Value Creation in Multinational Enterprise, Volume 7 (International Finance Review) (International Finance Review)

VALUE CREATION IN MULTINATIONAL ENTERPRISE INTERNATIONAL FINANCE REVIEW Series Editor: J. Jay Choi Volume 1: Asian F...

Author: J. Jay Choi | Reid W. Click

28 downloads 1135 Views 4MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

VALUE CREATION IN MULTINATIONAL ENTERPRISE

INTERNATIONAL FINANCE REVIEW Series Editor: J. Jay Choi Volume 1:

Asian Financial Crisis: Financial, Structural and International Dimensions Edited by: J. J. Choi

Volume 2:

European Monetary Union and Capital Markets Edited by: J. J. Choi & J. Wrase

Volume 3: Global Risk Management: Financial, Operational & Insurance Strategies Edited by: J. J. Choi & M. Powers Volume 4:

The Japanese Finance: Corporate Finance and Capital Markets in Changing Japan Edited by: J. J. Choi & T. Hiraki

Volume 5:

Latin American Financial Markets: Developments in Financial Innovations Edited by: H. Arbela´ez & R. W. Click

Volume 6:

Emerging European Financial Markets: Independence and Integration Post-Enlargement Edited by: J. A. Batten & C. Kearney

INTERNATIONAL FINANCE REVIEW VOLUME 7

VALUE CREATION IN MULTINATIONAL ENTERPRISE EDITED BY

J. JAY CHOI Temple University

REID W. CLICK George Washington University

Amsterdam – Boston – Heidelberg – London – New York – Oxford Paris – San Diego – San Francisco – Singapore – Sydney – Tokyo JAI Press is an imprint of Elsevier

JAI Press is an imprint of Elsevier The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA First edition 2007 Copyright r 2007 Elsevier Ltd. 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 written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent veriﬁcation of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN-13: 978-0-7623-1392-1 ISBN-10: 0-7623-1392-7 ISSN: 1569-3767 For information on all JAI Press publications visit our website at books.elsevier.com Printed and bound in The Netherlands 07 08 09 10 11 10 9 8 7 6 5 4 3 2 1

CONTENTS ABOUT THE SERIES

ix

LIST OF CONTRIBUTORS

xi

PART I: AN OVERVIEW INTRODUCTION TO VALUE CREATION IN MULTINATIONAL ENTERPRISE J. Jay Choi and Reid W. Click

3

PART II: MULTINATIONALS AND FOREIGN DIRECT INVESTMENT STRATEGIC AND FINANCIAL DETERMINANTS OF FOREIGN DIRECT INVESTMENTS Jongmoo Jay Choi and Eric C. Tsai

19

MODELING THE EVOLUTIONARY SEQUENCE OF INTERNATIONAL JOINT VENTURES Elmar Lukas

61

NEW TRENDS AND PERFORMANCES OF KOREAN OUTWARD FDI AFTER THE FINANCIAL CRISIS Seong-Bong Lee

75

v

vi

CONTENTS

PART III: STRATEGIES AND FIRM PERFORMANCE THE VALUE CREATION PERSPECTIVE OF INTERNATIONAL STRATEGIC MANAGEMENT Reid W. Click

99

OWNERSHIP STRUCTURE, DIVERSIFICATION STRATEGY, AND PERFORMANCE: IMPLICATIONS FOR ASIAN EMERGING MARKET MULTINATIONAL ENTERPRISES Juichuan Chang

125

SUCCESSFUL ADAPTATION STRATEGIES OF MULTINATIONAL ENTERPRISES IN CENTRAL AND EASTERN EUROPE Roxana Wright

149

THE RELATIONSHIP BETWEEN ORGANIZATIONAL STRUCTURES AND PERFORMANCE: THE CASE OF THE FORTUNE 500 Nicole Avdelidou-Fischer

169

PART IV: MERGERS AND ACQUISITIONS DIFFICULTIES IN VALUE CREATION: TELECOM NEW ZEALAND’S ACQUISITION OF AAPT LTD Alireza Tourani-Rad and Zoltan Toth

