in Double Taxation Treaties between Industrialised and Developing Countries: OECD and UN Models – a Comparison. Proceedings of a Seminar held in Stockholm in 1990 during the 44th Congress of the International Fiscal Association, pp. 9–13. (Deventer: Kluwer Law and Taxation Publishers). 16. B. Arnold, J. Sasseville, and E.M Zolt (2002), “Summary of the proceedings of an invitational seminar on tax treaties in the 21st century,” Bulletin for International Fiscal Documentation, 56, pp. 233–243. 17. J.H. Collins and D.A. Shackelford (1995), “Corporate, domicile and average effective tax rates: the cases of Canada, Japan, the United Kingdom, and the United States,” International Tax and Public Finance, 2, pp. 55–84. 18. J.R. Hines Jr. (1998), “Tax sparing” and direct investment in developing countries,” Working Paper, No. 6728 (Cambridge, MA: NBER). Also published in J.R. Hines, Jr., ed. (2001). International Taxation and Multinational Activity (Chicago: University of Chicago Press), pp. 39–72. 19. K. Vogel (1997), Double Taxation Conventions – a Commentary to the OECD-, UNand U.S. Model Conventions for the Avoidance of Double Taxation on Income and Capital with Particular Reference to German Treaty Practice, 3rd edition (London: Kluwer Law International). 20. Egger et al. (2004), p. 1.

664 eric neumayer

that profits from their investments are not doubly reduced by taxation in both host and residence country.21 As is the case with BITs, the conclusion of DTTs also sends a certain signal to foreign investors that goes beyond the mere issue of taxation, allowing the developing country partner to acquire “international economic recognition,”22 or, in the words of H. David Rosenbloom, a “badge of international economic respectability.”23 Certainly, policy makers in developing countries must believe that the conclusion of DTTs increases inward FDI, as otherwise they would not flock to the negotiating table to sign more such treaties. This is because, as already mentioned, the vast majority of DTTs concluded between developed and developing countries limit source-based taxation, which means that developing countries can only collect tax revenues from foreign investors to a limited extent. For DTTs among developed countries this does not matter so much as FDI flows more or less equally in both directions. Economic relations between developed and developing countries are highly unequal, however, with the developed country being almost exclusively the country of residence and the developing country almost exclusively the host country. The reduction in tax revenue following limits on source-based taxation clearly represents a cost to developing countries. This is the more so as developing countries typically have very unequal income distributions that governments stripped of financial resources will find difficult to address via transfer payments.24 Dagan goes further, arguing that DTTs serve the “cynical goal” of “redistributing tax revenues from poorer to the richer signatory countries.”25 Nonetheless, the tax loss is somewhat mitigated if the finding of Chisik and Davies holds true beyond the seven developing countries included in their sample.26 They show that country-pairs with highly asymmetric FDI patterns, typical for developed-developing country pairs, tend to negotiate higher withholding taxes. More importantly, any loss of tax revenue can be justified if the wider economic benefits of attracting more FDI, such as knowledge and technology spillovers, higher economic growth, increased employment, and improved living standards exceed these costs.27 In order to materialize, however, more FDI actually needs to be attracted with the help of DTTs. Some, like Figueroa, argue that taxes do not enter foreign investors’ investment decisions, which would mean 21. United Nations (2001), United Nations Model Double Taxation Convention between Developed and Developing Countries (New York: United Nations). 22. T. Dagan (1999), “The tax treaties myth,” New York University Journal of International Law and Policy, 32, p. 32. 23. Rosenbloom (1982) quoted in P.D. Reese (1987), “United States tax treaty policy toward developing countries,” UCLA Law Review, 35, p. 380. 24. Fitzgerald (2002). 25. Dagan (1999), p. 939; similarly, Figueroa (1992). 26. R. Chisik and R.B. Davies (2004), “Asymmetric FDI and tax-treaty bargaining: theory and evidence,” Journal of Public Economics, 88, pp. 1119–1148. 27. Reese (1987).

do double taxation treaties increase foreign direct investment 665

that by implication DTTs are ineffective in raising FDI flows.28 But, both Gastanaga, Nugent and Pashamova, in their analysis of FDI flows to developing countries, and Desai, Foley and Hines, in their analysis of investment decisions by American-owned foreign affiliates abroad, find that taxes have a noticeable and statistically significantly negative impact on investment.29 If taxes are important to foreign investors, and DTTs reduce double taxation, can it be shown that DTTs have a positive effect on FDI? Existing studies, to be reviewed next, paint a somewhat pessimistic picture. B. Review of Studies on the Effect of DTTs on FDI Blonigen and Davies (2002), in an analysis of bilateral FDI outflows and outbound stocks from OECD countries to other countries over the period 1982 to 1992, find that the existence of DTTs is associated with larger bilateral FDI flows and stocks in ordinary least squares (OLS) estimation. But, when older DTTs, which have often been concluded many years before the start of the study period, are distinguished from newer DTTs, which were concluded during the period of study, it appears that these newer treaties have no positive effect on FDI inordinary least squares estimation. In fixed-effects estimation, based on the withinvariation of the data only so that old treaties concluded before the start of the sample become irrelevant, the effect is even negative. Similarly, Blonigen and Davies, in an analysis of U.S. inbound and outbound FDI over the period 1980 to 1999, find that treaties concluded by the U.S. during this period had no statistically significant effect at best, and a negative effect at worst, on inbound and outbound FDI stocks.30 Davies confirms the non-significant and negative findings of both studies and, additionally, finds non-significant results if looking explicitly at treaty renegotiations.31 Egger et al. also find a negative effect of newly implemented DTTs in a differences-in-differences analysis of two years prior to and two years after treaty conclusion using dyadic FDI data over the period 1985 to 2000.32 In contrast to the findings reported above, two other studies find evidence suggesting indirectly that DTTs might work. First, Hines looks at the effect of 28. Figuerora (1992). 29. V.M. Gastanaga, J.B. Nugent, and B. Pashamova (1998). “Host country reforms and FDI inflows: how much difference do they make?” World Development, 26 (7), pp. 1299–1314; M.A. Desai, C.F. Foley, and J.R. Hines Jr. (2002). “Chains of ownership, regional tax competition, and foreign direct investment,” NBER Working Paper, No. 9224. (Cambridge, MA: NBER). 30. Blonigen and Davies (2004); Louie and Roussland (2002) find a non-significant effect of DTTs on the rates of return that American companies require for their foreign investment in the years 1992, 1994 and 1996. 31. R.B. Davies (2004), “Tax treaties and foreign direct investment: potential versus performance,” International Tax and Public Finance, 11, pp. 775–802. 32. Egger et al. (2004).

666 eric neumayer

“tax sparing” agreements, rather than DTTs, on Japanese FDI location.33 “Tax sparing” occurs when capital exporting countries exempt from taxation any extra income its firms earn from tax reduction incentives in foreign countries. The DTTs of most developed countries, with the notable exception of the U.S., with many, but not all, developing countries contain such agreements. In comparing Japanese and U.S. investment patterns in 1990, Hines estimates that FDI in developing countries with whom a “tax sparing” agreement exists is 1.4 to 2.4 times higher than what it would have been otherwise.34 Further, Di Giovanni analyzes merger and acquisition (M&A) deals rather than FDI over the period 1990 to 1999.35 He comes to the interesting conclusion that the existence of a DTT is associated with higher cross-border merger and acquisition flows. The major problem with existing studies that directly address the effect of DTTs on FDI is twofold: First, the simultaneous presence of both OECD and developing countries in the sample, as in Blonigen and Davies and Egger et al., can be problematic as FDI allocation decisions are likely to be based on drastically different motivations in both groups of countries.36 Second, the use of dyadic FDI data, which is otherwise a strength, necessarily leads to a sample that is restrictive and non-representative for OECD countries other than the U.S. This would not matter as much if the sample of countries for which data are available was a random one. This is not the case, however, since bilateral FDI data exist for practically all OECD countries, but for developing countries by and large only if their per capita income is relatively high or their population size is large. This excludes the very set of poor to lower middle-income and small to medium-sized developing countries for which the conclusion of a DTT can be an important instrument to woo foreign investors. The only exception to this problem is the United States, for which quite comprehensive dyadic data for a wide range of developing countries are available. Our research design aspires to overcome both problems. First, we use a sample that contains only developing countries to account for the fact that FDI allocation decisions in this group of countries is likely to be driven by different motivations than FDI allocation within OECD countries.37 Second, for the one developed country for which comprehensive outbound FDI data exist, namely

33. Hines (1998). 34. Ibid. 35. J. Di Giovanni (2005), “What drives capital flows? The case of cross-border M&A activity and financial deepening,” Journal of International Economics, 65, pp. 127–149. 36. Blonigen and Davies (2002); Davies (2004); Egger et al. (2004); B.A. Blonigen and M. Wang (2004), “Inappropriate pooling of wealthy and poor countries in empirical FDI studies,” Working Paper, No. 10378 (Boston, MA: National Bureau of Economic Research). 37. Note that we treat Mexico, South Korea and Turkey as well as the Eastern European countries of Czech and Slovak Republic, Hungary and Poland as developing countries despite their recent membership in the OECD.

do double taxation treaties increase foreign direct investment 667

the United States, we create a more representative sample of developing countries that also covers a longer time period than is the case in Blonigen and Davies.38 The larger sample is partly due to extending the sample backwards, from 1980 to 1970, and forwards, from 1999 to 2001, and partly because our control variables seem to have less missing data than the ones used by Blonigen and Davies.39 Third, as already mentioned, for other developed countries no truly representative dyadic sample can be created due to lack of dyadic FDI data. In order to circumvent this problem, we use non-dyadic FDI data, which are available for a large sample. This poses the immediate problem that we can no longer directly infer whether the FDI is covered by a DTT or not. To deal with this problem indirectly, we will use a measure of weighted cumulative DTTs a developing country has signed with OECD countries, where each DTT is weighted by the share of outward FDI flow the OECD country accounts for relative to total world outward FDI flow.40 The weighting is to account for differences in the size of potential FDI share, for which a developing country has double taxation provisions in place. Clearly, in an ideal world it would be better to have comprehensive dyadic FDI data for OECD countries other than the U.S., so one could do without weighting. But, in the absence of such data, we believe that the benefit of deriving results from a much larger and more representative sample outweighs the cost of accounting for the potential FDI inflow covered by DTTs indirectly via the weighting procedure described above. C. Research Design 1. Dependent variable For the United States outward FDI model, we use data on outbound FDI stocks over the period 1970 to 2001 from the U.S. Bureau of Economic Analysis41 converted to constant U.S.$ of 1996 with the help of the U.S. GDP deflator.42 We estimate the model in stocks rather than flows as this is common usage in the existing literature on U.S. DTTs. We use absolute FDI stocks because if one were to use FDI stocks as a percentage of host country’s GDP instead, the measure would capture changes in the relative importance of 38. Blonigen and Davies (2002). 39. Ibid. 40. Ideally, one would want to weight not by the total FDI outflow from OECD countries, but by the FDI outflow that goes to developing countries only. However, this information is again not available for many OECD countries over a long time period. 41. See http://www.bea.doc.gov/bea/di1.htm. 42. Dropping resource depletion as a control variable from the model would even allow one to extend the sample back to 1966. Results are hardly affected by doing so. Note that since we use year dummies in the estimations, other than for the summary descriptive statistics it makes no difference to the results reported below whether the dependent variable is held in nominal or in real terms since the deflator is of course absorbed in the year dummies for the FDI in absolute amounts variables or is cancelled out in the FDI share variables.

668 eric neumayer

foreign investment to the host country, but not changes in stocks directly. For the non-dyadic FDI model, we use both total inbound FDI stocks of developing countries as well as FDI inflows to developing countries in constant U.S.$ of 1996.43 The reason for using both stocks and flows is that UNCTAD (2005) provides non-dyadic FDI stock data only from 1980 onwards, whereas non-dyadic flow data are available from 1970 onwards. Also, a large part of the literature that analyses non-dyadic FDI is estimated in FDI flows rather than stocks. Quite possibly, the worldwide increase in the rate of the conclusion of DTTs is partly responsible for the increase in overall FDI going to developing countries. Nevertheless, there is always the danger that one finds a statistically significant relationship between two trending variables that is spurious. We deal with this potential problem in two ways: First, for the regressions on absolute FDI stocks or flows, we employ year-specific time dummies to absorb any yearto-year variation in total FDI unaccounted for by our explanatory variables and common to all developing countries, which should mitigate potential spuriousness of any significant results. Second, to test the effect of DTTs on the attractiveness of a developing country relative to other developing countries even more directly, we include as an alternative dependent variable the share of FDI stock (flow) relative to total stock (flow) in all developing countries. Since the share variables are not trending over time (they are bound from above and below and their mean is almost constant over time), no year-specific time dummies are needed in these sets of estimations. We take the natural log of the dependent variables to reduce the skewness of its distribution. This increases the model fit substantially. 2. Explanatory variables Our main explanatory variable for the regressions with the U.S. FDI outbound stock in developing countries as dependent variable is a dummy variable for the existence of a DTT. For the estimations on nondyadic FDI this is replaced by the cumulative number of DTTs a developing country has signed with OECD countries weighted by the share of outward FDI flow the OECD country accounts for relative to total world outward FDI flow. Data are taken from UNCTAD (2005) with information on DTTs for several OECD countries provided directly by UNCTAD’s International Arrangements Section. As mentioned above, the weighting is to account for differences in the size of potential FDI share, for which a developing country has double taxation provisions in place. We exclude DTTs signed between developing countries since these represent a rather recent development and FDI stocks among developing countries are relatively small over the entire study period. We should note, however, that FDI flows between developing countries have increased substantially lately, possibly accounting for up to one-third of FDI flows to developing

43. Data from UNCTAD (2005), Foreign Direct Investment Statistics (New York and Geneva: United Nations), http://www.unctad.org/Templates/StartPage.asp?intItemID=2921&lang=1.

do double taxation treaties increase foreign direct investment 669

countries in the 1990s.44 In principle, it would be useful to distinguish among DTTs in more detail, however, this would not only require an enormous research effort given the existence of hundreds of DTTs between developed and developing countries, but it would also be next to impossible to complete the study in a quantitative way. Even the agreed-upon tax withholding rates, which are quantifiable, typically only specify maximum allowable rates, not effective rates, rendering them uninformative.45 Our control variables are identical to the ones used in Neumayer and Spess and are very similar to the ones used by Halward-Driemeier and Tobin and RoseAckerman, who analyse the impact of BITs on FDI to developing countries.46 They are also among the ones more consistently found to be determinants of FDI.47 First, we include a dummy variable for the presence of a BIT with the United States in the set of regressions with the U.S. outbound FDI stock as dependent variable. For the non-dyadic FDI regressions, we use the weighted cumulative number of BITs a developing country has signed with developed countries, applying the same weighting procedure as for DTTs. The investor protection provisions contained in BITs are meant to increase FDI to signatory developing countries, and Neumayer and Spess find evidence that they work.48 Further, we include the natural log of per capita income, the log of total population size, and the economic growth rate as indicators of market size and market potential.49 Developing countries, which have concluded a free trade agreement with a developed country, might receive more FDI as it is easier to export goods back into the developed or other countries. Such agreements sometimes also contain provisions on policies that might be beneficial to foreign investors. We account for this with two variables: One is a dummy variable indicating whether a country is a member of the World Trade Organization. The other is a variable counting the number of bilateral trade agreements a developing country has

44. UNCTAD (2004 - WIR04), World Investment Report 2004: The Shift Toward Services (New York and Geneva: United Nations), p. 20. 45. Blonigen and Davies (2002). 46. E. Neumayer and L. Spess (2005), “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33 (10), pp. 1567– 1585; M. Hallward-Driemeier (2003), “Do bilateral investment treaties attract FDI? Only a bit . . . and they could bite,” World Bank Policy Research Paper, WPS 3121 (Washington, D.C.: World Bank); J. Tobin and S. Rose-Ackerman (2005), “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Yale Law School Center for Law, Economics and Public Policy Research Paper, No. 293 (New Haven, CT: Yale Law School). 47. A. Chakrabarti (2001), “The determinants of foreign direct investment: sensitivity analyses of cross-country regressions,” Kyklos, 54, pp. 89–114. 48. Neumayer and Spess (2005). 49. Data from World Bank (2003a). World Development Indicators CD Rom (Washington, D.C.: World Bank).

670 eric neumayer

concluded with the U.S., the European Community/European Union, or Japan, weighted by their respective shares of world trade.50 Data on agreements are taken from WTO (2004) and EU (2004), and trade share data come from WTO (2006).51 Note that for the estimation results with U.S. outbound FDI stock as the dependent variable, only bilateral trade agreements with the U.S. enter this variable and no weighting is necessary. The inflation rate is a proxy variable for macroeconomic stability. Data are taken from World Bank (2003a). We employ a measure of natural resource intensity to control for the fact that, all other things being equal, abundant natural resources are a major attractor to foreign investors. Our measure is equal to the sum of rents from mineral resource and fossil fuel energy depletion divided by gross national income, as reported in World Bank (2003b).52 Rents are estimated as (P–AC) × R, that is, as price minus average cost multiplied by the amount of resource extracted, an amount known as total Hotelling rent in the natural resource economics literature.53 There is a long tradition of studies analyzing the effect of political stability and institutional quality on FDI54 We use the political constraints (POLCON) index developed by Henisz, mainly because in comparison to alternative measures of institutional quality it has far larger availability across time and countries.55 Henisz has designed his index as an indicator of the ability of political institutions to make credible commitments to an existing policy regime, which

50. We do not include the Lomé Conventions or the follow-on Cotonou Agreement between the EU and 77 countries from Africa, the Caribbean and the Pacific (ACP) since it is highly unlikely that these had a major impact on FDI. 51. WTO (2004), Regional trade agreements (Geneva: World Trade Organization), www. wto.org; European Union (2004), EC Regional Trade Agreements (Brussels: European Union Directorate General Trade), http://europa.eu.int/comm/trade/index_en.htm; WTO (2006), World trade statistics (Geneva: World Trade Organization), www.wto.org. 52. World Bank (2003b), Adjusted Net Savings Data (Washington, D.C.: World Bank), http://lnweb18.worldbank.org/ESSD/envext.nsf/44ByDocName/Green AccountingAdjustedNetSavings. 53. H. Hotelling (1931), “The economics of exhaustible resources,” Journal of Political Economy, 39 (2), pp. 137–75. 54. F. Schneider and B. Frey (1985), “Economic and political determinants of foreign direct investment,” World Development, 13, pp. 161–175; A. Alesina, and R. Perotti (1996), “Income distribution, political instability, and investment,” European Economic Review, 40, pp. 1203–1228; D. Wheeler and A. Mody (1992), “International investment location decisions: the case of U.S. firms,” Journal of International Economics, 33, pp. 57–76; S. Globerman and D. Shapiro (2002), “Global foreign direct investment flows: the role of governance infrastructure,” World Development, 30, pp. 1899–1919; H. Louie and D.J. Rousslang (2002), “Host-country governance, tax treaties and American direct investment abroad,” Working Paper (Washington, DC: Office of Tax Analysis). 55. W.J. Henisz (2000), “The institutional environment for economic growth,” Economics and Politics, 12 (1), pp. 1–31.

do double taxation treaties increase foreign direct investment 671

he argues is the most relevant political variable of interest to investors. Building on a simple spatial model of political interaction, the index makes use of the structure of government in a given country and of the political views represented by the different levels of government (i.e. the executive and the lower and upper legislative chambers). It measures the extent to which political actors are constrained in their choice of future policies by the existence of other political actors with veto power who will have to consent to new legislation. Using information on party composition of the executive and the legislative branches allows us to into account how alignment across branches of government and the extent of preference heterogeneity within each legislative branch impacts the feasibility of policy change. Scores range from zero, which indicates that the executive has total political discretion and could change existing policies at any point of time, to one, which indicates that a change of existing policies is totally infeasible. Of course, in practice agreement is always feasible, so the maximum score is less than one. 3. Estimation technique One could use a random-effects or fixed-effects estimator. We suspect that there are factors making a country attractive to foreign investors that are not captured by our explanatory variables and that are (approximately) time-invariant, such as colonial history, culture, language, climate, geographical distance to the centers of the Western developed world, legal restrictions on inward FDI, etc. These are also likely to be correlated with the explanatory variables, which would render random-effects estimation biased. We report Hausman test results below, which confirm this suspicion and make fixedeffects estimation the preferred specification. Both random- and fixed-effects estimation results are based on robust standard errors (using Huber/White estimators of variance). To mitigate potential reverse causality problems, we lag all explanatory variables by one year. Ideally, one would like to tackle this problem more comprehensively with the help of instrumental variable regression, but practically all explanatory variables are potentially subject to reverse causality and it would be simply impossible to find adequate and valid instruments. Table 1 provides summary descriptive variable information. Variance inflation analysis did not suggest reason for concern with multicollinearity problems. As in any regression analysis, there is, of course, the possibility of omitted variable bias. For example, we cannot account for over-time changes in domestic legislation or fiscal policies encouraging or discouraging FDI other than that which is captured by BITs and DTTs because there is no comprehensive information available. Nevertheless, we see no reason to think that our results would be significantly biased, should this or any other potentially omitted variable be systematically correlated with our explanatory variables. D. Results Table 2 presents estimation results for the logged outbound stock of FDI in U.S.$ of 1996 from the United States to developing countries. Column I is based

672 eric neumayer table 1. descriptive statistical variable information united states fdi sample: Variable

Obs

Mean

ln FDI stock ln FDI stock share DTT with U.S. BIT with U.S. ln GDP p.c. ln Population Econ. Growth Inflation Resource rents Trade agreement with U.S. WTO membership POLCON

2086 2086 2086 2086 2086 2086 2086 2086 2086 2086 2086 2086

5.15 –6.35 0.39 0.16 8.03 16.03 0.01 71.36 6.04 0.00 0.73 0.20

Median Std. Dev. 5.35 –6.03 0 0 8.16 16.03 0.02 9.51 1.2 0 1 0.14

2.47 2.50 0.49 0.37 0.81 1.82 0.06 768.16 10.09 0.06 0.44 0.20

Min

Max

–0.11 –12.43 0 0 5.64 10.62 –0.28 –31.52 0 0 0 0

10.83 –1.38 1 1 9.72 20.99 0.64 26762 66.60 1 1 0.67

non-dyadic sample: Variable

Obs

Mean

Median

Std. Dev. Min

Max

ln FDI stock ln FDI stock share ln FDI flow ln FDI flow share DTTs (weighted)

2084 2084 2767 2767 2767

6.64 –6.47 3.92 –7.06 29.92

6.63 –11.28 4.07 –6.89 17.72

2.08 2.02 2.52 2.45 32.16

–3.15 –16.50 –4.69 –16.98 0

12.81 –1.41 10.78 –1.17 99.28

BITs (weighted) ln GDP p.c. ln Population Econ. Growth Inflation Resource rents Bilateral trade agr. (weighted) WTO membership POLCON

2767 2767 2767 2767 2767 2767 2767 2767 2767

23.97 7.93 15.72 0.01 65.78 5.55 2.54 0.67 0.18

14.62 8.02 15.80 0.01 9.13 0.8 0 1 0.08

26.98 0.83 1.88 0.07 684.46 9.76 9.59 0.47 0.20

0 5.64 10.62 –0.42 –31.52 0 0 0 0

99.34 9.72 20.99 0.78 26762 66.60 52.05 1 0.67

on a random-effects specification. The existence of a double taxation treaty is associated with a higher FDI stock. The treaty effect is estimated to be around 34%. Richer, more populous, resource-abundant countries and those with good institutional quality and with a trade agreement with the U.S. also have a higher FDI stock. Somewhat surprisingly, fast-growing economies and WTO members

do double taxation treaties increase foreign direct investment 673

table 2. estimation results (united states fdi outbound stock in developing countries)

DTT BIT ln GDP p.c. ln Population Econ. growth Inflation Resource rents Bilateral trade agreement WTO membership POLCON Observations Countries Time dummies R-squared Hausman test

I (Random)

II (Fixed)

III (Fixed)

0.343 (5.21)∗∗∗ –0.007 (0.11) 1.507 (16.30)∗∗∗ 0.607 (7.46)∗∗∗ –0.810 (2.60)∗∗∗ –0.000 (0.06) 0.017 (3.89)∗∗∗ 0.760 (5.22)∗∗∗ –0.119 (1.81)∗ 0.250 (1.95)∗ 2086 114 yes 0.58 95.57 (0.0000)

0.224 (3.74)∗∗∗ –0.026 (0.43) 1.201 (12.12)∗∗∗ –2.157 (7.05)∗∗∗ –0.757 (2.53)∗∗ –0.000 (0.25) 0.016 (3.66)∗∗∗ 0.618 (4.72)∗∗∗ –0.148 (2.31)∗∗ 0.319 (2.59)∗∗∗ 2086 114 yes 0.37

0.201 (3.37)∗∗∗ 0.030 (0.52) 1.296 (15.29)∗∗∗ –1.835 (15.01)∗∗∗ –0.714 (2.42)∗∗ –0.000 (0.35) 0.013 (3.13)∗∗∗ 0.691 (5.31)∗∗∗ –0.088 (1.40) 0.417 (3.34)∗∗∗ 2086 114 no 0.25

Notes: Dependent variable is logged FDI stock in columns I and II and logged FDI stock share in column III. Absolute z- and t-values in parentheses. Hausman test is asymptotically χ2 distributed with p-values in brackets. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

have a lower stock, and neither the inflation rate nor the presence of a bilateral investment treaty matter. The random-effects estimation results are problematic on two accounts: First, the Hausman test clearly rejects the random-effects assumptions. Second, on a conceptual level, for the DTT variable random-effects estimation fails to tell us whether the conclusion of a double taxation treaty is associated with a higher

674 eric neumayer

FDI stock. Instead, what the results show is that the presence of a DTT is associated with a higher stock; however, the DTT might have been concluded before the period of analysis started and the variable might be correlated with countryspecific effects. It is therefore important to check whether the statistically significantly positive effect of the DTT variable holds up in fixed-effects estimation. Based on the within-variation of the data only, any effect has to derive from treaties concluded during the period of study in fixed-effects estimation. Column II estimates the same model with fixed effects, whereas column III replaces the dependent variable with the share of FDI stock, also with fixed effects.56 Results suggest that the effect of DTTs is not driven by their correlation with country specific fixed effects. Instead, the conclusion of a DTT during the sample period is associated with an FDI stock that is around 22% higher and an FDI stock share that is around 20% higher. Results on the control variables are mostly consistent across the set of estimations. Population size is an important exception. Keeping in mind that the fixed effects estimation is based on the within-variation of the data in each country only, this can be interpreted to the effect that countries with a larger population size have a higher FDI stock, as suggested by the random-effects estimation results. But, as a country’s population grows, its FDI stock and FDI stock share become smaller rather than bigger conditional on the other explanatory variables and the country-specific fixed effects. In Table 3, we turn to the analysis of non-dyadic FDI data. Random-effects estimation results on FDI stocks in column I suggest that countries with a higher cumulative number of DTTs and BITs, and richer countries with larger populations, have a higher stock of FDI. So do countries that are more intensive in natural resource extraction, that exhibit greater institutional quality as measured by POLCON, that are WTO members, and that have a higher number of trade agreements with developed countries. The economic growth and inflation rate are statistically insignificant. The Hausman test again rejects the random effects assumption. The same is true for all the alternative dependent variables, which is why we concentrate on fixed-effects estimation from now on. The fixed-effects estimation results on FDI stocks are rather similar to the random-effects results (column II). In particular, higher cumulative numbers of DTTs and BITs remain positively associated with higher FDI stocks. The population variable switches signs again, whereas a higher inflation rate now has a negative effect. In column III, we replace the dependent variable with the share of FDI stock, but results remain largely consistent. As mentioned already, FDI flows are available for a longer time period than FDI stocks, and the literature on non-dyadic FDI is often estimated in flows. In columns IV and V, the log of FDI stock and FDI stock share is therefore replaced

56. Non-reported Hausman tests again clearly rejected the random-effects assumption for this alternative dependent variable.

do double taxation treaties increase foreign direct investment 675

table 3. estimation results (fdi non-dyadic stocks and inflows in developing countries)

DTTs (weighted) BITs (weighted) ln GDP p.c. ln Population Econ. growth Inflation Resource rents Bilateral trade agreements (weighted) WTO membership POLCON Observations Countries Time dummies R-squared Hausman test

I (Random)

II (Fixed)

III (Fixed)

IV (Fixed)

V (Fixed)

0.004 (3.34)∗∗∗ 0.005 (4.62)∗∗∗ 1.072 (12.55)∗∗∗ 0.577 (11.57)∗∗∗ 0.151 (0.53) –0.000 (1.47) 0.024 (6.06)∗∗∗ 0.009 (2.08)∗∗

0.002 (1.91)∗ 0.004 (3.78)∗∗∗ 0.814 (7.29)∗∗∗ –1.407 (3.53)∗∗∗ 0.190 (0.68) –0.000 (2.20)∗∗ 0.022 (4.53)∗∗∗ 0.011 (2.57)∗∗

0.003 (2.59)∗∗∗ 0.004 (3.34)∗∗∗ 0.820 (8.42)∗∗∗ –0.990 (5.06)∗∗∗ 0.157 (0.58) –0.000 (2.23)∗∗ 0.018 (3.92)∗∗∗ 0.010 (2.12)∗∗

0.009 (3.48)∗∗∗ 0.016 (6.14)∗∗∗ 0.462 (2.29)∗∗ –1.554 (3.13)∗∗∗ 1.128 (1.80)∗ –0.000 (3.18)∗∗∗ 0.030 (3.71)∗∗∗ 0.014 (1.85)∗

0.009 (3.54)∗∗∗ 0.012 (5.42)∗∗∗ 0.288 (1.55) –2.786 (12.66)∗∗∗ 1.337 (2.16)∗∗ –0.000 (3.28)∗∗∗ 0.027 (3.48)∗∗∗ 0.015 (1.99)∗∗

0.268 (5.15)∗∗∗ 0.201 (2.12)∗∗ 2115 120 yes 0.68 79.48 (0.0000)

0.182 (3.40)∗∗∗ 0.157 (1.60) 2115 120 yes 0.50 –

0.170 (3.14)∗∗∗ 0.152 (1.54) 2145 120 no 0.14 –

0.169 (1.54) 0.534 (2.23)∗∗ 2767 120 yes 0.22 –

0.103 (0.98) 0.426 (1.82)∗ 2767 120 no 0.09 –

Notes: Dependent variable is logged FDI stock in columns I and II, logged FDI stock share in column III, logged FDI flows in column IV, and logged FDI flow share in column V. Absolute z- and t-values in parentheses. Hausman test is asymptotically χ2 distributed with p-values in brackets. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

by the log of FDI inflows and FDI inflow share. Results are remarkably consistent with the previous ones for FDI stocks. In particular, higher cumulative numbers of DTTs and BITs remain positively associated with higher FDI flows. The main difference is that faster growing economies attract higher flows, whereas the WTO membership dummy variable loses its statistical significance.

676 eric neumayer table 4. fixed-effects estimation results for separate low-income and middle-income developing country samples (united states fdi outbound stock)

DTT BIT ln GDP p.c. ln Population Econ. growth Inflation Resource rents Bilateral trade agreement WTO membership POLCON Income group Observations Countries Time dummies R-squared

I

II

III

IV

–0.184 (1.21) 0.432 (3.43)∗∗∗ 1.574 (8.86)∗∗∗ –1.879 (2.23)∗∗ –1.816 (4.10)∗∗∗ –0.000 (0.36) 0.014 (1.82)∗

0.276 (4.12)∗∗∗ –0.199 (2.98)∗∗∗ 1.086 (8.99)∗∗∗ –1.826 (4.89)∗∗∗ 0.031 (0.08)

–0.201 (1.39) 0.364 (3.04)∗∗∗ 1.468 (9.91)∗∗∗ –2.584 (15.44)∗∗∗ –1.492 (3.45)∗∗∗ –0.000 (0.23) 0.015 (1.86)∗

0.154 (2.22)∗∗ –0.122 (1.87)∗ 1.099 (10.27)∗∗∗ –2.581 (13.53)∗∗∗ 0.119 (0.31)

–0.511 (2.53)∗∗ –0.195 (0.74) Low 785 46 yes 0.33

–0.000 (0.13) 0.017 (3.15)∗∗∗ 0.495 (4.00)∗∗∗ –0.070 (1.03) 0.474 (3.48)∗∗∗ Middle 1301 68 yes 0.44

–0.433 (2.07)∗∗ 0.266 (1.03) Low 785 46 no 0.46

–0.000 (0.80) 0.010 (2.07)∗∗ 0.645 (6.47)∗∗∗ 0.085 (1.27) 0.608 (4.15)∗∗∗ Middle 1301 68 no 0.24

Notes: Dependent variable is logged FDI stock in columns I and II and logged FDI stock share in columns III and IV. Absolute t-values in parentheses. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level.

In tables 4 and 5 we explore whether the effect of DTTs holds up for two subsets of developing countries, namely low-income versus middle-income countries, using country classification according to World Bank criteria (see appendix 1). The tables report the set of fixed-effects estimation results for both sub-samples of countries. From table 4, it is clear that the positive effect of DTTs on FDI is

(fdi non-dyadic stocks and inflows)

DTTs (weighted) BITs (weighted) ln GDP p.c. ln Population Econ. growth Inflation Resource rents

I

II

III

IV

V

VI

VII

VIII

–0.000 (0.03) 0.003 (1.51) 1.152 (6.36)∗∗∗ –1.491 (2.29)∗∗ 0.008 (0.02) –0.000 (1.57) 0.007 (1.46)

0.005 (3.47)∗∗∗ 0.005 (3.33)∗∗∗ 0.653 (3.77)∗∗∗ –1.891 (3.38)∗∗∗ 0.079 (0.20) –0.000 (1.80)∗ 0.030 (4.07)∗∗∗

–0.000 (0.12) 0.006 (3.11)∗∗∗ 1.311 (7.61)∗∗∗ –0.602 (3.00)∗∗∗ 0.096 (0.23) –0.000 (1.17) 0.004 (0.84)

0.006 (4.07)∗∗∗ 0.006 (3.95)∗∗∗ 1.073 (7.02)∗∗∗ –1.085 (2.45)∗∗ 0.055 (0.14) –0.000 (1.36) 0.025 (3.17)∗∗∗

0.008 (1.58) 0.006 (1.22) 1.281 (3.86)∗∗∗ 1.557 (1.13) 0.417 (0.38) –0.000 (3.09)∗∗∗ –0.004 (0.37)

0.008 (2.89)∗∗∗ 0.020 (5.97)∗∗∗ –0.158 (0.57) –1.671 (2.63)∗∗∗ 1.603 (2.35)∗∗ –0.000 (1.02) 0.057 (5.15)∗∗∗

0.009 (1.61) 0.000 (0.07) 0.972 (2.79)∗∗∗ –2.653 (8.61)∗∗∗ 0.740 (0.69) –0.000 (3.24)∗∗∗ –0.004 (0.34)

0.008 (2.97)∗∗∗ 0.017 (6.11)∗∗∗ –0.310 (1.36) –2.146 (5.04)∗∗∗ 1.860 (2.76)∗∗∗ –0.000 (1.04) 0.052 (4.98)∗∗∗

Continued

do double taxation treaties increase foreign direct investment 677

table 5. fixed-effects estimation results for separate low- and middle-income developing country samples

(fdi non-dyadic stocks and inflows) (cont’d...)

Bilateral trade agr. (weighted) WTO membership POLCON Income group Observations Countries Time dummies R-squared

I

II

III

IV

V

VI

VII

VIII

– – 1.028 (8.52)∗∗∗ 0.113 (0.71) Low 911 50 yes 0.53

0.013 (2.94)∗∗∗ -0.087 (1.34) –0.079 (0.63) Middle 1204 70 yes 0.52

– – 1.042 (8.58)∗∗∗ 0.162 (1.14) Low 919 50 no 0.23

0.013 (3.21)∗∗∗ –0.030 (0.47) 0.042 (0.30) Middle 1226 70 no 0.15

– – 0.735 (3.15)∗∗∗ 0.460 (1.06) Low 1179 50 yes 0.14

0.009 (1.28) –0.113 (0.90) 0.334 (1.19) Middle 1588 70 yes 0.31

– – 0.663 (3.02)∗∗∗ 0.369 (0.91) Low 1179 50 no 0.12

0.010 (1.40) –0.142 (1.18) 0.336 (1.21) Middle 1588 70 no 0.10

Notes: Dependent variable is logged FDI stock in columns I and II, logged FDI stock share in columns III and IV, logged FDI flows in columns V and VI, and logged FDI flow share in columns VII and VIII. Absolute t-values in parentheses. ∗ significant at .1 level ∗∗ at .05 level ∗∗∗ at .01 level

678 eric neumayer

table 5. fixed-effects estimation results for separate low- and middle-income developing country samples

do double taxation treaties increase foreign direct investment 679

exclusive to the group of middle-income countries.57 Similarly, from table 5, it is clear that the positive effect of the cumulative number of DTTs on non-dyadic FDI stocks and inflows is also exclusive to the group of middle-income developing countries, even though the variable is close to marginal statistical significance for low-income countries in the FDI flow models.58 One must keep in mind, of course, that with few exceptions, such as China and India, countries in the low-income group have concluded few, if any, DTTs with developed countries, whereas countries in the group of middle-income countries have many more DTTs in existence. The average value of the weighted treaty variable in the year 2001 in the group of middle-income countries is almost double the average of the low-income group (38.3 versus 20).

conclusion Developing countries that sign a DTT with the United States benefit from a higher FDI stock, and share of FDI stock, originating from U.S. investors. Unfortunately, due to lack of data, no representative sample of dyadic FDI data for other capital-exporting developed countries can be constructed. Nonetheless, our estimation results with non-dyadic FDI and a cumulative number of DTT treaty variables weighted by the relative importance of the developed country treaty partner as a capital exporter suggest that the effect is a general one. Developing countries with more DTTs with major capital-exporting developed countries benefit from a higher overall FDI stock and share of stock and receive more FDI inflows as well as a higher share of inflows. The message to developing countries, therefore, is that succumbing to the restrictions on their authority to tax corporate income from foreign investors typically contained in DTTs does have the desired payoff in terms of higher FDI. To our knowledge, ours is the first study to provide robust empirical evidence that DTTs increase FDI to developing countries; however, once we split the sample of developing countries into low-income and middle-income countries, we found that DTTs are only effective in the group of middle-income countries. Future research should explore this finding in greater detail. Statistical significance is not equivalent to substantive importance. We therefore need to know the strength of the effect of the DTT variables on FDI. How much more FDI can a developing country expect if it signs a DTT with the U.S. or if it aggressively engages in a program to sign DTTs with developed countries? For the U.S. FDI model, the estimated coefficient would suggest that concluding 57. None of the low-income countries has a bilateral trade agreement with the U.S. such that the variable is dropped from the estimations in the relevant regressions. 58. Again, none of the low-income countries has a bilateral trade agreement with the U.S., the European Union, or Japan.