209

ANALYSIS OF GLOBAL COMPETITORS’ REACTION TO MEGA MERGER ANNOUNCEMENTS BY AN MNC: THE CASE OF THE CITICORP–TRAVELERS MERGER Isaac Otchere and Suhadi Mustopo

229

Contents

vii

A COMBINED CASCADE MODEL TO EXPLAIN INDUSTRIAL CONSOLIDATION: THEORY AND AN APPLICATION TO STEEL Huaichuan Rui

255

PART V: FINANCE AND GOVERNANCE MEASURING AND MANAGING THE FOREIGN EXCHANGE EXPOSURE OF CHINESE COMPANIES Patrick J. Schena UK MEASURES OF FIRM-LIVED EQUITY DURATION Richard A. Lewin, Marc J. Sardy and Stephen E. Satchell

285

307

MULTINATIONALS AND EXCHANGE RATE PASS-THROUGH Alexandra Lai and Oana Secrieru

339

CORPORATE GOVERNANCE IN RUSSIA: A CASE STUDY OF TIMELINESS OF FINANCIAL REPORTING IN THE TELECOM INDUSTRY Robert W. McGee

365

PART VI: THE FINANCIAL SERVICES SECTOR MERGERS AND CONSOLIDATION OF FINANCIAL SERVICE FIRMS: GLOBAL TRENDS AND STRATEGIES FOR VALUE CREATION Edward C. Boyer and Jongmoo Jay Choi

393

viii

CONTENTS

CROSS-BORDER INVESTMENT IN THE LATIN AMERICAN BANKING SECTOR Jesu´s Arteaga-Ortiz, Harvey Arbela´ez and Wendy M. Jeffus

419

PART VII: THE INFLUENCE OF THE STATE GOVERNANCE AND POLITICAL RISK Arvind K. Jain

441

STRATEGIC TRADE POLICY FOR SMALL STATES: A POLITICAL ECONOMY PERSPECTIVE Jaleel Ahmad

461

STRATEGIC IPO UNDERPRICING: THE ROLE OF CHINESE STATE OWNERSHIP Yong Wang and Xiaotian (Tina) Zhang

475

MULTINATIONAL VERSUS STATE POWER IN AN ERA OF GLOBALIZATION: THE CASE OF MICROSOFT IN CHINA, 1987–2004 Jean-Marc F. Blanchard

497

INTERNATIONALIZATION AND VALUE CREATION IN THE GLOBAL TEXTILES AND APPAREL INDUSTRY: A COMPARATIVE ANALYSIS OF LITHUANIA AND MOLDOVA Sanford L. Moskowitz

535

STRATEGY DURING AN INDUSTRY CRISIS: THE POST-QUOTA EXPERIENCE OF TURKISH APPAREL MANUFACTURERS Liesl Riddle

565

ABOUT THE SERIES International Finance Review is an annual book series in the international ﬁnance area (IFR, broadly deﬁned). The IFR will publish theoretical, empirical, institutional or policy-oriented articles on multinational business ﬁnance and strategies, global capital markets and investments, global risk management, global corporate ﬁnance and institutions, currency markets and international ﬁnancial economics, emerging market ﬁnance, or related regional or country-speciﬁc issues. In general, each volume will have a particular theme. Those interested in contributing an article or editing a volume should contact the Series Editor, J. Jay Choi at E-mail: [email protected]