680 eric neumayer

a DTT with the U.S. is predicted to increase the FDI stock from U.S. investors by 22% and the FDI stock share by 20%. These are large, but not implausibly high, effects. For non-dyadic FDI, we look at a one standard deviation increase in the DTT variable. This is equivalent to an increase of around 32.2 in the weighted cumulative DTT variable, which runs from 0 to 99.3 and is close to the sample mean of 29.9, but almost twice the sample median of 17.7. In the year 2002, a standard deviation increase would be approximately equivalent to signing a DTT with the United States and a country like Canada or Spain, or, alternatively, it would be equivalent to signing a DTT with Italy, France, Netherlands, and Japan. Based on the estimations in table 3, a country experiencing a one standard deviation increase in the DTT variable, is predicted to increase its FDI stock by about 6%, its share of FDI stock by about 9% and its FDI inflow, as well as FDI share relative to the total inflow to developing countries, by about 29%. Clearly, these are non-negligible increases following a substantial increase in DTT activity. But whether the demonstrated benefits of signing up to DTTs in the form of increased FDI are higher than the substantial costs developing countries incur in negotiating, signing, and concluding DTTs together with the loss in tax revenues is impossible to tell. What we do know is that DTTs fulfil the purpose of attracting FDI, and that those developing countries that have signed more DTTs with major capital exporting developed countries are likely to have received more FDI in return.

do double taxation treaties increase foreign direct investment 681

appendix 1 List of Countries Included in Sample with Number of DTTs with Oecd Countries in 2002 in Brackets Low-income countries (Gross National Income in 2001 less than or equal to $745): Angola (0), Armenia (4), Azerbaijan (3), Bangladesh (10), Benin (2), Burkina Faso (1), Burundi (0), Cambodia (3), Cameroon (0), Central African Republic (1), Chad (0), China (21), Comoros (1), Congo (Dem. Rep.) (0), Congo (Rep.) (2), Côte d’Ivoire (8), Ethiopia (1), Gambia (4), Ghana (2), Guinea (0), GuineaBissau (0), Haiti (0), Honduras (1), India (20), Indonesia (17), Kenya (7), Kyrgyz Republic (2), Lesotho (1), Madagascar (1), Malawi (6), Mali (1), Mauritania (1), Moldova (1), Mozambique (1), Nepal (1), Nicaragua (0), Niger (1), Nigeria (6), Pakistan (16), Rwanda (0), São Tomé and Principe (0), Senegal (4), Sierra Leone (2), Tanzania (5), Togo (1), Uganda (4), Vietnam (12), Yemen (0), Zambia (12), Zimbabwe (8). Middle-income countries (Gross National Income in 2001 more than $745): Albania (5), Algeria (6), Antigua and Barbuda (4), Argentina (16), Barbados (8), Belarus (4), Belize (4), Bolivia (6), Botswana (2), Brazil (15), Bulgaria (17), Cape Verde (1), Chile (3), Colombia (1), Costa Rica (2), Croatia (6), Czech Republic (20), Dominica (5), Dominican Republic (1), Ecuador (5), Egypt (14), El Salvador (1), Equatorial Guinea (0), Estonia (15), Fiji (5), Gabon (1), Georgia (3), Grenada (3), Guatemala (0), Guyana (3), Hungary (19), Iran (4), Jamaica (7), Jordan (4), Kazakhstan (10), Korea (Rep.) (20), Latvia (16), Lebanon (4), Lithuania (15), Macedonia FYR (6), Malaysia (17), Mauritius (6), Mexico (17), Morocco (15), Namibia (4), Panama (2), Papua New Guinea (3), Paraguay (0), Peru (2), Philippines (17), Poland (20), Romania (19), Russian Federation (18), Seychelles (3), Slovak Republic (20), South Africa (17), Sri Lanka (14), St. Kitts and Nevis (4), St. Lucia (3), St. Vincent and the Grenadines (4), Swaziland (2), Syria (1), Thailand (17), Trinidad and Tobago (9), Tunisia (15), Turkey (12), Ukraine (10), Uruguay (2), Uzbekistan (8), Venezuela (11).

682 eric neumayer

references Alesina, A. and R. Perotti (1996). “Income distribution, political instability, and investment,” European Economic Review, 40, pp. 1203–1228. Arnold, B. (2004). “Tax treaties and tax avoidance: the 2003 revisions to the commentary to the OECD Model,” Bulletin for International Fiscal Documentation, 58, pp. 244–260. Arnold, B., J. Sasseville and E.M Zolt (2002). “Summary of the proceedings of an invitational seminar on tax treaties in the 21st century,” Bulletin for International Fiscal Documentation, 56, pp. 233–243. Belsley, D.A., E. Kuh, and R.E.Welsch (1980). Regression Diagnostics (New York: John Wiley). Blonigen, B.A. and M. Wang (2004). “Inappropriate pooling of wealthy and poor countries in empirical FDI studies,” Working Paper, No. 10378 (Boston, MA: National Bureau of Economic Research). Blonigen, B.A. and R.B. Davies (2002) “Do bilateral tax treaties promote foreign direct investment?” Working Paper, No. 8834 (Boston, MA: National Bureau of Economic Research). Also published in E.K. Choi and J.C. Hartigan, eds. (2005), Handbook of International Trade Volume II (London: Blackwell), pp. 526–546. —— (2004). “The effects of bilateral tax treaties on U.S. FDI activity,” International Tax and Public Finance, 11, pp. 601–622. Chakrabarti, A. (2001). “The determinants of foreign direct investment: sensitivity analyses of cross-country regressions,” Kyklos, 54, pp. 89–114. Chisik, R. and R.B. Davies (2004). “Asymmetric FDI and tax-treaty bargaining: theory and evidence,” Journal of Public Economics, 88, pp. 1119–1148. Collins, J.H. and D.A. Shackelford (1995). “Corporate, domicile and average effective tax rates: the cases of Canada, Japan, the United Kingdom, and the United States,” International Tax and Public Finance, 2, pp. 55–84. —— (2002). “Do U.S. multinationals face different tax burden than other companies?” Working Paper (Chapel Hill, NC: University of North Carolina). Dagan, T. (1999). “The tax treaties myth,” New York University Journal of International Law and Policy, 32, pp. 939–996. Davies, R.B. (2004). “Tax treaties and foreign direct investment: potential versus performance,” International Tax and Public Finance, 11, pp. 775–802. Desai, M.A., C.F. Foley, and J.R. Hines Jr. (2002). “Chains of ownership, regional tax competition, and foreign direct investment,” NBER Working Paper, No. 9224. (Cambridge, MA: NBER). De Soysa, I., and J.R. Oneal (1999). “Boon or bane? Reassessing the productivity of foreign direct investment,” American Sociological Review, 64, pp. 766–782. Di Giovanni, J. (2005). “What drives capital flows? The case of cross-border M&A activity and financial deepening,” Journal of International Economics, 65, pp. 127–149. Dunning, J.H. (1998). “The changing geography of foreign direct investment,” in N. Kumar, ed., Globalization, Foreign Direct Investment and Technology Transfers: Impacts on and Prospects for Developing Countries (New York: Routledge). Easson, A. (2000). “Do we still need tax treaties?” Bulletin for International Fiscal Documentation, 54, pp. 619–625. Egger, P., M. Larch, M. Pfaffermayer, and H. Winner (2004). “The impact of endogenous tax treaties on foreign direct investment: theory and empirical evidence,” Working Paper (University of Innsbruck, Department of Economics and Statistics).

do double taxation treaties increase foreign direct investment 683

European Union (2004). EC Regional Trade Agreements (Brussels: European Union Directorate General Trade), http://europa.eu.int/comm/trade/index_en.htm. Figueroa, A.H. (1992). “Comprehensive tax treaties,” in Double Taxation Treaties between Industrialised and Developing Countries: OECD and UN Models – a Comparison. Proceedings of a Seminar held in Stockholm in 1990 during the 44th Congress of the International Fiscal Association, pp. 9–13. (Deventer, The Netherlands: Kluwer Law and Taxation Publishers). Fitzgerald, V. (2002). “International tax co-operation and capital mobility,” Oxford Development Studies, 30 (3), pp. 251–266. Gastanaga, V.M., J.B. Nugent, and B. Pashamova (1998). “Host country reforms and FDI inflows: how much difference do they make?” World Development, 26 (7), pp. 1299–1314. Globerman, S. and D. Shapiro (2002), “Global foreign direct investment flows: the role of governance infrastructure,” World Development, 30, pp. 1899–1919. Hallward-Driemeier, M. (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and they could bite,” World Bank Policy Research Paper, WPS 3121 (Washington, D.C.: World Bank). Henisz, W. J. (2000). “The institutional environment for economic growth,” Economics and Politics, 12 (1), pp. 1–31. Hermes, N. and R. Lensink (2003). “Foreign direct investment, financial development and economic growth,” Journal of Development Studies, 40 (1), pp. 142–163. Hines, J.R. Jr. (1998). “Tax sparing” and direct investment in developing countries,” Working Paper, No. 6728 (Cambridge, MA: NBER). Also published in J.R. Hines, Jr., ed. (2001). International Taxation and Multinational Activity (Chicago: University of Chicago Press), pp. 39–72. Hotelling, H. (1931). “The economics of exhaustible resources,” Journal of Political Economy, 39 (2), pp. 137–75. Louie, H. and D.J. Rousslang (2002). “Host-country governance, tax treaties and American direct investment abroad,” Working Paper (Washington, DC: Office of Tax Analysis). Neumayer, E. and L. Spess (2005). “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33 (10), pp. 1567–1585. Noorbakhsh, F., A. Paloni, and A. Youssef (2001). “Human capital and FDI inflows to developing countries: new empirical evidence,” World Development, 29, pp. 1593–1610. Reese, P.D. (1987). “United States tax treaty policy toward developing countries,” UCLA Law Review, 35, pp. 369–397. Schneider, F. and B. Frey (1985). “Economic and political determinants of foreign direct investment,” World Development, 13, pp. 161–175. Taylor, C.J. (2000). “The impact of host country government policy on U.S. multinational investment decisions,” The World Economy, 23, pp. 635–647. Tobin, J. and S. Rose-Ackerman (2005). “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Yale Law School Center for Law, Economics and Public Policy Research Paper, No. 293 (New Haven, CT: Yale Law School). UNCTAD (1998). Bilateral Investment Treaties in the Mid-1990s (New York and Geneva: United Nations).

684 eric neumayer —— (2001). Bilateral Investment Treaties 1959–1999 (New York and Geneva: United Nations). —— (2003 - WIR03). World Investment Report 2003: FDI Policies for Development National and International Perspectives (New York and Geneva: United Nations). —— (2004 - WIR04). World Investment Report 2004: The Shift Toward Services (New York and Geneva: United Nations). —— (2005). Foreign Direct Investment Statistics (New York and Geneva: United Nations), http://www.unctad.org/Templates/StartPage.asp?intItemID=2921&lang=1 United Nations (2001). United Nations Model Double Taxation Convention between Developed and Developing Countries (New York: United Nations). Vogel, K. (1997). Double Taxation Conventions – a Commentary to the OECD-, UN- and U.S. Model Conventions for the Avoidance of Double Taxation on Income and Capital with Particular Reference to German Treaty Practice, 3rd edition (London: Kluwer Law International). Wheeler, D. and A. Mody (1992). “International investment location decisions: the case of U.S. firms,” Journal of International Economics, 33, pp. 57–76. World Bank (2003a). World Development Indicators CD Rom (Washington, D.C.: World Bank). —— (2003b). Adjusted Net Savings Data (Washington, D.C.: World Bank), ™http://lnweb18.worldbank.org/ESSD/envext.nsf/44ByDocName/GreenAccounting AdjustedNetSavings. WTO (2004). Regional trade agreements (Geneva: World Trade Organization), www.wto.org. —— (2006). World trade statistics (Geneva: World Trade Organization), www.wto.org. Zebregs, H. (1998). “Can the neoclassical model explain the distribution of foreign direct investment across developing countries?” IMF Working Paper, WP/98/139 (Washington, D,C,: International Monetary Fund). Zhang, X. and K.H. Zhang (2003). “How does globalisation affect regional inequality within a developing country? Evidence from China,” Journal of Development Studies, 39 (4), pp. 47–67.

part four exploring the impact of tax and foreign investment treaties on foreign direct investment flows

This page intentionally left blank

24. the effect of tax and investment treaties on bilateral fdi flows to transition economies∗ tom coupé, irina orlova and alexandre skiba introduction According to the Monterrey Consensus on development financing (2002) rich countries are committed to undertake measures to increase foreign direct investment and other private flows to low-income and transition economies. The idea behind this political commitment is that all sorts of resources must be mobilized for development. Specifically, the Consensus states that a central challenge is to create the necessary domestic and international conditions to facilitate direct investment flows. The question, then, is how foreign direct investment (FDI) flows can be promoted by rich countries, and, further, which policies or measures can enhance FDI outflows to a given poor country. Among the priority areas the Monterrey Consensus mentions are the development of a regulatory framework for promoting and protecting investments, including the areas of human resource development; the avoidance of double taxation; and the signing of investment agreements. Also, the World Investment Report, the United Nations Conference on Trade and Development (UNCTAD) (2003) focuses on policies that can help to stimulate FDI, focusing on such instruments as the adoption of bilateral investment treaties (BITs) and bilateral treaties for the avoidance of double taxation (DTTs). The Monterrey Consensus was adopted in 2002, when there was a continuing downturn in global investment. Already in 2004 the rise in flows to developing countries, South-East Europe (SEE), and Commonwealth of Independent States (CIS) not only put an end to the downturn that had begun in 2001, but it also represented the highest ever level of investment flows to these countries. The question is whether this rise was attributable to the increasing efforts of the international community to enable more foreign investment. Even though this particular historical episode escapes our empirical analysis, with this research we try to shed light on the relation between specific policies and FDI.

∗ This chapter was reprinted with permission from the authors. The chapter was originally published as “The Impact of Policies on FDI flows to Transition Countries,” in Bob Lucas, T.N. Srinivasan and Lyn Squire, The Impact of Rich Countries’ Policies on Poverty in Poor Countries (Edward Elgar, forthcoming 2008). The authors acknowledge financial support from the Global Development Network.

688 tom coupé, irina orlova and alexandre skiba

In order to contribute to the ongoing debate, we focus our study on the effect of international treaties, more specifically bilateral investment treaties and bilateral tax treaties, on FDI. Bilateral investment treaties are agreements between two countries for the reciprocal encouragement, promotion, and protection of investments in each other’s territories by companies based in either country. The general purpose of double taxation treaties is to provide a favorable investment regime by the exclusion of double taxation of an investor. According to UNCTAD (2000), double taxation can be defined as the levy of taxes on income/capital in the hands of the same tax payer in more than one country in respect to the same income or capital for the same period. DTTs arise to avoid such a hardship. The existing studies analyzing the effect of treaties on FDI focus either on bilateral investment treaties or on bilateral tax treaties, but not on both types of treaties simultaneously.1 The studies on the effect of BITs provide mixed results. Using a dyadic approach, Hallward-Driemeier (2003) finds little evidence for the effect of BITs. Using the same approach, Salacuse and Sullivan (2005), however, find a positive effect for BITs signed by the United States, but no effect for BITs signed by other Organization for Economic Cooperation and Development (OECD) countries. Neumayer and Spess (2005) use a non-dyadic approach and provide evidence that a greater number of BITs signed by a host country increases the FDI a developing country receives. Tobin & Rose-Ackerman (2005) find a positive effect when using a non-dyadic approach, but no effect when using a dyadic approach. There are also very few studies that analyze the effect of tax treaties on FDI. The fundamental work on the issue belongs to Blonigen and Davies (2000, 2002, 2004). The basic point of their studies is that there is no empirical evidence of the standard view that treaties increase foreign direct investment. In this study, we empirically examine the effect of both treaties, BITs and DTTs, on foreign direct investment simultaneously. Since previous studies consider either bilateral investment treaties or tax treaties, but do not include both in one regression, their results suffer from omitted variable bias. This omitted variable bias can be substantial since the correlation between the two treaties is quite high; in our sample it is 0.27.2 We further separate tax treaties into three categories: income and capital tax treaties (ICT), income tax treaties (IT), and social security treaties (SS). This division is based on the categorization by the International Bureau of Fiscal Documentation. Using several dummies, rather than the single catch-all dummy

1. An exception is the unpublished MA thesis of Goryunov (2004). His results are problematic however because he does not include year dummies in his regressions, thus his results confound year effects and treaty effects. 2. This correlation is computed between a BIT dummy and a tax treaty dummy which is one if there is any tax treaty and zero otherwise (see more about this below).

the effect of tax and investment treaties on bilateral fdi flows 689

used in previous studies, is important since we thereby use more information— the extent to which a specific treaty influences FDI might depend on its specific category—and can get more precise estimates. Some studies also distinguish between tax treaties, but based on a time criterion, as Blonigen and Davies (2004) include two treaty dummies: one for new treaty and the other for old treaty, differentiating them by the date of signature. Finally, not all studies consider the potential endogeneity issue—FDI might cause treaties to be signed or not signed rather than vice versa. We provide an extensive treatment of this issue by experimenting with several possible instrumental variables. Our study provides strong evidence for a positive BITs effect: throughout numerous specifications, we get a positive significant estimate of BITs. On the other hand, our tax treaty dummies do not show such consistency: different specifications give different results. In our preferred specifications most tax treaty effects are insignificant. This chapter is structured as follows: the first section describes bilateral investment and tax treaties, and reviews existing empirical studies on both BITs and DTTs. Next, we describe the methodology and data, followed by a discussion of the estimation technique and results, and our conclusion. A. BITs and DTTs 1. Bilateral investment treaties (BITs) The general purpose of BITs is the promotion and protection of investments from one country to another country. While all BITs are very similar in their major provisions, some variations exist between the two basic models of BITs that have emerged so far. The “European model” appeared first; it was endorsed by the OECD Ministers in 1962. Then in the early 1980s the “North American model” was developed. Both models cover the following areas: admission and treatment, transfers, key personnel, expropriation, and dispute settlement. The two models differ in several aspects. In the “European model” treatment provisions apply only to an investment after establishment, while in the “North American model” treatment provisions also apply to the investment at the pre-establishment phase (this refers to the entry of investments and investors of a contracting party into the territory of another contracting party). In addition, the U.S. model has provisions that are more elaborate on some matters (e.g., right of entry) than the European BITs. As to protection of established investments, both models contain more or less the same concepts. These include national treatment (terms no less favorable than those that apply to domestic investors) and most-favored-nation treatment (terms no less favorable than those that apply to investors from third countries), free transfers of funds, adequate and effective compensation in the case of expropriation, full protection and security of investments, and dispute settlement mechanisms. Most BITs have been signed between developed capital-exporting countries and developing capital-importing countries. The majority of BITs were signed

690 tom coupé, irina orlova and alexandre skiba

during the last two decades, and UNCTAD has been actively monitoring and analyzing the increase in the number of BITs. The total number of treaties quintupled, rising from 385 at the end of the 1980s to 1,857 at the end of the 1990s. During the same period, the number of treaties concluded by developing countries and by Central and Eastern European (CEE) countries increased from sixtythree to 833. According to UNCTAD statistics, the total number of BITs reached 2,265 in 2003 (see Appendix A). They now involve 177 economies. The recent proliferation of bilateral investment treaties suggests that they are playing an increasingly important role in international investment relations. 2. Double taxation treaties (DTTs) Tax treaties exist between countries on a bilateral basis to prevent double taxation (taxes levied twice on the same income, profit, capital gain, inheritance, or other item). There are a number of model tax treaties published by various national and international bodies, such as the UN and the OECD. The OECD Model Convention on Income and on Capital serves as a model used by countries when negotiating bilateral tax agreements. The Convention is dynamic in that it is constantly monitored and updated as economies evolve and new tax questions arise. In recent years, for example, special reports had to be made on tax treatment of software and treaty characterization issues arising from e-commerce. According to the International Bureau of Fiscal Documentation, 2,390 double taxation treaties had been signed by the end of 2005. These treaties include income treaties, income and capital treaties, and social security treaties. Over two thirds of DTT’s (1,632 treaties) have been signed since 1980 with a notable surge in the activity in the 1990s when 731 treaties were signed. A sizeable portion of all treaties is signed by Central and Eastern European (CEE) countries. CEE countries participate in 41% of all signed treaties. Most of them, about 70%, were signed since 1985. As outlined by Blonigen and Davies (2004), tax treaties perform four primary functions. The first is to standardize tax definitions of treaty partners. Differing tax definitions can lead to double taxation and inefficient capital flows. The second is to reduce transfer pricing and other forms of tax avoidance. The third goal of tax treaties is to prevent treaty shopping. Treaty shopping is described as the routing of income through particular countries in order to take advantage of treaty benefits that were designed to be given only to residents of the contracting countries. The most common rules regarding treaty shopping restrict treaty benefits if more than 50% of a corporation’s stock is held by a third, non-treaty country’s residents (Doernberg, 1997). Finally, tax treaties affect the actual taxation of multinational corporations. They do so through the provisions for double taxation relief and the rules that reduce maximum allowable withholding taxes on three types of remitted income: dividend payments, interest payments, and royalty payments. From the above discussion, we can see that theoretically tax treaties can both promote and reduce investment. On the one hand, tax treaties can promote

the effect of tax and investment treaties on bilateral fdi flows 691

investment by reducing uncertainty about the tax environment abroad; on the other hand, a tax treaty that reduces the ability to transfer price or somehow prevents other types of tax evasion also reduces the incentive to invest (if investment activity is purely for tax minimization reasons). Empirically, the question about what dominates remains open. 3. Literature review A dramatic increase of foreign direct investment during the 1990s led to a boom in economic research studying the forces affecting FDI. A part of this literature looks at the relation of government policies and FDI. In this section, we review the literature that deals with international treaties and their effect on FDI. While most economic texts use the assumption that treaties encourage FDI, one can find little evidence in support of this statement in the empirical literature. We start with the review of studies that deal with the impact of tax treaties on FDI. These are not very numerous. Blonigen and Davies (2000) were the first to directly explore the effects of tax treaties on foreign direct investment. They use data over the period 1966–1992 on U.S. inbound and outbound FDI. Several approaches are introduced to capture the effect of treaties. A first approach uses a simple dummy variable that indicates whether there is a treaty or not for a specific country pair. In the second approach, the authors use a treaty age variable equal to the number of years that a treaty had been in effect. In both cases, the authors find positive and significant effects and conclude that tax treaties have a strong, positive impact on FDI. One problem with this approach is that it combines the effects of older treaties, in place long before their sample period began, with more recent ones. Since the old treaty partners of the United States (Europe, Japan, Canada, Australia, New Zealand) are also the largest hosts and homes for U.S. FDI, this means that treaty variables may have been capturing some unobserved differences between these countries and other nations. Therefore, in the more recent version (Blonigen and Davies, 2004) they separate the treaty dummy and include two dummy variables: one for old treaties, the other for new treaties. With two dummies they find that the old treaty dummy has a positive and significant effect, while the new treaty dummy has a negative and significant effect. Blonigen and Davies employ the same approach in their 2002 paper, in which they use inbound and outbound FDI stock and flow data for OECD countries from 1982–1992. Their findings are similar to those in the previous work: when old and new treaties are not separated, the data indicates a positive and significant effect; after separating treaties, the old treaty estimate is positive and significant, whereas the new treaty estimate is negative and significant. Louie and Rousslang (2002) use a different approach but also do not find strong support for FDI encouragement. They use 1990s income tax return data for U.S. multinational enterprises (MNEs) to calculate the rate of return for foreign subsidiaries. They test whether this rate of return changes after a treaty is ratified. They do not separate the effects of old and new treaties, but only count

692 tom coupé, irina orlova and alexandre skiba

those treaties that had been in force prior to 1987. The coefficient on their treaty dummy is significant and negative; however, when authors include in the regressions proxies for corruption and political instability, the significance of the treaty dummy fades entirely. Thus, they attribute their initial result to the omitted variable bias and conclude that good governance attracts both FDI and tax treaties, but that treaties have no effect on FDI. Not all studies considered above account for treaties signed throughout the 1990s, while most transition countries emerged during that period. Moreover, CEE countries concluded the majority of tax treaties during the 1990s. On the contrary, studies that investigate the influence of BITs on FDI flows do cover the 1990s. UNCTAD (1998) used a dataset for 133 countries and 200 BITs to show that there is a rather weak correlation between the existence of a treaty and an increase in FDI. In this research, only cross-sectional data was used (year 1995). Hallward-Driemeier (2003) used data on FDI flows from twenty OECD countries to thirty-one developing countries over the period of 1980– 2000. She also finds little evidence of a strong positive correlation between bilateral investment treaties and increase in FDI flows. Salacuse and Sullivan (2005) provide three cross-sectional analyses of FDI inflow to ninety-nine developing countries in the years 1998, 1999 and 2000, as well as a fixed-effects estimation of the bilateral flow of FDI from the U.S. to thirty-one developing countries over the period of 1991–2000. They find that signing a BIT with the United States is associated with higher FDI inflows, whereas the number of BITs signed with other OECD countries is statistically insignificant. But since this analysis is cross-sectional, it cannot detect how a higher number of BITs raises the flow of FDI to developing countries over time. In addition, in their fixed-effects regression the authors do not include year dummies, which could mean that the treaty dummies reflect year effects. Neumayer and Spess (2005) use a larger and more representative sample compared to previous work. It covers the period from 1970–2001 and includes 119 developing countries. Their main explanatory variable is the cumulative number of BITs a developing country has signed with OECD countries. They find a strong positive effect of BITs on FDI inflows. They use a non-dyadic research design, which means that the authors analyze total OECD FDI flows into a developing country, instead of looking at country-to-country flows. They criticize Hallward-Driemeier (2003) and Salacuse and Sullivan (2005) on the grounds of using a dyadic modeling, which cannot capture the potential of BITs to attract more FDI from other developed nonsignatory countries as well, and therefore may underestimate the effect of BITs. Tobin and Rose-Ackerman (2005) analyze the impact of BITs using nondyadic FDI inflows, in a panel from 1980–2000. Their general conclusion is that a higher number of BITs lowers the FDI a country receives at high levels of political risk and raises the FDI only at low levels of risk. In an additional dyadic analysis

the effect of tax and investment treaties on bilateral fdi flows 693

of fifty-four countries, they fail to find any statistically significant effect of BITs on FDI flows from the United States to developing countries. Summarizing the above studies on BITs, different approaches tend to lead to different conclusions: authors that use a dyadic approach find no effect, while authors that use a non-dyadic approach do find an effect. In our research we adhere to the dyadic design, since it controls for country pair specific effects, and hence allows us to control better for omitted variables. In contrast to previous studies that use a dyadic design and despite the fact that we work with a smaller sample, include both BIT and DTTs and tackle the endogeneity issue, we do find robust empirical evidence of a positive effect of BITs on FDI. B. Methodology and data 1. Methodology The empirical specification is based on the so-called gravity model. The gravity model was developed by Tinbergen (1962) and Poyhonen (1963). Originally, it was used to analyze only trade flows between countries. Theoretical work on the gravity model for FDI, however, is scant. One prominent study is Bergstrand and Egger (2005). The general idea is the following: the amounts of bilateral resource flows will positively depend on size of source/ destination countries (usually represented by GDP, sometimes by population size or land area, or even all mentioned factors simultaneously), which just reflects potential supply/demand, and negatively by transportation costs (that is inversely proportional to physical distance between countries). Usually, the gravity equation takes a log-linear form:

where FDIij – FDI inflows from country i to country j GDPi – GDP of country i GDPj – GDP of country j DISTij – distance between the capitals of countries i and j Dij – dummy variables (treaty dummy, common border, common language dummy) k – indicates the number of dummy variables included in the regression

In this study, we will focus on transition countries. There are multiple reasons for this focus. First, a lot of newly concluded treaties involved transition countries; hence, a sample of transition countries is one where there is a lot of variation in the ‘treatment’ variable over time. Such variability is important to get precise (more efficient) estimates. Second, by focusing on transition countries we get a relatively homogenous sample; while we sacrifice degrees of freedom by restricting our sample, we decrease the extent of omitted variables. Third, for the transition countries, the European Bank for Reconstruction and Development (EBRD) reform ratings are available which provide us with a good proxy for a wide range of home policies.

694 tom coupé, irina orlova and alexandre skiba

Previous studies either include a BIT dummy or a tax treaty dummy as an explanatory variable. In this study, we include both dummies simultaneously to avoid omitted variable bias. In addition, instead of including just one tax treaty dummy variable as is usually done, we classify tax treaties on the basis of their content and include three dummies. The International Bureau of Fiscal Documentation categorizes tax treaties in the following manner: income/capital tax treaties, social security treaties, administrative assistance, inheritance/gift, and transport tax treaties. We pick up three major categories out of this classification, which constitute the vast majority of treaties. These are: 1) income and capital tax treaty; 2) income tax treaty and 3) social security treaty. Income and income and capital tax treaties can be differentiated on the bases of taxes covered. Thus, income and capital tax treaties cover taxes on both income and capital, while income tax treaties cover taxes on income only. For instance, Ukraine has an agreement with the Federal Republic of Germany for the avoidance of double taxation with respect to taxes on income and capital (income and capital tax treaty). This treaty covers the following taxes in Germany: income tax, corporation tax, capital tax, and trade tax; and in Ukraine it covers the following taxes: tax on profit of enterprises, income tax on citizens, tax on property of enterprises, and tax on immovable property of citizens. Whereas with Sweden, Ukraine has concluded only an income tax treaty. This treaty covers taxes on profit of enterprises and personal income earned in Ukraine. The same treaty covers the following taxes in Sweden: state income tax, special income tax for non-residents, special income tax for non-resident artistes, and the communal income tax. As to social security agreements, they are often signed with tax treaty partners. For example, United States has social security agreements with many of its tax treaty partners. Under these agreements, many people who work or have worked for both countries can receive credit for work performed in both countries under the social security system of one country. We allow for more differentiation between tax treaties because it does matter what treaty or treaties among the three mentioned above has/have been signed. Different treaties imply different degrees of integration between the countries. For instance, the existence of an income tax treaty between two countries implies a lower degree of integration as opposed to the case where an income and capital tax treaty is in place. Moreover, a country with an income and capital tax treaty might be more attractive than a country with just an income tax treaty. In our sample there can be, at most, two tax treaties in place between any two countries, since income tax treaties and income and capital tax treaties are mutually exclusive. We incorporate this information in our research design with the following method: We include three dummies: income and capital tax treaty dummy, income tax treaty dummy and social security treaty dummy. The signs and magnitudes of correlations between these tax treaty dummies vary substantially. The ICT dummy and IT dummy are negatively correlated (–0.68) which makes sense since these treaties are mutually exclusive; IT and SS treaties

the effect of tax and investment treaties on bilateral fdi flows 695

are negatively correlated (–0.17) while there is a positive correlation between ICT and SS treaties (0.19). 2. Variables Our dependent variable is the annual bilateral flow of FDI, converted to constant U.S.$ of 2000 with the help of the U.S. Implicit Price Deflator. Since we take the natural log of the dependent variable, we lose all zero FDI flows. In our analysis we disregard zero flows (259 out of 1,224 observations). But if instead we set all zero FDI flows to positive FDI flows of U.S. $1, as do Neumayer and Spess (2005), our results are hardly affected. We only include flows from OECD countries to transition countries since we are interested in how FDI streams to developing countries can be stimulated (the list of countries can be found in Appendix C). Our main explanatory variables are four bilateral treaty dummies: (1) BIT dummy capturing the effect of bilateral investment treaties; (2) ICT dummy capturing the effect of income and capital tax treaties; (3) IT dummy capturing the effect of income tax treaties; and (4) SS dummy capturing the effect of social security treaties. All four variables are constructed as zero/one dummies. Thus, we have one bilateral investment treaty dummy and three bilateral tax treaty dummies. Since income tax treaty (IT) and income and capital tax treaty (ICT) are mutually exclusive, the maximum number of tax treaties per country pair is two. The possible combinations are: (1) social security treaty and income tax treaty and (2) social security treaty and income and capital tax treaty. Each dummy takes on the value of one for every year starting from the ratification year. To have an effect on FDI attraction, a treaty needs to be ratified. Nevertheless, there can be initial mover advantages and investors might start investing long before the actual date of ratification. Therefore, as a robustness check, we also run a specification with a dummy based on the date of signature. Our control variables are those conventionally considered as the determinants of FDI in a gravity model. We include the natural logs of host and source GDPs; host GDP (GDPj) is used as an indicator of market size, while source GDP (GDPi) is proportional to the pool of potential investors from source country. Natural log of host’s GDP per capita (GDPCAPj) proxying for production costs is also included. However, as discussed by Globerman and Shapiro (2002), the sign of the impact of GDP per capita is ambiguous, since this variable both reflects the level of development (encouraging FDI) and the level of wages which can discourage inward FDI if not compensated by productivity. The distance variable (DISTij) is calculated as a mean distance between the main towns of each country. These basic variables are complemented with a dummy for contiguity (CONTij) or common border and another one for common language (COMLGij). Like distance, these two variables account for various transaction costs incurred when investing abroad. Common border and common language data are taken from CIA World Factbook. Since transition countries that are members of the World Trade Organization (WTO) might receive more FDI as it is easier to export goods back to the home

696 tom coupé, irina orlova and alexandre skiba

country, we account for this with a dummy variable (WTOj) indicating whether a host country is a member of the WTO or not. To control for host country policies and quality of institution, we include a composite transition index (INDEXj) compiled from the EBRD Transition Indicators. These are nine qualitative country-by-country indicators with the scale from 1 (little/no progress) to 4.33 (standards of advanced market economy) covering first phase reforms (small scale privatization, price and trade liberalization) and second phase reforms concerning institutional development (large scale privatization, governance, competition, infrastructure, and financial institutions). The transition index is constructed as the first principal component of the EBRD Transition Indicators for the host countries over the whole sample period. The first component accounts for 80.55% of the variance in the nine EBRD Transition Indicators. Thus, our baseline equation is the following:

3. Data description We have created a database of tax treaties involving transition countries. The data includes the dates of signature, ratification, and termination. We find a few cases3 when a treaty was terminated, but in the same year a new treaty was ratified. In these cases since the treaties were re-signed within a year of termination the dummies do not have an interruption. We have also considered the difference between the old/inherited and new treaties. Inherited treaties are ones which were in place before our sample began.4 Inherited treaties may complicate identification of the treaty effect. If we get a positive correlation between the tax treaty variable and FDI, it will not be clear whether other unobservable characteristics may be leading to increased FDI activity. This occurs because the tax treaty variable will pick up any residual effects on

3. Renewed tax treaty dummies constitute less than 3% of observations. 4. These are SS, ICT, and IT, which were mostly signed during the 1960s and 1970s, as well COMECON or CMEA tax treaties. The Council for Mutual Economic Assistance (CMEA) was dissolved in 1990. However, the tax authorities of the signatories to the multilateral CMEA treaties have agreed to observe the provisions of the treaties amongst themselves until new bilateral treaties are in place. This agreement took place before the dissolution of the USSR. Original parties to this treaty were: Bulgaria, Czechoslovakia, Hungary, German Democratic Republic, Mongolia, Poland, Romania and the USSR. Following the dissolution of the USSR, the Members of the Commonwealth of Independent States (CIS) have, in principle, agreed to honor the international treaties, including income tax treaties, concluded by the USSR until new treaties have been negotiated in their own name. However, some exceptions took place. For example, Estonia, Latvia, and Lithuania have announced their general disapproval of USSR treaties.

the effect of tax and investment treaties on bilateral fdi flows 697

FDI that are not measured by other regressors. The majority of treaties in our sample, however, were signed after our sample data began. We will call them new treaties. Inherited treaties comprise only about 6% of our data. Thus, in this respect our treaties can be divided into two categories: inherited have only crosscountry variation and no time-series variation, however, cross-country variation is sufficient; and new treaties have both cross-country and time-series variation. These new treaties have a better opportunity to measure the impact of a tax treaty, as we have data on FDI activity before and after the treaty. The data on tax treaties is taken from International Bureau of Fiscal Documentation tax treaties database. As to bilateral investment treaties, this data is drawn from UNCTAD database on Bilateral Investment Treaties. FDI data is drawn from OECD Direct Investment Database. The source for OECD data are the Central Banks and Statistical Offices of the FDI home countries, which follow the recommendations of the International Monetary Fund (IMF) Balance of Payments Manual and the OECD Benchmark Definition of Foreign Direct Investment. OECD data is considered more reliable compared to other databases compiled on the basis of host country statistics, since it is based on IMF methodological guidelines and therefore tends to be more uniform. Host country source of data in case of FDI is less reliable because of large discrepancies. Even though new European Union (EU) member states and the candidate countries now follow the IMF definition of foreign direct investment, deviations were frequent in the past. For instance, most Western Balkan countries still fail to report all the forms of FDI (equity investment, reinvested profits, other investment). There are discrepancies due to the very definition of FDI as well. According to IMF Balance of Payments Manual, Revison 5, capital investment abroad is regarded as foreign direct investment if the purpose is to establish and maintain permanent equity relations with a foreign company and at the same time to exercise a noticeable influence on the management of that company; the share of the foreign investor must make up at least 10% of the target firm’s equity capital. However, not all the countries apply the 10% equity threshold. Thus, as mentioned above we consider OECD data to be more reliable. Also, it is observed that most of the outward FDI emerges from developed countries and about 87% of this comes from the OECD countries, where the top three positions are taken by United States, United Kingdom and France (see Appendix B). Our dataset covers seventeen OECD (home) countries and nine transition (host) countries over the period of 1990-2001. We should mention here as well that OECD data does not contain negative values of FDI (disinvestment). The OECD database contains FDI data in nominal terms; in local currencies. With the help of annual average exchange rates and Implicit Price Deflator (IPD) we convert flow data to U.S. dollars 2000. Data on GDP is borrowed from the World Development Indicators (WDI) database. Information about common borders and common language is available in the CIA World Factbook. Common border may affect the amounts of capital flows.

698 tom coupé, irina orlova and alexandre skiba figure 1. cumulative number of bilateral investment treaties Bulgaria

Czech Republic

Hungary

Poland

Romania

Russia

Slovakia

Slovenia

Ukraine

Cumilative number of signed BITs (by host)

15 10 5 0 15 10 5 0 15 10 5 0 1990

1995

2000 1990

Graphs by FDI host country

1995

2000 1990

1995

2000

Year

Distances between countries’ capitals are calculated using online source http:// www.indo.com/cgi-bin/dist/. Data on WTO dates of membership is taken from the World Trade Organization official Web site http://www.wto.org/english/thewto_e/whatis_e/tif_e/org6_e.htm. The composite transition index is compiled on the basis of EBRD Transition Indicators borrowed from the EBRD Web site. Information on treaties is summarized in figures 1 through 3. Figure 1 with the cumulative number of bilateral investments treaties by FDI host country demonstrates the differences in the treaty signing patterns among transitioning countries. We will exploit these differences to identify the effect of treaties on FDI. Hungary, Poland, Czech Republic come into our data having investment treaties with most source countries while during the same period there are some country pairs that do not sign a treaty. Bulgaria, Romania, Slovenia, and Ukraine show the most activity and sign most of the treaties between 1990 and 2002. Even among the most dynamic countries there is some variation in the timing of activity. Romania and Ukraine sign most of the treaties in the first half of the 90’s while Slovenia and Bulgaria are still signing treaties towards the end of the decade. Figure 2 shows the cumulative number of investment treaties and cumulative number of the double taxation treaties (IT and ICT). The figure suggests that the periods of signing of both types of treaties often coincide and both exhibit a fair amount of variation over time. Figure 3 describes components of the total number of BITs by groupings of FDI source countries. Slovenia starts signing treaties with the non-G7 European countries without a common border and does

the effect of tax and investment treaties on bilateral fdi flows 699

Cumulative number of signed treaties

figure 2. cumulative number of investment and double taxation treaties Bulgaria

Czech Republic

Hungary

Poland

Romania

Russia

Slovakia

Slovenia

Ukraine

20 15 10 5 0 20 15 10 5 0 20 15 10 5 0 1990

1995

2000 1990

1995

2000 1990

1995

2000

Year IT and ICT

BIT

Graphs by FDI host country

Cumulative number of signed BITs

figure 3. cumulative number of bilateral investment treaties by groupings of fdi source countries Bulgaria

Czech Republic

Hungary

Poland

Romania

Russia

Slovakia

Slovenia

Ukraine

15 10 5 0 15 10 5 0 15 10 5 0 1990

1995

2000 1990

1995

2000 1990

1995

2000

Year total

with G7 countries

with adjacent countries

with European countries

Graphs by FDI host country

Notes: 1) Canada is not among G7 countries due to unavailability of FDI flows data; 2) non-European countries are the Republic of Korea, Japan, and the U.S.