EDITORIAL ADVISORY BOARD M. Adler Columbia University NY, USA

V. Errunza McGill University, Montreal, Quebec, Canada

W. Bailey Cornell University, Ithaca, NY, USA

R. Grosse Thunderbird Business School, Glendale, AZ, USA

I. Cooper London Business School, UK

Y. Hamao University of Southern California, Los Angeles, CA, USA

J. Doukas Old Dominion University/European Financial Management, Norfolk, VA, USA

C. R. Harvey Duke University, Durham, NC, USA

G. Dufey University of Michigan, Ann Arbor, MI, USA

R. Hawkins Georgia Institute of Technology, Atlanta, GA, USA ix

x

ABOUT THE SERIES

J. E. Hodder University of Wisconsin, Madison, WI, USA

R. Marston University of Pennsylvania, Philadelphia, PA, USA

M. Levi University of British Columbia, Vancouver, B C, Canada

R. Roll University of California at Los Angeles, CA, USA

D. Logue Dartmouth College, Hanover, NH, USA

A. Saunders New York University, NY, USA

J. Lothian Fordham University, NY, USA

R. Sweeney Georgetown University, Washington, DC, USA

LIST OF CONTRIBUTORS Jaleel Ahmad

Concordia University, Montreal, Canada

Harvey Arbela´ez

Monterey Institute of International Studies, Monterey, CA, USA; Groupe ESC Lille, Lille, France

Jesu´s Arteaga-Ortiz

Universidad de Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain

Nicole Avdelidou-Fischer

Queen Mary, University of London, London, UK

Jean-Marc F. Blanchard

San Francisco State University, San Francisco, CA, USA

Edward C. Boyer

Temple University, Philadelphia, PA, USA

Juichuan Chang

Ling Tung University, Taichung City, Taiwan

J. Jay Choi

Temple University, Philadelphia, PA, USA

Reid W. Click

George Washington University, Washington, DC, USA

Arvind K. Jain

Concordia University, Montreal, Canada

Wendy M. Jeffus

Southern New Hampshire University, Manchester, NH, USA

Alexandra Lai

Bank of Canada, Ottawa, Canada

Seong-Bong Lee

Korea Institute for International Economic Policy, Seoul, Korea

Richard A. Lewin

University of Cambridge, Cambridge, UK xi

xii

LIST OF CONTRIBUTORS

Elmar Lukas

University of Paderborn, Paderborn, Germany

Robert W. McGee

Barry University, Miami Shores, FL, USA

Sanford L. Moskowitz

Saint John’s University, Collegeville, MN, USA

Suhadi Mustopo

Deutsche Bank AG, Jakarta, Indonesia

Isaac Otchere

Carleton University, Ottawa, Canada

Liesl Riddle

George Washington University, Washington, DC, USA

Huaichuan Rui

Brunel University, Uxbridge, UK

Marc J. Sardy

Rollins College, Winter Park, FL, USA

Stephen E. Satchell

University of Cambridge, Cambridge, UK

Patrick J. Schena

The Fletcher School, Tufts University, Medford, MA, USA

Oana Secrieru

Bank of Canada, Ottawa, ON, Canada

Zoltan Toth

Auckland University of Technology, Auckland, New Zealand

Alireza Tourani-Rad

Auckland University of Technology, Auckland, New Zealand

Eric C. Tsai

State University of New York at Oswego, Oswego, NY, USA

Yong Wang

Temple University, Philadelphia, PA, USA

Roxana Wright

Keene State College, Keene, NH, USA

Xiaotian (Tina) Zhang

Temple University, Philadelphia, PA, USA

PART I: AN OVERVIEW

This page intentionally left blank

INTRODUCTION TO VALUE CREATION IN MULTINATIONAL ENTERPRISE J. Jay Choi and Reid W. Click OVERVIEW In a fundamental sense, creation of value is the purpose of a ﬁrm. Values – measured by proﬁts, cash ﬂows, stock prices, or some strategic objectives – are the ultimate reasons why a ﬁrm exists. The ongoing and ever-expanding discussion of globalization, whether based on trade ﬂows or ﬁnancial ﬂows, draws attention to the value of multinational enterprise. Existing empirical work on the impact of multinational ﬁrms, however, is inconclusive. Some observers point to the valuation discount with international operations due to the costs of agency and control and the difﬁculty of coordinating complex organizations and cultures. Others emphasize the value of a multinational network and the operational efﬁciency of a multinational enterprise (MNE). Thus, issues related to value creation are important and lively areas of business and ﬁnance. In fact, value creation is now at the frontier between the functional areas of ﬁnance and strategy. In international ﬁnance, the topic also interacts with economics in the areas of strategic trade policy and exchange rate behavior, as well as business strategy, as it relates to the management of an MNE. The role of government policy is also part of the debate, because the importance of public policy and the behavior of Value Creation in Multinational Enterprise International Finance Review, Volume 7, 3–15 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07001-4

3

4

J. JAY CHOI AND REID W. CLICK

policymakers are elevated when ﬁnance and business become international, as evidenced by consistent attention to political risk. The papers in this volume of International Finance Review provide a reﬂection on the role of international ﬁnance – and its relationship to strategy, economics, political science, and public policy – in examining value creation in multinational enterprise. These are 22 original papers, not published elsewhere, submitted speciﬁcally for this volume based on its theme. The papers present a breadth of methodologies, including theoretical, empirical, conceptual, and case study approaches. Several papers offer combinations of these different categories. Among the empirical papers, there are many kinds of data sets analyzed, ranging from macroeconomic data to ﬁrm-level ﬁnancial data to survey data. In addition, the data sets are rigorously analyzed in many different ways. This volume also takes a broad perspective on multinational enterprise, which not only allows discussion of traditional areas in the study of MNEs as corporations but also includes topics related to multinational enterprise as an undertaking and not just as corporation. For example, there is attention to small- and medium-sized companies as well as larger MNEs. There are also papers that consider exporting enterprises and the environment of multinational enterprise. In this spirit, the volume covers multinational enterprise from a variety of perspectives, including views from private corporations and government policymakers, and the authors of the papers include both academics and practitioners. Altogether, the papers offer insights into value creation through a variety of lenses. One prominent feature of the volume is that the papers collectively cover nearly all areas of the world. As some acknowledgment of its current importance in the world economy, there are four papers on China, ranging from ﬁnancial studies to industry studies. There are speciﬁc country inquiries for Korea, Russia, the U.K., Spain, Turkey, Lithuania and Moldova, and Australia and New Zealand. There are also regional studies of Asia, Latin America, and Central and Eastern Europe. In addition, four papers consider general foreign operations of U.S. enterprises. Another feature of the volume is that many different industries are examined. There are naturally several papers on ﬁnancial services, but there are also papers on telecommunications ﬁrms, the steel industry, and textiles and apparel. The papers are divided into seven parts, including, as Part I, this overview. Part II begins our examination of value creation in multinational enterprise with broad attention to MNEs and foreign direct investment (FDI). In Part III, there is greater focus on strategies and ﬁrm performance. Parts IV–VI offer deeper analysis in three speciﬁc, yet prominent, areas of

Introduction to Value Creation in Multinational Enterprise

5

value creation and MNEs: mergers and acquisitions, ﬁnance and governance, and the ﬁnancial services sector. Part VII returns to a broader picture, examining the inﬂuence of the state in value creation in multinational enterprise, as a way to wrap up the volume.