700 tom coupé, irina orlova and alexandre skiba

not sign any treaties outside Europe. Romania signs treaties with G7 European countries before expanding to other countries in Europe. Ukraine starts with G7 Europeans, continues to other European countries before signing outside the continent. 4. Empirical estimation and results a. Estimation procedures. Our sample can be grouped in two different ways: by country pairs (153 groups) and by host countries (9 groups). We do the estimation for both groupings. When we do the host country grouping we keep host characteristics constant and look across source countries. When we do country pair grouping we take a pair and look within the pair trying to identify whether the timing of the effect played a role. Our estimation starts with several tests, which allow us to decide whether fixed effects, random effects, or simple OLS on pooled data should be preferred. First, we look at panel versus simple OLS on pooled data estimation. Since we suspect country-specific effects in the data, we perform fixed-effects, and the F-test applied after carrying out the fixed-effects estimation answers the question of fixed-effects vs. pooled OLS. Second, if F-test suggests that panel estimation should be preferred, the Hausman specification test is applied to discriminate between fixed and random effects estimation of the panel data. The tests mentioned above suggest fixed effects specification for host country grouping and random effects specification for country pair grouping. There is potentially a problem of endogeneity: countries with larger bilateral FDI flows are also more likely to enter into an agreement. One way to tackle endogeneity is to use instrumental variables estimation. Successful implementation of this approach hinges on availability of valid instruments. What we need are variables that affect the probability that two countries conclude a treaty but are not directly related to the volumes of FDI between these two countries. We will refer to these factors as country’s propensity to conclude treaties. One possibility is to use the number of other tax/investment treaties a host has entered into with countries other than the source country being considered. We will further refer to these treaties as outside treaties. This is exactly what Hallward-Driemeier (2003) does. She writes: “The willingness of a host to ratify a BIT, as measured by the number of outside BITs, should be correlated with the probability it signs with this particular host country, but shouldn’t affect the amount of FDI that particular source country would send. Thus, when U.S. investors are considering investing in India, their decision would not be affected by whether India has ratified treaties with the UK or France. However, that India has entered other treaties would be expected to influence their willingness to enter such a treaty with the U.S.” Note that those who claim the non-dyadic approach is better, like Neumayer and Spess (2005) believe that BITs do have positive spillover effects because of what they call the signaling effect. The signing of BITs (especially with major capital exporting countries) sends out a signal to potential investors that the developing country is generally serious about the protection of foreign investment. The authors assume that it is difficult to say

the effect of tax and investment treaties on bilateral fdi flows 701

how important the signaling effect is (which benefits investors from all countries), compared to the commitment effect (which only relates to investors from BIT partner countries). Still, if the signaling effect is important, how can one explain, for example, Ukraine signing sixty-six tax treaties and fifty-eight BITs with other countries. Indeed if spillovers were important we would observe that countries would sign only treaties with the most important capital exporting nations. Therefore, in our opinion, the dyadic approach is better because it allows us to control for omitted variables (unlike the non-dyadic approach) and the propensity of signing treaties can be used as a valid instrument. Instead of using just the number of outside treaties a host country has, we construct another instrument—the sum of outside treaties a host and a home country have. A pair of countries is more likely to conclude an agreement, when both countries, not just the host country, have the habit of signing such agreements. Our instruments are intended to reflect the overall propensity of countries to sign such types of treaties. The coefficients of correlation between treaty dummies and suggested instruments range from 42% (SS) to 64% (ICT). We have tried other possible options for instruments. Among them the conventionally used number of outside treaties a host country has, as well as the number of outside treaties a home country has, and possible combinations between the two. We have also tested the option of using the number of treaties signed during the last five years for both home and host countries. The logic behind this is that those countries that have been more active signing treaties in the last five years are more likely to conclude such agreements now. The highest correlations between the instruments and instrumented variables are found for the sum of outside treaties that both host and home countries have for all years. Previous studies based on IV methodology, use the number of outside treaties a host country has signed. To test whether the instruments we use (sums) are stronger than the traditional alternative (outside treaties of hosts) we compare first-stage F-statistics.5 They are significantly higher for the instrument we use, especially for ICT and IT. It is worth noting that we cannot use measures of openness or the transition index to instrument for investment treaties because those measures inevitably correlate with the FDI flows. Another potential problem is a statistically significant relationship between two upward trending variables that is spurious. We deal with this problem by including year dummies to account for any year-to-year variation in FDI flows unaccounted for by our other explanatory variables. Finally, we explore the timing of the treaty effects on FDI. A dummy for each of the three years prior to and after the ratification of the treaties is included. Another dummy takes on the value of unity for years after the third year from ratification. We estimate the effects of timing in specifications with host and pair

5. F-statistics reported in the Appendix D.

702 tom coupé, irina orlova and alexandre skiba

random effects, and pooled OLS. The coefficients on the timing dummies should be interpreted in terms of difference with the reference period. In our case the reference period comprises all years up to 4 years prior to ratification. The timing of the treaty effect has been investigated in a similar fashion by HallwardDriemeier (2003). The difference is that we include a separate dummy for all years after three years from ratification of each treaty. The usefulness of our approach lies in separating the periods before and after the 3-year window, which otherwise are lumped together in the reference category and forces the effect of treaties to die out. Our approach allows the coefficient on a treaty timing dummy to be interpreted as the difference between the level of FDI in that year and the level of FDI from years earlier than three years prior to ratification. We expect that the effect of the third year prior and after the ratification will be estimated less precisely because there are fewer observations for these dummies. b. Results. Table (1) reports pooled OLS estimation results for the logged amount of FDI in U.S. dollars of 2000 flowing from OECD country to a transition country as the dependent variable. As a starting date for our treaty dummies in columns (1) and (2) we use the date of ratification and then the date of signature as a robustness check in column (3). Almost all control variables are significant and of expected sign. The larger the home country and the host country, the larger is the FDI flow. Flows are also higher to host countries with smaller GDP per capita, which makes sense if we think of GDP per capita as a proxy for production costs. Our distance variable is negative, while having a common language is positive, both are significant and in accordance with theoretical expectations. Contiguity (common border) dummy is insignificant. This might be because few neighboring countries are present in our sample. The transition index as a measure of institutional quality in host countries shows a positive significant effect. Our variables of interest are significant at least in columns (1) and (2); in column (3) only the bilateral investment treaty dummy and the social security treaty dummy remain significant. Since column (3) presents the results based on the date of signature, this supports the hypothesis that to have an effect on FDI a treaty needs to be ratified. All treaty dummies but SS have a positive effect. Thus, according to the results in column (1), the existence of a BIT between two contracting parties increases the FDI flow by 44%. If ICT or IT is in place we get a 94% or 66% increase correspondingly, while SS treaty decreases FDI flow by 54%. There could be a possible explanation of this negative effect based on the tax evasion provisions, which can serve as disincentives to engage in FDI activity.6 In column (2) we interact the existence of a BIT with the institutional quality (proxied by the Transition Index). We do so to test whether BITs are only valuable within a country with a certain level of overall institutional development.

6. A detailed explanation can be found below.

the effect of tax and investment treaties on bilateral fdi flows 703

table 1. ols estimation results (logged fdi flows in $u.s. 2000) Variables

OLS (rat)

Interaction with transition index

OLS (sign)

lnGDP host

BIT∗Transition Index

0.71068 (0.000)∗∗∗ 0.91832 (0.000)∗∗∗ −0.22465 (0.099) −0.79725 (0.000)∗∗∗ 2.26368 (0.000)∗∗∗ −0.19366 (0.327) 0.47526 (0.000)∗∗∗ −0.05892 (0.741) 0.44207 (0.001)∗∗∗ 0.93916 (0.004)∗∗∗ 0.66218 (0.037)∗∗ −0.5422 (0.001)∗∗∗ –

0.7603 (0.000)∗∗∗ 0.9478 (0.000)∗∗∗ −0.3485 (0.007)∗∗ −0.88202 (0.000)∗∗∗ 2.1495 (0.000)∗∗∗ −0.2832 (0.161) 0.54209 (0.000)∗∗∗ −0.0322 (0.860) 0.2829 (0.052)∗∗∗ 0.2482 (0.380) −0.10102 (0.719) −0.5446 (0.000)∗∗∗ –

N. Obs R2

962 0.4878

0.73006 (0.000)∗∗∗ 0.9161 (0.000)∗∗∗ −0.2749 (0.044)∗∗ −0.83469 (0.000)∗∗∗ 2.2405 (0.000)∗∗∗ −0.28437 (0.152) 0.6187 (0.000)∗∗∗ 0.7427 (0.681) 0.70463 (0.000)∗∗∗ 0.9908 (0.002)∗∗∗ 0.6652 (0.034)∗∗ −0.5934 (0.000)∗∗∗ −0.22456 (0.000)∗∗∗ 962 0.4832

lnGDP home lnGDPCAP host lnDistance Common language Contiguity Transition Index WTO BIT ICT IT SS

962 0.4788

Note: p-values in parentheses; ∗− significance at 10%; ∗∗− 5%; ∗∗∗− 1% Robust standard errors; year dummies not reported.

Two possibilities exist: BITs can act either as complements or as substitutes for strong domestic protection of property rights. A positive interaction term on institutional quality and the ratification of a BIT would favor the former interpretation, while a negative interaction term would favor the latter. The results show a negative interaction, thus revealing that BITs per se will have a stronger effect in a weak institutional setting. Because tax treaties are not really aimed at

704 tom coupé, irina orlova and alexandre skiba

complementing or substituting domestic protection of property rights, we do not interact them with the institutional quality. When significant the per capita GDP of the host country negatively affects the investments inflows. This finding is expected if FDI is carried out to take advantage of the low labor cost, (vertical FDI) rather than to access the market (horizontal FDI). Producers seeking lower cost locations are more likely to channel their investments into countries with lower incomes. Another possible explanation comes from the fact that the low income counties experience relatively fast growth in the FDI inflows because they start from a lower stock of FDI. In that case the coefficient on per capita GDP captures the effect of catching up by the lower income economies. Table (2) reports panel estimation results. Hausman test fails to reject the random-effects assumption (time-invariant factors are uncorrelated with the explanatory variables) for country pair grouping; but rejects this assumption for host country grouping. We therefore present the random-effects estimation results for pair grouping in column (1) and fixed-effects for host grouping in column (2). For both of these specifications we interact BIT with Transition Index in columns (3) and (4). For the random effects pair specification, as a robustness check, we use the date of signature instead of the date when a treaty was ratified (column 5). Throughout all the specifications control variables stay significant and of the expected sign with the exception to host GDP p.c, contiguity, and WTO variables, which are insignificant. The WTO dummy turns negative significant in host fixed effects specifications. This is a rather strange result, because the WTO dummy is expected to have a positive effect. But due to the fact that all countries in our sample entered the WTO in 1995 (except for Bulgaria—1996—and Russia and Ukraine, which have not yet entered the WTO), variation in this variable is low, and we probably should not pay too much attention to this. Among the variables of interest only BIT and SS show consistency in the signs and significance throughout all specifications.7 BITs encourage FDI flows and this effect ranges from 42% to 83%, which is close to what we get in pooled OLS estimation. All specifications give a negative significant SS result, which is puzzling because SS should increase FDI if they remove the possibility of paying social security contributions twice. There could be a possible explanation originating from legal literature on tax treaties, however, which assumes that treaties are intended to reduce tax evasion by MNEs, not reduce double taxation. In particular, the matter concerns transfer pricing provisions and the exchange of tax information between governments. These measures can offset the FDI-promoting effects of tax treaties.

7. If we use pair Fixed effects, BIT and SS remain significant, while the other tax treaties are insignificant.

the effect of tax and investment treaties on bilateral fdi flows 705

table 2. panel estimation results (logged fdi flows in $u.s. 2000) Variables

Pair RE (rat)

Host FE (rat)

Pair RE Host FE Pair RE (interaction) (interaction) (sign)

lnGDP host lnGDP home lnGDPCAP host lnDistance

0.68945 (0.000)∗∗∗ 0.93828 (0.000)∗∗∗ 0.20578 (0.300) −0.84545 (0.000)∗∗∗ 2.4291 (0.007)∗∗∗ −0.49278 (0.232) 0.24028 (0.001)∗∗∗ −0.25549 (0.135) 0.42502 (0.005)∗∗∗ 0.45236 (0.087)∗ 0.17418 (0.524) −0.62424 (0.012)∗∗ –

1.09312 (0.023)∗∗ 0.93966 (0.000)∗∗∗ −0.739452 (0.468) −0.8528 (0.000)∗∗∗ 1.87444 (0.000)∗∗∗ −0.04306 (0.833) −0.0848 (0.562) −0.5515 (0.052)∗ 0.50338 (0.000)∗∗∗ 0.79268 (0.001)∗∗∗ 0.55490 (0.018)∗∗ −0.50398 (0.004)∗∗∗ –

0.7193375 (0.000)∗∗∗ 0.9390046 (0.000)∗∗∗ 0.1945107 (0.321) −0.8865569 (0.000)∗∗∗ 2.387142 (0.007)∗∗∗ −0.5718995 (0.157) 0.3283385 (0.000)∗∗∗ −0.1929713 (0.262) 0.5234234 (0.001)∗∗∗ 0.4807682 (0.067)∗ 0.1947564 (0.474) −0.6498642 (0.008)∗∗∗ −0.140099 (0.012)∗∗

Common language Contiguity Transition Index WTO BIT ICT IT SS BIT∗ Transition Index N. Obs R2

962 962 within=0.35 within=0.41 between=0.52 between=0.02 overall=0.47 overall=0.22

1.276357 (0.007)∗∗∗ 0.9360072 (0.000)∗∗∗ −0.3901564 (0.699) −0.8995257 (0.000)∗∗∗ 1.881151 (0.000)∗∗∗ −0.1636568 (0.421) 0.0599106 (0.685) −0.5970304 (0.034)∗∗ 0.8333026 (0.000)∗∗∗ 0.8302537 (0.000)∗∗∗ 0.5394726 (0.020)∗∗ −0.5165181 (0.003)∗∗∗ −0.294981 (0.000)∗∗∗

0.7201158 (0.000)∗∗∗ 0.9555236 (0.000)∗∗∗ 0.1449517 (0.466) −0.8962245 (0.000)∗∗∗ 2.421931 (0.009)∗∗∗ −0.5446089 (0.194) 0.2686834 (0.000)∗∗∗ −0.2463965 (0.150) 0.4527339 (0.006)∗∗∗ − 0.0689651 (0.770) −0.3384541 (0.197) −0.5626095 (0.011)∗∗ –

962 962 962 within=0.35 within=0.43 within=0.35 between= 0.53 between=0.07 between=0.50 overall=0.47 overall=0.25 overall=0.45

Note: p-values in parentheses; ∗− significance at 10%; ∗∗− 5%; ∗∗∗− 1% Year dummies not reported.

706 tom coupé, irina orlova and alexandre skiba

The ICT dummy displays sufficient consistency too. It is positive and significant in the first four specifications, and turns insignificant only in the last column, where we do a specification based on the date of signature. This again supports the hypothesis that to have an effect on FDI a treaty needs to be ratified. The IT dummy is significant (with positive sign) only in host fixed effects specifications, but not in country pair random effects specifications. Including a traditional tax treaty dummy (one if there is any tax treaty and zero if there is no treaty) like in previous works shows no significant effect. The interacted BIT∗Transition Index gives a negative significant result, which is consistent with what we get for pooled OLS specification. Table (3) presents Instrumental Variables estimation results. It starts with IV (2SLS) regressions followed by fixed-effects IV regression. Hausman test suggests country pair random effects, as a robustness check we also do pair fixed effects (column 5). In column (4) the interacted BIT∗Transition Index is included. We simultaneously instrument four treaty dummies: ICT, IT, SS, and BIT. As discussed above, the instruments used are the sums of outside treaties that both host and home countries have corresponding to each treaty type (ICT, IT, SS, BIT). Tax treaty dummies are insignificant throughout IV specifications, except for SS treaty, which is negative significant in columns (1) and (2). BIT is the only treaty that stays significant throughout all specifications, and the effect of BIT is unchangeably positive. As can be seen from the table below, instrumenting treaties produces a very large coefficient in case we do not control for country pair effects, while the pair fixed effects and random effects estimates remain fairly reasonable. Given that the size of the coefficient increases, high FDI flows would decrease the chances of having a treaty. Indeed, if there is already a high FDI inflow, why would one want to sign a treaty? c. Timing of the treaty effect. There is no clear theoretical prediction about the timing of the treaty effect. A treaty can foster investment or itself be a response to an increase in the investment activity between two countries. The exact length of the lag or lead in the treaty effect is also an empirical question. In order to explore the timing of the treaty effect we augment the parsimonious specification with the dummies for each of the three years prior and each of the three years after the treaty was ratified. All years after the third year are designated with a separate dummy. The three year window is chosen based on the limited length of our panel because for a wider window the estimates of the effects closer to the beginning and end of the window become less precise. In this setup our reference category corresponds to years prior to the third year before the earliest treaty was ratified. In other words, a positive and significant coefficient on “BIT Yr ratify +1” implies that in the year after a BIT was ratified there was an increase in annual FDI flows relative to the period more than three years prior to the ratification. The results are presented in Table 4. We estimate three specifications: pair random effects, host random effect, and a pooled OLS specification. Our findings are catalogued below along with possible explanations.

the effect of tax and investment treaties on bilateral fdi flows 707

table 3. iv estimation results (logged fdi flows in $u.s. 2000) Variables

lnGDP host

2SLS (rat)

0.0318697 (0.905) lnGDP home 1.233196 (0.000)∗∗∗ lnGDPCAP 2.3945 host (0.014)∗∗ lnDistance −0.150128 (0.653) Common 0.50637 language (0.738) Contiguity 2.0292 (0.022)∗∗ Transition −.52427 Index (0.152) WTO −0.09522 (0.816) 6.4553 BIT (0.003)∗∗∗ ICT −0.18744 (0.947) IT −1.55113 (0.617) SS −5.5825 (0.011)∗∗ BIT∗Transition – Index N. Obs 962

2SLS (sign)

Pair RE IV (rat)

Pair RE IV (interaction)

Pair FE IV (rat)

0.3938 (0.481) 1.1777 (0.000)∗∗∗ 1.4212 (0.272) −0.1538 (0.723) 1.7379 (0.194) 1.7283 (0.107) −0.0874 (0.791) 0.2946 (0.685) 6.8703 (0.001)∗∗∗ −3.6337 (0.690) −5.0019 (0.566) −3.8844 (0.075)∗ –

1.013625 (0.130) 0.6852966 (0.614) −0.1580569 (0.916) −0.3545127 (0.874) 3.785506 (0.611) −0.4026737 (0.879) −0.1123104 (0.354) −0.4881483 (0.144) 1.672409 (0.009)∗∗∗ 0.8162004 (0.946) −0.9587139 (0.932) 0.0011532 (0.999) –

1.4177 (0.071)∗ 0.2858 (0.850) −1.1032 (0.408) dropped

962

962

0.4605846 (0.058)∗ 1.089217 (0.003)∗∗∗ 0.43603 (0.374) −0.5040456 (0.511) 2.006902 (0.466) 0.1156395 (0.912) −0.1554021 (0.483) −0.6211766 (0.025) 1.677447 (0.002)∗∗∗ −1.605046 (0.716) −3.282159 (0.456) −1.0321 (0.496) 0.3204584 (0.219) 962

dropped dropped −0.18403 (0.173) −0.45214 (0.119) 1.4616 (0.061)∗ 0.9662 (0.920) −1.0424 (0.901) −0.51178 (0.760) – 962

Note: p-values in parentheses; ∗ − significance at 10%; ∗∗ − 5%; ∗∗∗ − 1% Year dummies not reported.

First, generally our results do not reveal an increase in investment before ratification. If anything, in one case we observe just the opposite and the levels of FDI drop before ratification, which is consistent with investors delaying their investments till ratification. Second, only the BITs encourage FDI. The FDI flows increase in the year of ratification and remain higher afterwards. “BIT Yr ratify +3” is significant only at the 10.2% level in the pair random effect specification. This lower significance

708 tom coupé, irina orlova and alexandre skiba table 4. timing of the treaty effect Variables

Pair random effects

Host random effects

OLS

lnGDPhost lnGDPhome lnGDPPChost EBRD Index lnDIST Language Contiguity WTO BIT Yr ratify −3 BIT Yr ratify −2 BIT Yr ratify −1 BIT Year ratify BIT Yr ratify +1 BIT Yr ratify +2 BIT Yr ratify +3 BIT After +3 ICT Yr ratify −3 ICT Yr ratify −2 ICT Yr ratify −1 ICT Year ratify ICT Yr ratify +1 ICT Yr ratify +2 ICT Yr ratify +3 ICT After +3 IT Yr ratify −3 IT Yr ratify −2 IT Yr ratify −1 IT Year ratify IT Yr ratify +1 IT Yr ratify +2 IT Yr ratify +3 IT After +3 SS Yr ratify −3 SS Yr ratify −2 SS Yr ratify −1 SS Year ratify SS Yr ratify +1 SS Yr ratify +2

0.712 (0.000) 0.954 (0.000) 0.199 (0.324) 0.259 (0.000) −0.797 (0.000) 2.467 (0.004) −0.356 (0.362) −0.161 (0.399) 0.350 (0.203) 0.213 (0.423) 0.353 (0.166) 0.855 (0.000) 0.487 (0.044) 0.601 (0.011) 0.402 (0.102) 0.591 (0.017) −0.549 (0.336) −0.965 (0.042) 0.303 (0.497) 0.421 (0.350) 0.486 (0.270) 0.181 (0.679) 0.064 (0.876) 0.359 (0.359) −0.830 (0.176) −0.762 (0.161) 0.052 (0.918) 0.367 (0.426) −0.278 (0.548) 0.358 (0.424) −0.177 (0.673) −0.160 (0.683) −0.090 (0.834) −0.141 (0.745) −0.220 (0.613) −0.962 (0.024) −0.645 (0.216) −0.506 (0.243)

0.684 (0.000) 0.886 (0.000) −0.206 (0.147) 0.460 (0.000) −0.723 (0.000) 2.170 (0.000) −0.057 (0.788) 0.024 (0.906) 0.027 (0.935) −0.032 (0.919) 0.162 (0.58) 0.665 (0.01) 0.425 (0.075) 0.433 (0.057) 0.255 (0.261) 0.538 (0.002) 0.226 (0.741) −0.476 (0.383) 0.412 (0.405) 0.721 (0.136) 0.535 (0.235) 0.500 (0.264) 0.367 (0.365) 1.226 (0.000) 0.025 (0.972) −0.535 (0.394) 0.119 (0.835) 0.574 (0.255) 0.104 (0.832) 0.679 (0.152) 0.321 (0.444) 0.830 (0.007) 0.218 (0.685) 0.153 (0.777) 0.171 (0.751) −0.639 (0.213) −0.071 (0.909) 0.044 (0.930)

0.684 (0.000) 0.886 (0.000) −0.206 (0.147) 0.460 (0.000) −0.723 (0.000) 2.170 (0.000) −0.057 (0.788) 0.024 (0.906) 0.027 (0.935) −0.032 (0.919) 0.162 (0.581) 0.665 (0.010) 0.425 (0.075) 0.433 (0.057) 0.255 (0.261) 0.538 (0.002) 0.226 (0.741) −0.476 (0.383) 0.412 (0.405) 0.721 (0.136) 0.535 (0.235) 0.500 (0.264) 0.367 (0.365) 1.226 (0.000) 0.025 (0.972) −0.535 (0.394) 0.119 (0.835) 0.574 (0.255) 0.104 (0.832) 0.679 (0.153) 0.321 (0.444) 0.830 (0.007) 0.218 (0.685) 0.153 (0.777) 0.171 (0.751) −0.639 (0.213) −0.071 (0.909) 0.044 (0.930)

the effect of tax and investment treaties on bilateral fdi flows 709

Variables

Pair random effects

Host random effects

OLS

SS Yr ratify +3 SS After +3 N.obs R-sq

−0.220 (0.632) −0.769 (0.010) 962 within = 0.369 between = 0.520 overall = 0.470

0.126 (0.805) −0.652 (0.001) 962 within = 0.412 between = 0.927 overall = 0.502

0.126 (0.805) −0.652 (0.001) 962 0.502

Notes: p-values are in parentheses. All specifications include year dummies. The choice between fixed and random effects was made based on a Hausman test of null hypothesis that there is no systematic difference between coefficients.

could be partially due to lack of variation because treaties ratified in or after 1999 will not have a Year+3 dummy by construction. Third, ICT and IT have a similar effect on FDI. Pair RE estimates reveal no effect at conventional levels of significance in all but one case. There seems to be a drop in FDI 2 years prior to a ratification of an ICT. The host RE (and OLS) suggest that ICT and IT both have a positive long term effect on FDI even though the exact timing is not statistically distinguishable. Fourth, there are indications that the effect of the social security treaty (SS) is negative. The preferred pair RE estimate picks up a decrease in the year of ratification. All specifications show a decrease in the level of FDI for the period of 4 years or more after the ratification.

conclusion The majority of economic texts stress the intuitive notion that bilateral tax treaties and bilateral investment treaties should promote FDI. However, not all the empirical studies support either hypothesis. The most prominent works on tax treaties, on the contrary, even find negative impacts on FDI. Whether bilateral investment treaties have an impact is more ambiguous. Papers using a dyadic approach do not find any evidence of BITs effect on FDI, in contrast to papers that use a non-dyadic approach, which tend to find a positive effect on FDI. Using OECD data we find that transition countries that have BITs with developed countries receive more FDI inflows from these countries. The effect is robust to various model specifications. On the contrary, tax treaties do not reveal any robust effect on FDI flows. This finding is particularly valuable for three main reasons: First, to our knowledge, this is the first study examining both bilateral investment treaties

710 tom coupé, irina orlova and alexandre skiba

and tax treaties simultaneously in one model. Second, it focuses on transition countries, which allows for more homogeneity in the sample. This is especially important, since then we can account for possible omitted variable bias. Thirdly we tackle the endogeneity issue. This study also provides evidence that BITs function to some extent as substitutes for institutional quality. In the majority of estimations that we do with interaction, the interaction term between the BIT dummy and a proxy for institutional quality is negative and statistically significant, implying that the net effect of a BIT is smaller (but still positive) if the quality of a host country’s institutions is better. We do not find any robust effect of tax treaties on FDI. When we include a traditional tax treaty dummy, which is one if there is any tax treaty and zero if there is no treaty, the result is consistent with what we get in general when doing tax treaty categorization – no significant effects are found. This can be explained by the two effects of tax treaties on FDI that might offset each other. One of them reduces double taxation, which should encourage FDI; the other prevents tax evasion through setting constraints on transfer pricing by MNEs, which might discourage FDI. Summing up, the main message we have for policy makers of developed and developing countries alike is that signing and ratifying BITs with developed countries does have the desired payoff of higher FDI flows.

the effect of tax and investment treaties on bilateral fdi flows 711

appendices appendix a Number of BITs and DTTs Concluded, Cumulative and Year to Year, 1990–2004 250

3000 2500

200

2000 150 1500 100 1000 50

500

0

0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 BITs (total per year: left scale) Total BITs (cumulative: right scale)

DTTs (total per year: left scale) Total DTTs (cumulative: right scale)

Source: WIR 2005

appendix b Direct Investment Cumulative Outflows from OECD Countries 1990–2003, Millions of Us Dollars United States/Etats-Unis United Kingdom/Royaume-uni France BLEU Germany/Allemagne Netherlands/Pays-Bas Japan/Japon Canada Spain/Espagne Switzerland/Suisse Sweden/Suède Italy/Italie Korea/Corèe Denmark/Danemark Finland/Finlande Australia/Australie Norway/Norvege Austria/Autriche Portugal Ireland/Irelande Mexico/Mexique Hungary/Hongrie Greece/Grèce Turkey/Turquie New Zealand/Nouvelle Zèlande Iceland/Islande Czech Rep./Rèp. Tchèque Poland/Pologne Slovak Rep./Rèp. Slovaque 0

200000

400000

600000

800000

1000000

1200000

1400000

1600000

712 tom coupé, irina orlova and alexandre skiba

appendix c List of Countries Included in Sample Home countries

Host countries

Austria, Belgium, Denmark, Finland, France, Germany, Italy, Japan, the Republic of Korea, Netherlands, Poland, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States of America

Bulgaria, Czech Republic, Hungary, Poland, Romania, Russian Federation, Slovakia, Slovenia, Ukraine

appendix d First-stage F-statistics Instruments

BIT

ICT

IT

SS

Sums Outside treaties of hosts

36.21 23.55

186.86 7.99

132.10 6.63

41.95 31.87

the effect of tax and investment treaties on bilateral fdi flows 713

references Bergstrand, J. and P. Egger. (2005). “A-knowledge-and-physical-capital model of international trade, foreign direct investment, and outsourcing: part I developed countries,” University of Notre Dame Working Paper. Blonigen, B. and R. Davis (2000). “The effects of bilateral tax treaties on U.S. FDI activity,” NBER Working Paper No. 7929 (Cambridge, MA: NBER). —— (2002). “Do bilateral tax treaties promote foreign direct investment?” NBER Working Paper, No. 8834 (Cambridge, MA: NBER). —— (2004). “The effects of bilateral tax treaties on U.S. FDI activity,” International Tax and Public Finance, 11 (5), pp. 601–622. Davis, R. (2004). “Tax treaties and foreign direct investment: potential versus performance,” International Tax and Public Finance, 11 (5), pp. 775–802. Doernberg, R. (1997). International Taxation in a Nutshell (St. Paul: West Publishing). Globerman, S. and D. Shapiro (2002). “Global foreign direct investment flows: the role of governance infrastructure,” World Development, 30, pp. 1898–1919. Goryunov, D. (2004). “The effectiveness of FDI promotion in transition countries,” EERC M.A. Thesis. Hallward-Driemeier, M. (2003). “Do bilateral investment treaties attract FDI? Only a bit . . . and they could bite,” (Washington, D.C.: World Bank, DECRG). Louie, H. and D. Rousslang D. (2002). “Host country governance, tax treaties, and American direct investment abroad,” mimeo. Neumayer, E. and L. Spess (2005). “Do bilateral investment treaties increase foreign direct investment to developing countries?” World Development, 33, 10, pp. 1567–1585. Monterrey Consensus 2002, UN International Conference on Financing for Development, 18–22 March 2002, Monterrey. Salacuse, Jeswald and Nicholas Sullivan (2005). “Do BITs really work? An evaluation of bilateral investment treaties and their grand bargain,” Harvard International Law Journal, 46 (1), pp. 67–130. Tobin J. and S. Rose-Ackerman (2005). “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Yale Law School Center, Economics and Public Policy Research Paper No. 293. UNCTAD (1998). Bilateral Investment Treaties in the Mid-1990s (New York: United Nations). —— (2003—WIR03). World Investment Report 2003: FDI Policies for Development: National and International Perspectives (New York and Geneva: United Nations). —— (2005—WIR05). World Investment Report 2005: Transnational Corporations and the Internationalization of R&D (New York and Geneva: United Nations).

This page intentionally left blank

selected bibliography on bilateral investment treaties and double taxation treaties 1 lisa e. sachs This bibliography is of literature that focuses specifically on BITs and DTTs. The reprinted and original articles included in the present volume are not included in this bibliography.

bilateral investment treaties Banga, Rashmi (2003). “Impact of government policies and investment agreements on FDI inflows,” Indian Council for Research on International Economic Relations, Working Paper, No. 116. Bernardini, Piero (2001). “Investment protection under bilateral investment treaties and investment contracts,” Journal of World Investment, 2. —— (2007). “Nationality requirements under BITs and related case law,” in Ortino, Federico, Liberti, Lahra. et al., eds., Investment Treaty Law: Current Issues, 2 (London: British Institute of International and Comparative Law). Bubb, Ryan and Susan Rose-Ackerman (2007). “BITs and bargains: strategic aspects of bilateral and multilateral regulation of foreign investment,” International Review of Law and Economics, 2. Burke-White, William and Andreas von Staden (2008). “Investment protection in extraordinary times: the interpretation and application of non-precluded measures provisions in bilateral investment treaties,” Virginia Journal of International Law, 48. Burkhardt, Hans-Martin (1986). “Investment protection treaties: recent trends and prospects,” Aussenwirtschaft: Schweizerische Zeitschrift für internationale Wirtschaftsbeziehungen, 41 (Special Issue on ‘Promotion of Direct Investment in Developing Countries’). Büthe, Tim and Helen Milner (2008). “The politics of foreign direct investment into developing countries: increasing FDI through trade agreements?” American Journal of Political Science, 52. Chen, An (2006). “Should the four great safeguards in Sino-foreign BITs be hastily dismantled? Comments on provisions concerning dispute settlement in model U.S. and Canadian BITs,” Journal of World Investment & Trade, 7.

1. The editors thank Anne van Aaken, Andrea Bjorklund, Anne Hoffman, Mark Kantor, Abba Kolo, Lahra Liberti, Lars Markert, Pierre Poret, Jan Peter Sasse, Stephan Schill, Stephen M. Schwebel, Michael Waibel, Thomas Walde, Joerg Weber, and the contributors to this volume for their input into this bibliography.

716 lisa e. sachs Congyan, Cai (2007). “Change of the structure of international investment and the development of developing countries’ BIT practice—towards a third way of BIT practice,” Journal of World Investment & Trade, 8. Connolly, Kelley (2007). “Say what you mean: improved drafting resources as a means for increasing the consistency of interpretation of bilateral investment treaties,” Vanderbilt Journal of Transnational Law, 40. Desai, Mihir A. and Alberto Moel (2006). “Czech mate: expropriation and investor protection in a converging world,” ECGI Working Paper, No.62/2004. Desbordes, Rodolphe and Vincent Vicard (2007). “Foreign direct investment and bilateral investment treaties: an international political perspective,” Working Paper. Diehl, Alexandra N. (2008). “Tracing a success story, or, ‘the baby boom of BITs’: characteristics and particularities of the tight net of bilateral investment treaties existing today,” in Reinisch, August and Knahr, Christina, eds., International Investment Law in Context (Utrecht: Eleven Publishing). Dolzer, Rudolf and M. Stevens (1995). Bilateral Investment Treaties (The Hague: Martinus Nijhoff Publishers). Dolzer, Rudolf and Christoph Schreuer, Principles of International Investment Law (Oxford: Oxford University Press, 2008). Elkins, Zachary, Andrew Guzman, and Beth Simmons (2008). “Competing for capital: the diffusion of bilateral investment treaties, 1960–2000,” Illinois Law Review, 1. Faruque, Abdullah Al (2004). “Creating customary international law through bilateral investment treaties: a critical appraisal,” Indian Journal of International Law, 44. Fietta, Stephen (2005). “Most favoured nation treatment and dispute resolution under bilateral investment treaties: a turning point,” International Arbitration Law Review, 8. Franck, Susan D. (2005). “The nature and enforcement of investor rights under investment treaties: do investment treaties have a bright future?” U.C. Davis Journal of International Law and Policy, 12. Freyer, Dana H. et al. (1998). “Arbitration under bilateral investment treaties: an often overlooked tool,” Mealey’s International Arbitration Report. —— (2006). “Bilateral investment treaties and arbitration,” in Carbonneau, Thomas E. and Jeanette A. Jaeggi, eds., AAA Handbook on International Arbitration and ADR (New York: Juris Publishing). Füracker, Matthias (2006). “Relevance and structure of bilateral investment treaties— the German approach,” German Arbitration Journal. Gaffney, John and James Loftis (2007). “The ‘effective ordinary meaning’ of BITs and the jurisdiction of treaty-based tribunals to hear contract claims,” Journal of World Investment & Trade, 8. Gallagher, Norah and Laurence Shore (2004). “Bilateral investment treaties: options and drawbacks,” International Arbitration Law Review, 7. Ginsburg, Tom (2005). “International substitutes for domestic institutions: bilateral investment treaties and governance,” International Review of Law and Economics, 25. Greig, Robert T. et al. (2008). “How bilateral investment treaties can protect foreign investors in the Arab world or Arab investors abroad,” Journal of International Arbitration, 25. Guzman, Andrew (1998). “Why LDCs sign treaties that hurt them: explaining the popularity of bilateral investment treaties,” Virginia Journal of International Law, 38. Haftel, Yoram Z. (2008). “The Effect of U.S. BITs on FDI Inflows to Developing Countries: signaling or credible commitment?” Manuscript.

selected bibliography 717

Hamilton, Calvin A. (2005). “Trade and investment: foreign direct investment through bilateral and multilateral treaties,” New York International Law Review, 18. Hindelang, Steffen (2004). “Bilateral investment treaties, custom and a healthy investment climate: the question of whether BITs influence customary international law revisited,” Journal of World Investment & Trade, 5. Hoffmann, Anne K. (2007). “The investor’s right to waive access to protection under a bilateral investment treaty,” ICSID Review—Foreign Investment Law Journal, 22(1). Jandhyala, Srividya et al. (2006). “A bilateral analysis of bilateral investment treaties,” (Philadelphia, PA: University of Pennsylvania), Manuscript. Juillard, Patrick (1979). “Les conventions bilatérales d’investissement conclues par la France,” Journal du Droit International, 22(1). Karl, Joachim (1996). “The promotion and protection of German foreign investment abroad,” ICSID Review, 22(1). Kishoiyian, Bernard (1994). “The utility of bilateral investment treaties in the formulation of customary international law,” Northwestern Journal of International Law and Business, 14. Koremenos, Barbara (2007). “If only half of international agreements have dispute resolution provisions, which half needs explaining?” Journal of Legal Studies, 36. Kurtz, Jürgen (2004). “The MFN standard and foreign investment—an uneasy fit,” Journal of World Investment and Trade, 5(6). Liebeskind, Jean-Christophe (2002). “State-investor dispute settlement clauses in Swiss bilateral investment treaties,” ASA Bulletin, 5(6). Mann, F.A. (1981). “British treaties for the promotion and protection of investments,” The British Yearbook of International Law, 52. Mayeda, Graham (2007). “Playing fair: the meaning of fair and equitable treatment in bilateral investment treaties,” Journal of World Trade, 41. Montt, Santiago (2007). “The BIT generation’s emergence as a collective action problem: prisoner’s dilemma or network effects?” Latin American and Caribbean Law and Economics Association Annual Papers, No. 043007–3 (Berkeley, CA: University of California). —— (forthcoming). State Liability in the BIT Generation (Oxford: Hart Publishing). Mosoti, Victor (2005). “Bilateral investment treaties and the possibility of a multilateral framework on investment at the WTO: are poor economies caught in between?” Northwestern Journal of International Law and Business, 26. Muchlinski, Peter T. (2000). “The rise and fall of the multilateral agreement on investment: where now?” The International Lawyer, 34. —— (2007). Multinational Enterprises and the Law, 2nd ed. (Oxford: Oxford University Press). Neumayer, Eric (2006). “Self-interest, foreign need and good governance: are bilateral investment treaty programs similar to aid allocation?” Foreign Policy Analysis, 2. Ocran, T. Mobido (1987). “Bilateral investment protection treaties: a comparative study,” New York Law School Journal of International and Comparative Law, 8. Organization for Economic Co-operation and Development (OECD) (2008). “Definition of investor and investment in international investment agreements,” (prepared by Lahra Liberti and Catherine Yannaca-Small), in OECD, International Investment Law: Understanding Concepts and Tracking Innovations (Paris: OECD). —— (2008). “Interpretation of the umbrella clause in investment agreements,” (prepared by Catherine Yannaca-Small), in OECD, International Investment Law: Understanding Concepts and Tracking Innovations (Paris: OECD).

718 lisa e. sachs —— (2008). “International investment agreements: a survey of environmental, labour and anti-corruption issues,” (prepared by Kathryn Gordon and Monica Bose), in OECD, International Investment Law: Understanding Concepts and Tracking Innovations (Paris: OECD). —— (2006). “Novel features in recent OECD bilateral investment treaties,” (prepared by Marie-France Houde), in OECD, International Investment Perspectives (Paris: OECD). —— (2005). “ ‘Indirect expropriation’ and the ‘right to regulate’ in international investment law,” (prepared by Catherine Yannaca-Small), in OECD, International Investment Law: A Changing Landscape (Paris: OECD). —— (2005). “Fair and equitable treatment standard in international investment law,” (prepared by Catherine Yannaca-Small), in OECD, International Investment Law: A Changing Landscape (Paris: OECD). —— (2005). “Most-favoured nation treatment in international investment law,” (prepared by Marie-France Houde and Fabrizio Pagani), in OECD, International Investment Law: A Changing Landscape (Paris: OECD). —— (2004). “Relationships between international investment agreements,” (prepared by Marie-France Houde and Catherine Yannaca-Small), in OECD, International Investment Perspectives (Paris: OECD). Parra, Antonio R. (2000). “ICSID and bilateral investment treaties,” News from ICSID, 52. Pekar, Rostislav and Ondre Sekanina (2007). “China’s bilateral investment treaties with EU member states: gaining a competitive advantage through investment protection,” China Law & Practice, 21(7), pp. 31–33. Peters, Paul (1998). “Some serendipitous findings in BIT’s: the Barcelona Traction Case and the reach of bilateral investment treaties,” in E. Denters and N. Schrijver, eds., Reflections on International Law from the Low Countries (The Hague: Kluwer). —— (1991). “Dispute settlement arrangements in investment treaties,” Netherlands Yearbook of International Law, 22. Peterson, Luke E. (2004). Bilateral Investment Treaties and Development Policy-Making (Winnipeg: International Institute for Sustainable Development). Pilch, Gennady (1992). “The development and expansion of bilateral investment treaties,” American Society of International Law Proceedings, 86. Radi, Yannick (2007). “The application of the most-favoured-nation clause to the dispute settlement provisions of bilateral investment treaties: domesticating the ‘trojan horse,’” European Journal of International Law, 18, p. 757. Reinisch, August, ed. (2008). Standards of Investment Protection (Oxford: Oxford University Press). Robbins, Joshua (2006). “The emergence of positive obligations in bilateral investment treaties,” University of Miami International and Comparative Law Review, 13, p. 403. Sacerdoti, Giorgio (1998). “Bilateral treaties and multilateral instruments on investment protection,” Receuil des Cours, 21(7), p. 251. —— (2000). “The admission and treatment of foreign investment under recent bilateral and regional treaties,” Journal of World Investment, 1, p. 105. —— (2004). “Has the proliferation of BITs gone too far? Is it now time for a multilateral investment treaty?” The Journal of World Investment & Trade, 5(1), p. 97. —— (2008). “The proliferation of BITs: conflicts of treaties, proceedings and awards,” in Karl P. Sauvant and Michael Chiswick-Patterson, eds., Appeals Mechanism in International Investment Disputes (Oxford: Oxford University Press).

selected bibliography 719

Salacuse, Jeswald (1990). “BIT by BIT: the growth of bilateral investment treaties and their impact on foreign investment in developing countries,” The International Lawyer, 24, pp. 655–675. —— (2007). “Is there a better way? Alternative methods of treaty-based, investor-state dispute resolution,” Fordham International Law Journal, 31, p. 138. —— (2007). “The treatification of international investment law,” Law and Business Review of the Americas, 13, p. 155. Schill, Stephan W. (2007). “Tearing down the Great Wall: the new generation investment treaties of the People’s Republic of China,” Cardozo Journal of International and Comparative Law, 15, p. 73. —— (2007). “Fair and equitable treatment as an embodiment of the rule of law,” in R. Hofmann and C. Tams, eds., The International Convention for the Settlement of Investment Disputes (ICSID): Taking Stock After 40 Years, p. 31. —— (2007). “Do investment treaties chill unilateral state regulation to mitigate climate change?” Journal of International Arbitration, 24(5), p. 469. Schlemmer, E.C., (2001). “Investor protection and forum selection in bilateral investment treaties,” South African Yearbook of International Law, 5(1), p. 174. —— (2004). “Foreign direct investment and umbrella clauses under BITs,” South African Yearbook of International Law, 29, p. 255. Schwebel, Stephen M. (2004). “Investor-state disputes and the development of international law: the influence of bilateral investment treaties on customary international law,” Proceedings of the annual meeting—American Society of International Law, 98, pp. 27–30. —— (2005). “The reshaping of the international law of foreign investment by concordant bilateral investment treaties,” in Steve Charnovitz, Debra P. Steger, and Peter van den Bossche, eds., Law in the Service of Human Dignity: Essays in Honour of Florentino Feliciano (Cambridge: Cambridge University Press). —— (2005). “The United States 2004 model bilateral investment treaty: an exercise in the regressive development of international law,” in Gerald Aksen, et al. eds., Global Reflections on International Law, Commerce and Dispute Resolution: Liber amicorum in Honour of Robert Briner (Paris: International Chamber of Commerce). Schreuer, Christoph (2004). “Travelling the BIT route: of waiting periods, umbrella clauses and forks in the road,” The Journal of World Investment & Trade, 5(2), p. 231. Siqueiros, Jose Luis (1994). “Bilateral treaties on the reciprocal protection of foreign investment,” California Western Iinternational Law Journal, 24, p. 255. Sornarajah, M. (1986). “State responsibility and bilateral investment treaties,” Journal of World Trade Law, 20, p. 79. Srur, Muradu A. (2004). “The international investment regime: towards evolutionary bilateral and regional investment treaties?” Manchester Journal of International Economic Law, 1, p. 54. Suda, Ryan (2006). “The effect of bilateral investment treaties on human rights enforcement and realization,” in Olivier De Schutter, ed., Transnational Corporations and Human Rights (Oxford: Hart Publishing). Tobin, J. and Susan Rose-Ackerman (2005). “Foreign direct investment and the business environment in developing countries: the impact of bilateral investment treaties,” Yale Law School, Economics and Public Policy Research Paper No. 293. —— (2008). “When BITs have some bite: the political economic environment for bilateral investment treaties,” (New Haven: Yale University).

720 lisa e. sachs UNCTAD (1998). Bilateral Investment Treaties in the Mid-1990s (New York and Geneva: United Nations). —— (2001). Bilateral Investment Treaties 1959–1999 (New York and Geneva: United Nations). —— (2005). Bilateral Investment Treaties 1995–2005: Trends in Investment Rulemaking (New York and Geneva: United Nations). Unegbu, O. Carl (1999). “BITs and ICC arbitration: portent of a new wave?” Journal of International Arbitration, 16, p. 93. United Nations Centre on Transnational Corporations (1988). Bilateral Investment Treaties (London and Boston: Published in co-operation with the United Nations by Graham & Trotman). Vandevelde, Kenneth J. (1996). “Arbitration provisions in the BITs and the Energy Charter Treaty,” in Thomas W. Wälde, ed., The Energy Charter Treaty: Aan East–West Gateway for Investment and Trade (London and Boston: Kluwer Law International). —— (1998). “The political economy of a bilateral investment treaty,” American Journal of International Law, 92, p. 621. —— (1998). “Investment liberalization and economic development: the role of bilateral investment treaties,” Columbia Journal of Transnational Law, 36, p. 501. —— (2000). “The economics of bilateral investment treaties,” Harvard International Law Journal, 41, p. 469. Vesel, Scott (2007). “Clearing a path through a tangled jurisprudence: most-favorednation clauses and dispute settlement provisions in bilateral investment treaties,” Yale Journal of International Law, 32, p. 125. Wong, Jarrod (2006). “Umbrella clauses in bilateral investment treaties: of breaches of contract, treaty violations, and the divide between developing and developed countries in foreign investment disputes,” George Mason Law Review, 14, p. 135. —— (2008). “The application of most-favored-nation clauses to dispute resolution provisions in bilateral investment treaties,” Asian Journal of WTO & International Health Law and Policy, 3, p. 171. Yackee, Jason Webb (2005). “Are BITs such a bright idea? Exploring the ideational basis of investment treaty enthusiasm,” U.C. Davis Journal of International Law and Policy, 12, p. 195. —— (forthcoming, 2008). “Conceptual difficulties in the empirical study of bilateral investment treaties,” Brooklyn Journal of International Law.

double taxation treaties Altman, Zvi Daniel (2006). Dispute Resolution Under Tax Treaties (Amsterdam: IBFD Publications). Arnold, Brian (2004). “Tax treaties and tax avoidance: the 2003 revisions to the commentary to the OECD model,” Bulletin for International Fiscal Documentation 58, p. 244. —— with J. Sasseville and E.M. Zolt (2002). “Summary of the proceedings of an invitational seminar on tax treaties in the 21st century,” Bulletin for International Fiscal Documentation, 56, p. 233. Ault, H.J. (1994). “The role of the OECD commentaries in the interpretation of tax treaties,” Intertax, 22(4), p. 144.

selected bibliography 721

Avery Johnes, J. F. et al. (1984). “The interpretation of tax treaties with particular reference to Article 3(2) of the OECD Model,” British Tax Review, 1, pp. 14–54 and pp. 90–108. —— (1993). “Article 3(2) of the OECD Model Convention and the commentary to it: treaty interpretation,” European Taxation, 33(8), p. 252. —— (2001). “The ‘one true meaning’ of a tax treaty,” Bulletin for International Fiscal Documentation, 55(6), p. 220. —— (2002). “The effect of changes in the OECD commentaries after a treaty is concluded,” Bulletin for International Fiscal Documentation, 56(3), p. 102. Avi-Yonah, Reuven S. (2007). International Tax as International Law: An Analysis of the International Tax Regime (New York: Cambridge University Press). Baistrocchi, Eduardo (forthcoming, 2008). “Tax treaty interpretation in the emerging world: theory and implications,” British Tax Review. —— (2007).“The structure of the asymmetric tax treaty network: theory and implications,” Bepress Legal Series, Working Paper, No. 1991. Baker, Philip (2001). Double Taxation Conventions: A Manual on the OECD Model Tax Convention on Income and on Capital (London: Sweet & Maxwell). Becker, Helmut and Felix Wurm (1988). Treaty Shopping: An Emerging Tax Issue and its Present Status in Various Countries (Deventer and Boston: Kluwer Law and Taxation). Borrego, Vega and Félix Alberto (2006). Limitation on Benefits Clauses in Double Ttaxation Conventions (The Hague: Kluwer Law International). Chisik, R. and R.B. Davies (2004). “Asymmetric FDI and tax-treaty bargaining: theory and evidence,” Journal of Public Economics, 88, p. 1119. —— (2004). “Gradualism in tax treaties with irreversible foreign direct investment,” International Economic Review, 45, p. 113. Dagan, Tsilly (2000). “The tax treaties myth,” NYU Journal of International Law and Politics, 32, p. 939. Davies, R.B (2003). “Tax treaties, renegotiations, and foreign direct investment,” Economic Analysis and Policy, 32, p. 251. —— (2003). “The OECD Model Tax Treaty: tax competition and two-way capital flows,” International Economic Review, 44, p. 725. —— (2004). “Tax treaties and foreign direct investment: potential versus performance,” International Tax and Public Finance, 11, p. 775. —— with P. Norbäck and A. Tekin-Koru (2007). “The effect of tax treaties on multinational firms: new evidence from microdata,” World Economy. De Mooij, R.A. and S. Ederveen (2003). “Taxation and foreign direct investment: a synthesis of empirical research,” International and Tax Public Finance, 10, p. 673. Diamond, Walter and Dorothy Diamond (1998). International Tax Treaties of All Nations (New York: Ocean Publications). Doernberg, R. (1997). International Taxation in a Nutshell, 3rd ed. (St. Paul: West Publishing). Easson, Alex (1999). Taxation of Foreign Direct Investment: An Introduction (The Hague: Kluwer Law International). —— (2000). “Do we still need tax treaties?” Bulletin for International Fiscal Documentation, 54, p. 619. Edmiston, K, et al. (2003). “Tax structures and FDI: the deterrent effects of complexity and uncertainty,” Fiscal Studies, 24, p. 341.

722 lisa e. sachs Ellis, M. J. (2000). “The influence of the OECD commentaries on treaty interpretation— response to Prof. Dr. Klaus Vogel,” Bulletin for international fiscal documentation, 54(12), p. 617. —— (2006). “The role of the commentaries on the OECD Model in the tax treaty interpretation process—response to David Ward,” Bulletin for International Taxation, 60(3), p. 103. Engelen, F. A. (2004). Interpretation of Tax Treaties under International Law (Amsterdam: IBFD). —— (2006). “Some observations on the legal status of the Commentaries on the OECD Model,” Bulletin for International Taxation, 60(3), p. 105. Figueroa, A.H. (1992). “Comprehensive tax treaties,” in Double Taxation Treaties Between Industrialized and Developing countries: OECD and UN Models—A Comparison, Proceedings of a Seminar held in Stockholm in 1990 during the 44th Congress of the International Fiscal Association (Deventer: Kluwer Law and Taxation Publishers). Fitzgerald, V. (2002). “International tax co-operation and capital mobility,” Oxford Development Studies, 30, 251. Gifford, W.C. (1978). “Permanent establishments under the nondiscrimination clause in income tax treaties,” 11 Cornell International Law Journal, 11, p. 51. Gordon, R. and J.R. Hines (2002). “International taxation,” in A.J. Auerbach and M. Feldstein, eds., Handbook of Public Economics (Amsterdam: Elsevier). Gravelle, Pierre (1988). “Tax treaties: concepts, objectives and types,” International Bureau of Fiscal Documentation Bulletin, 42, p. 522. Gresik, Thomas A. (2001). “The taxing task of taxing transnationals,” Journal of Economic Literature, 39, 800. Haug, Simone M. (1996). “The United States policy of stringent anti-treaty-shopping provisions: a comparative analysis,” Vanderbilt Journal of Transnational Law, 29, p. 191. Hines Jr., James R. and Kristen L. Willard (1992). “Trick or treaty? Bargains and surprises in international tax agreements,” Manuscript. —— (2001). “ ‘Tax sparing’ and direct investment in developing countries,” in J.R. Hines Jr., ed., International Taxation and Multinational Activity (Chicago: University of Chicago Press). Holmes, Kevin (2007). International Tax Policy and Double Tax Treaties: An Introduction to Principles and Application (Amsterdam: IBFD). Janeba, Eckhard (1995). “Corporate income tax competition, double taxation treaties, and foreign direct investment,” Journal of Public Economics, 56, p. 311. Jones, John F. (1999). “Are tax treaties necessary,” Tax Law Review, 53, p. 1. Kulle, Kristina (2004). “Avoidance & limitation: bi-lateral treaties and double taxation,” New England Journal of International and Comparative Law, 11, p. 89. Lang, Michael (1998). Multilateral Tax Treaties: New Developments in International Tax Law (London and Boston: Kluwer Law International). —— (2001). Tax Treaty Interpretation (The Hague and Boston: Kluwer Law International). Lang, Michael and Mario Züger (2002). Settlement of Disputes In Tax Treaty Law (The Hague and New York: Kluwer Law International). O’Brien, J.G. (1978).“The nondiscrimination article in tax treaties,” Law and Policy in International Business, 10, p. 545. Ocran, T. Mobido (1989). “Double taxation treaties and transnational investment: a comparative study,” Transnational Lawyer, 2, p. 131.

selected bibliography 723

Organisation for Economic Co-operation and Development (1994). Tax Information Exchange Between OECD Member Countries: A Survey of Current Practices (Paris: OECD Committee on Fiscal Affairs). —— (forthcoming, 2008). Model Tax Convention on Income and on Capital (Paris: OECD Committee on Fiscal Affairs). Panayi, Christiana (2007). Double Taxation, Tax Treaties, Treaty-Shopping and the European Community (The Netherlands: Kluwer Law International). Passos, Adelaide (1986). Tax Treaty Law (Kenwyn, South Africa: Juta in association with Divaris Stein Publishers). Reece, Sharon A. (1992). “Arbitration in income tax treaties: ‘to be or not to be,’” Florida Journal of International Law, 7, p. 277. Reese, P.D. (1987). “United States tax treaty policy toward developing countries,” UCLA Law Review, 35, p. 369. Rohatgi, R. (2005). Basic International Taxation, vol. 1 (Richmond: Richmond Law & Tax). —— (2007). Basic International Taxation, vol. 2 (London: BNA International). Roin, Julie (1995). “Rethinking tax treaties in a strategic world with disparate tax systems,” Virginia Law Review, 81, p. 1753. Shelton, Ned (2004). Interpretation and Application of Tax Treaties (London and Dayton, Ohio: LexisNexis). Skaar, Arvid Aage (1991). Permanent Establishment: Erosion of a Tax Treaty Principle (Deventer; Boston: Kluwer Law and Taxation Publishers). Smith, Robert T. (1996). “Tax treaty interpretation by the judiciary,” Tax Lawyer, 49, p. 845. Srinivasan, K and B.P. Bhargava (1992). Guide to Double Taxation Avoidance Agreements (New Delhi: Vidhi Foundation). Tillinghast, David (2003). “Arbitration of disputes under income tax treaties: a panel overview” Proceedings of the annual meeting- American Society of International Law, 97, p. 107. Townsend, John A. (2001). “Tax treaty interpretation,” Tax Lawyer, 55, p. 219. Trelles, Oscar M. (1979). “Double taxation/fiscal evasion and international tax treaties,” Indiana Law Review, 12, p. 341. UNCTAD (2000). Taxation. UNCTAD Series on Issues in International Investment Agreements (New York and Geneva: United Nations). United Nations (2001). United Nations Model Double Taxation Convention between Developed and Developing Countries (New York: United Nations). Vann, Richard, ed. (1996). Tax Treaties: Linkages Between OECD Member Countries and Dynamic Non-Member-Economies (Paris: OECD). Vogel, K. (1997). Double Taxation Conventions—A Commentary to the OECD, UN and US Model Conventions for the Avoidance of Double Taxation on Income and Capital with Particular Reference to German Treaty Practice, 3rd ed. (London: Kluwer Law International). —— (2000). “The influence of the OECD commentaries on treaty interpretation,” Bulletin for international fiscal documentation, 54(12), p. 612. Wälde, Thomas and Abba Kolo (2007). “Investor-state disputes: the interface between treaty-based international investment protection and fiscal sovereignty,” Intertax, 35, p. 424. —— (2008). “Economic crisis, capital transfer restrictions and investment protection under modern investment treaties,” Capital Market Law Journal, 3(2), p. 154.

724 lisa e. sachs Ward, David A. (1993). “Abuse of tax treaties,” in H. Alpert and Kees van Raad, eds., Essays on International taxation (Deventer; Boston: Kluwer Law and Taxation Publishers). —— (2005). The Interpretation of Income Tax Treaties with Particular Reference to the Commentaries on the OECD Model (Kingston, Ont.: International Fiscal Association). —— (2006). “The role of the commentaries on the OECD Model in the tax treaty interpretation process,” Bulletin for international taxation, 60(3), p. 97. Wattel, P. J. and O. Marres (2003). “The legal status of the OECD commentary and static or ambulatory interpretation of tax treaties,” European Taxation, 43(7/8), p. 222.

index A AALCC. See Asian-African Legal Consultative Committee Agreement on Trade Related Investment Measures, 20–21, 231 Andean Free Trade Agreement, 297 Arbitration ad hoc, 66, 131–132 binding, provisions, 74, 91,183, 185–187, 212, 353, 380 and BIT clauses and provisions, 18, 28, 172, 210, 312, 381 BIT critics and, 155 BITs as customary law and, 157 choice between ICSID and, 67 compulsory, 131, 188 consent to international, 56 costly, 29, 313 direct investor-state dispute mechanisms and, 395 extra national, 398–399 ICSID, 19, 68, 131–132, 175, 185, 187, 353, 380–381, 398 in Latin American countries, 6 non-binding, 104 OECD model of, 104 panels, 183–184, 188, 205, 210 securing international, 58, 112 under investor-state treaty, 30, 32, 398, 423 under NAFTA, 48 under UNCITRAL rules, 48–49, 66, 132, 398 unfair and inequitable treatment in, 48–49 Asian-African Legal Consultative Committee, 47 model “A” BIT, 60–62 model “B” BIT, 62 model treaties, 47

B Bilateral investment treaties, 3 alternatives to, 382–383

arbitration under, 381 associated with large increases in foreign investment, 440 basic structure of, 123 binding investor-to-state dispute settlement provisions under, 230 commitment to investment protection of, 15 compensation for expropriation under, 61–64 compensation for losses due to armed conflict or internal disorder under, 61 compliments to good institutional quality, 226 conditions to stimulate foreign direct investment and, 311–312 content of, 37–71 core elements, 296 cross sectional analysis of FDI inflows under, 233 customary international law and, 156–158 defined, 109–110, 352–355 design of, 253 in developed countries, 13 discriminatory treatment against foreign investors in, 230 disputes between the contracting parties under, 66–67 dispute settlement clauses and procedures contained in, 65, 439 dominant vehicle through which investment is regulated, 73–75 early, 67 econometric research study of, 140–143 effective resolution of credible commitment problems and, 382 effect of, on FDI location, 225–227 effect of, on foreign direct investment, 188–190 efficiency implications of, 91–93 evaluation of, 109–110 ex ante information and, 184

726 index Bilateral investment treaties, 3 (cont.) ex post enforcement and, 186–188 ex post information and, 185–186 failure to negotiate a multilateral agreement on investment and, 317–320 FDI related issues regulated by, 253 first, 253 and foreign direct investment increases to developing countries, 225–252 free trade agreements, 297 free transfer of payments under, 59–61 general principles of law of, 158–159 government policies and, 231 history of, 317–320 impact of, 75, 93–96, 147–155, 253–272 impact of, on customary international law, 75, 93–96 impact of, on general nondiatic FDI inflows, 232–233 impact on foreign direct investment flows, 109–394 importance of institutions and property rights and, 356–359 important instrument of protection to foreign investors, 231 incorporation of international dispute resolution mechanisms for, 395–596 increased global efficiency, 75 international arbitration and, 230–231 international arbitration under, 19 international investment insurance and, 382–383 international legal instrument, 109 investment contracts and, 382 investment promotion under, 138–140 inward flow of FDI and, 225–227 lack of investor awareness of, 381–383 large FDI inflows and, 225–227 marginal impact of, 311–312 method of governing relationship between investors and governments, 74–75 model, 30 modern, 115 as most common type of international investment agreement, 37 most economically important aspect of, 395

most favored nation treatment under, 17 movement, 110–113 national treatment under, 17 network of, 15–16 objectives of, 109 observance of obligations under, 54–59 popularity of, 73–97, 228 potential endogeneity of, 396–397 as a potent tool for attracting FDI, 311–312 primary purpose of, 114–115 prior United Nations studies of, 143–147 programs, 14 property rights and, 352–355 protected tariffs of tax incentives for foreign investors and, 231 for protecting and influencing foreign investment, 109 protections provided by, 16 provisions on common set of investment issues of, 438–443 purpose of, 225–227 rise to prominence of the, 74–75 risk of investment and, 313–314 role of, in promoting FDI, 227 signaling effect of, 312–313 signing, 439–457 specific standards of treatment under, 59–75 studies of, location of FDI, 232–234 trends in, 355–356 uniformity across countries, 296 welfare implication of, 93 Bilateral investment treaties, core elements of, 296 Bilateral investment treaties and foreign direct investment alternative estimation techniques for statistical analysis of, 200–202 causation between, 395–398 correlation between, 395–398 empirical findings from monadic analysis concerning, 179–180 empirical models of, 176–178 political analysis of, 171– 224 political theory of, 180–184 statistical analysis of, 190–196 statistical findings of, 196–199

index 727

studies on the effect of, on foreign direct investment, 173–176 Bilateral investment treaties as means of attracting foreign investment as a commitment device, 350–352 studies on growth and investment relating to, 351–352 Bilateral investment treaty movement development of the, 114 evolution of the, 113–118 goals of the, 118–122 history of the, 110–113 impetus for the, 110 in the late 1980s. 117 new face in the history of the, 117–118 Bilateral investment treaty regime, 60, 88–89, 91–93, 120, 210, 311–321 global, 27, 311–321 rise of the, 312, 317–320 Bilateral investment, treaty signing and analysis exploring the economic effects of, 440–443 countries, 443–448 developing countries and, 440–443 economic determinants including FDI stocks and flows of, 449–450 economic determinants of, 444–448 economic evidence on value of, 452–453 the effect of country identity and, 451 effect on FDI flows, 453–455 effects of, 440–443 empirical results of, 444–448 identifying economic effects of, 450 interpretation of further issues on, 455–457 investment liberalization and, 441–442 leads and lags in investment reactions and, 451 values of, 452–455 Bilateral investment treaty structure compensation for losses from armed conflict or internal disorder, 129–130 conditions for the entry of foreign investments, 126 general standards of treatment of foreign investment, 126–128 investment dispute settlement 130–133 monetary transfers, 128–129

operational conditions, 129 protection of dispossession, 130 scope of application, 123–125 Bilateral tax treaties, 461–466 Bilateral treaties, 4 availability of, as commitment mechanism, 318 bilateral flows based on signed, 359 Bilateral Treaties of Friendship, Commerce and Navigation, 4, 7 colonial era, 8 investment protection for, 10 modern, 17 most favored nation treatment and, 17 national treatment and, 17 postwar, 8–10 submission of disputes to an ad hoc arbitral tribunal under, 17 Bilateral treaty making, 115–117 BIT. See Bilateral investment treaties

C Calvo Doctrine defined, 5 end of, 22 Latin America and, 5 Central American Free Trade Agreements (CAFTA), 297 Charter of Economic Rights and Duties of States (CERDS), 12 adoption of, 14 creation, 12 expropriation issues and, 14 regime, 87–88, 92–93 Credible commitments ability of political institutions to make, 236, 670–671 to attract more foreign direct investment, 85, 215, 229, 289 bilateral investment agreements as, 206–207, 211–213, 318, 380–382, 389 contractual, 90 cost of, 91 customary law and, 91 dynamic inconsistency problem and, 91 lack of, 92 liberal economic policies defined as, 182 logic of, 176 mechanism, 82, 92, 206

728 index Credible commitments (cont’d...) problems with, 217, 382 to property rights, 314 risk premium of, 382 Customary international law, 5, 30–31 Calvo clause in, 5 deficiencies of, 13 under the Hull Rule, 75 lack of consensus on, 118–119 as a means of protecting international investment, 13–14

D Developing countries aggregate FDI flows to, 160–161 arbitration in, 155–156 BITs as investment vehicles that promote FDI in, 225–250 BITs concluded by, 356–357 capital importing, 26–27, 689–690 competing for FDI, 350–352, 437–443 credible commitment by, 312 DTTs and, 659–679 decolonized, 111 economic determinants of BIT signing by, 445–449 effect of BITs on inward FDI flows to, 197 expropriation, privatization and harassing of FDI issues in, 130, 134, 312, 318 FDI as a percentage of gross capital formation in, 441 FDI flows towards, 21, 73–74,147–148, 165, 169, 175–176, 313, 359, 440, 687–692 FDI in and among, 357, 362, 441 FDI inward flows to. See Developing countries, FDI flows towards FDI regime and, 320–321 growth of BITs and FDI in, 118, 439 increased use of BITs between, 71 integration agreements between, 52 liberalization of, 119, 133, 136 literature and studies on, 356–358 national or most favored nation treatment in, 128 OECD countries and, 177, 317, 355, 359 368, 374, 376, 692

popularity of BITs in, 313 private market for credit and, 21–22 property rights in, 175, 362 protection of FDI in, 176 rule of law in, 172, 177 share of FDI to, 359–360 socialist countries and, 12 sovereignty issues and investment in, 29–30, 112–113 statistical analysis of FDI flows to, 190–217 tax treaties with the U.S. signed by, 616 trade between developed countries and, 36 U.S. FDI flows to, 141, 177–178, 204, 692 Developed countries, 13 capital exporting, 120, 232, 679–680 DDTs and, 572, 662–666, 669, 679, 615, 626 development aid from, 227, 661 as dominant source and major recipient of FDI, 227, 661 dyadic sample of, 667 FDI and, 194 investment treaties signed between, 118, 120 liberal market model in, 119 OECD model treaty and, 602 tax issues in, 580, 582, 604, 606, 609, 616, 632–633 total number of BITs in force in, 313 Dispute resolution fora (forum), 172 international, 395 investor-state, 23, 29, 395–396 mechanisms, 315, 350, 353, 395–398 procedures, 353 provisions in FCNs, 9–10 WTO system of, 64–65 Doha Round of Multilateral Trade Negotiations, 34 declaration, 442 investment agreements and the, 34, 442 negotiation of a MIA during the, 34, 155–156, 442 statement, 20, 442 Double taxation treaties defined, 690

index 729

developing countries and, 659, 679–680 exploring the impact or effect of, on FDI flows, 461–538, 659–660, 665–667 and FDI, 660–665 growth of, 660 introduction, 99–106 models, 99–101 network, 99, 101,104, 660 number of, concluded 1990–2004, 711 number of, with OECD countries, 681, 710 OECD model of, 99–101 statistical analysis of, 667–679, 690 timing of, effect of FDI flows, 708 United Nations model, 99–101 web, 99, 101,104, 660 Dynamic inconsistency problem defined, 78 developing nations attempting to attract FDI and, 230 efficiency and, 82–83 and foreign direct investment, 78–79 global perspective of, 82 host country and, 82 impact of, 82 reputational effect of, 82

E Espousal, 19 as a customary law mechanism, 5 defined, 5 as a diplomatic process, 6 as a dispute resolution mechanism, 112 procedure, 5–6 satisfactory resolution of, 5–6 Economic integration, 35 agreements, 25 in the past, 26 regional and interregional, 26 Effect of bilateral investment treaties on foreign direct investment. See Impact of bilateral investment treaties on foreign direct investment Effect of bilateral tax treaties. See impact of bilateral investment treaties on foreign direct investment. Empirical methodology and data to estimate the effect of BITs, 491–496

Empirical link between investment treaties and foreign direct investment econometric analysis of the, 379–383 revisiting the 379–392 Endogenous tax treaties, 513–516

F FCN. See Bilateral Treaties of Friendship, Commerce and Navigation Foreign direct investment (FDI) balancing financial accounts and, 295 bilateral outward stocks, 258 Central and Eastern Europe, 273–391 dynamic inconsistency and, 78–79 economic benefits, 437 effect of taxation on, 461–466 empirical studies of the effects of taxation on, 485 flows, 73–74 global flows of, 73–74 growth and growth rates, 395 how treaties affect, 467–469 impact of BITs on, 312–314 impact of endogenous tax treaties on, 513–516 key determinants of, 298–302 location, 273–291 negotiating BITs as a strategy to attract, 287 nominal outward stocks, 258 outward stocks, 255, 258 protection of, 4–7 regulations on, 287 response to tax incentives of, 461–466 studies concerning extent of FDI response to tax incentives, 461–466 vertical, 254–256 Foreign direct investment flows global, 273 inward, 273–276, 311–312 Foreign investment alternatives to, 21 developing countries, 11–12 form of neocolonialism, 11 Germany and, 13 liberalization of, 3 political risks of, 14 protection of, 3 socialist block and the Soviet Union, 11

730 index Fourth Ministerial Meeting, Doha, 2001. See Doha Round Free Trade Agreements, 32

G GATS. See General Agreement on Trade in Services GATT. See General Agreement on Tariffs in Trade General Agreement on Tariffs and Trade 6, 10 framework, 6 as a trade negotiations forum, 6 trade relations and, 6 General Agreement on Trade in Services 20 Good neighbor policy, 6. See also Roosevelt Administration

H Havana Charter, 114. See also International Trade Organization Hull Rule customary international law and, 75 fall of,76–78 international community and, 79 international law and, 79 opposition to, 74–75 protection for foreign investors and, 228

I International investment agreement, 3, 37 changes in, 19 colonial era, 3–6 content of, 34 covered persons and entities under, 42–45 definition of investments covered by, 39–40 divorced from the concept of liberalization, 35 fair and equitable treatment under, 46–52 first historical era of, 3–6 framework of, 28 general standards of treatment in, 46 global era, 19–28 history of, 3–35

increasingly universal, 35 major and most common provisions in, 37–38 postcolonial era, 3, 7–19 preambles to, 38 principle provisions of, 37–38 as a protection against political risk, 35 provisions defining the scope of application of the treaty, 39–46 standards of treatment to be applied to, 46 subject matter covered in, 39–42 temporal application, 45–46 territorial scope of, 45 ICRG. See International Country Risk Guide ICSID. See International Centre for Settlement of Investment Disputes, 19 IIA. See International investment agreement IMF. See International Monetary Fund Impact of bilateral investment treaties on FDI, 485–487 analysis of time series data to determine the, 326–331 caveats concerning the, 323–324 change in the exchange rate as a variable in determining the, 333 country risk as a variable in determining the, 334 cross sectional analysis to determine the, 331–337 evidence regarding the, 396–398 inflation as a variable in determining the, 333 literature on, 253–355 market size and growth as a variable in determining the, 332–333 other variable in determining the, 334–335 rate of capital formation as variable in determining the, 333–334 statistical analysis of the, 323, 326–348 theoretical considerations surrounding the, 324–326 Intellectual property rights, 21 International Centre for Settlement of Investment Disputes, 19 International Country Risk Guide (as an index of political risk), 315

index 731

International Court of Justice Barcelona traction case and the, 110–113 submission of disputes to, 17 International investment colonial era, 4 protection of, 4 regime, 6 International investment law bilateral investment treaties and, 109–110 historical development of, 115 modern, 115 transformation of, 109–110 International investment regime, 35 colonial era, 6 features of, 6 postcolonial era, 10 trade and property, 6 treaty-based, 32–33 virtually universal, 35 International law, 91–96 International legal framework, 115 International Monetary Fund, 60 International Trade Organization, 114 International trade relations, 6 Investment protection, 19, 37–71 Investor-host conflicts, 75 Investor protection regime. See Regime, Investor protection

K Key determinants of foreign direct investment analysis of, 323–326 in developing countries, 297–303 difference between sub-regions, 308 econometric model to identify, 303–305 factor endowments as one of, 298 factors relating to geography as one of, 299–300 factors relating to infrastructure as one of, 299–300 importance of, 323 in Latin America and the Caribbean, 305–308 literature on, 297–302 macroeconomic stability of one of, 298–299 market size as one of, 298

policy environment as one of, 299 political stability as one of, 299 privatization as one of, 302 substantial differences in, 308 value of BITs and, 314

L Least Developed Countries (LDC), 76–91 behavior of the, 83–85 collective interests of the, 86–88 conduct of the, 88–90 in early 1960s, 77 economic interpretation concerning, 90–91 interests of individual, 85–86 paradoxical behavior of the, 83–91 strategic analysis of, 85–88 Liberalization. See Market liberalization

M MAI. See Multilateral Agreement on Investment, 33 Market liberalization an absolutist evaluation of, 134–136 investment and, 133–134 a relativist evaluation of, 136–138 trend toward, 23 MFN. See Most favored nation Military Action. See Military force Military force, 19 to collect debts, 13 as a legitimate means of protecting country’s interests, 35 to protect investment, 13, 19 use of, 13 MNE. See Multinational enterprises Monroe Doctrine, 6 Most favored nation, 4 obligations, 30 treatment, 4, 16 Multilateral Agreement on Investment, 33 Multilateral agreements, 6 Multilateral investment agreement, 312, 317–320 Multinational enterprises, 70 horizontal, 254 institutional environment and, 275 investment strategy and, 275 vertical, 254

732 index N NAFTA. See North American Free Trade Agreement National treatment,16 New institutional economics and foreign direct investment bilateral investment treatments concerning, 278–279 corruption and, 277–278 currency evaluation and, 281–282 FDI control variables and, 280–287 literature on the, 276–277 market size and the, 282–283 political risks inherent to the, 280 privatization and the, 279–280 regulations on the, 279 rule of law and the, 283 statistical results concerning, 285–287 theory on, 276–280 New International Economic Order, 12 adoption of, 14 declaration of, 12 discussions of a, 22 NIEO. See New International Economic Order North American Free Trade Agreement, 25–28 arbitral decisions under the, 47 chapter eleven, 26, 231 interpretive notice under, 31 investment protection provisions under, 231 investment related provisions of, 26 section on investment, 117

O Office of Investment Affairs, Bureau of Economic Business Affairs, 253–254 Organisation for Economic Co-operation and Development (OECD), 114

P Postcolonial era. See International investment agreement, history of

Promotion of foreign direct investment effect of bilateral tax treaties on the, 469–474 empirical framework for statistical analysis of the, 467–469 statistical analysis and results of the, 474–481 Property protection provisions 6, 8 Protectionist policies, 6

R Regional trade agreements, 23

S Socialist countries, 12 Standard of expropriation, 15–16

T Tax treaties bilateral, 487–491 functions of, 487–491 history of, 487–491 list of United States, 489 primary functions of, 488–490 primary goal of, 485–486 process of creating, 487–491 Theoretical trade literature 254 Trade liberalization 6 Trade protection provisions 6, 8 Trade-Related Investment Measure (TRIM), 137–138 Transition countries, 266–267

U United Nations Conference on Trade and Development (UNCTAD), 254 U.S. bilateral investment treaty program basic aims, 253–254 fact sheet, 253

W World Trade Organization (WTO), 20