MULTINATIONALS AND FOREIGN DIRECT INVESTMENT Part II examines multinationals and FDI from both empirical and theoretical perspectives. Jay Choi and Eric Tsai open the discussion with ‘‘Strategic and Financial Determinants of Foreign Direct Investments.’’ They point out that conventional theories regard FDIs as strategic moves based on operational or industrial organization considerations. Choi and Tsai, however, demonstrate that ﬁnancial factors are also important in corporate FDI decisions. The paper offers a comprehensive assessment through an integrated model with both strategic and ﬁnancial factors, and empirically investigates the factors using ﬁrm-level data from the FDIs of U.S. corporations. The ﬁnancial factors concern internal capital market strength and corporate governance and include exchange rate changes, internal and external ﬁnancing cost, risk diversiﬁcation, and agency costs. The results of the study demonstrate that there is variability in the signiﬁcance of ﬁnancial variables depending on industries and destinations. The integrated model with both strategic and ﬁnancial factors is shown to be superior to either component model in explaining FDIs, and ﬁnancial factors are shown to be no less important in their ability to explain the prevailing FDI phenomena than strategic or operational variables. Speciﬁc attention to international joint ventures is provided in Elmar Lukas’s ‘‘Modeling the Evolutionary Sequence of International Joint Ventures.’’ The paper is a theoretical contribution to the literature on FDI under uncertainty and underscores the importance of modeling the evolutionary sequence pattern of foreign market entry. The model in the paper helps reﬁne the application of real options in the international joint venture context by providing a closed-form solution in continuous time to value their overall strategic ﬂexibility. Moreover, the analysis provides a novel perspective on existing empirical results and generates a number of new testable predictions. The ﬁnal paper in this section is an examination of FDI from a different part of the world. Seong-Bong Lee examines the empirical patterns of

6

J. JAY CHOI AND REID W. CLICK

Korean FDI in ‘‘New Trends and Performances of Korean Outward FDI after the Financial Crisis’’ using macroeconomic data and ﬁrm-level data aggregated to the macro level. Korean outward FDI increased rapidly during the 1990s, and the Korean chaebols were the main players in this increase. However, Lee points out that outward FDI decreased signiﬁcantly during the three to four years after the ﬁnancial crisis and then reversed in 2002, increasing once again. Lee documents that a feature of the current rise in outward FDI is that not only are large enterprises engaging in FDI, but small and medium-sized companies are too. Therefore, Korean outward FDI now shows a mixed investment pattern in which two different types of investment coexist: one is performed by big companies to secure large local markets and the other by small- and medium-sized companies attempting to ensure low labor costs. This new pattern reﬂects structural changes in the Korean economy as well as the growing trend of reciprocal inﬂuences. Lee not only analyzes these new outward FDI patterns, but also examines the performance of outward FDI after the ﬁnancial crisis in Korea. Lee also offers a concluding discussion of the anticipated impact of these changes in outward FDI on the Korean economy.

STRATEGIES AND FIRM PERFORMANCE With a broad picture generally established, Part III moves on to consider strategies and ﬁrm performance. The four papers in this section address the issue of value creation in MNEs through empirical investigations. Two papers utilize data on the foreign operations of U.S. MNEs and the other two utilize data from the Asia/Paciﬁc and Central/Eastern European regions. A general background to strategy and ﬁrm performance is provided in the ﬁrst paper of this section, ‘‘The Value Creation Perspective of International Strategic Management.’’ In this paper, Reid W. Click uses the principle of value creation to analyze the strategic management of MNEs. International strategic management is ﬁrst deﬁned as the process through which value is created by managers operating across a national border. The domain of international strategic management is thus determined by activities that distinguish international management from domestic management in the process of value creation. Click then uses this perspective on value creation to answer three questions pertaining to international strategic management. First, how important is international strategic management? Simple statistics presented in the paper demonstrate that the international component of value creation is important in the U.S. economy. Second, what is the

Introduction to Value Creation in Multinational Enterprise

7

domain of international strategic management? The paper presents a framework in which international strategic management is the aggregation of value created through international production, marketing, and ﬁnancial activities, and reveals that the domain of international management is vast. Third, does international strategic management make the whole MNE worth more than the sum of its parts? Empirical evidence uncovered using stock market returns suggests that the answer is yes, at least for U.S. multinationals in the early 1990s. The second paper in this section builds upon the ﬁrst paper by addressing ‘‘Ownership Structure, Diversiﬁcation Strategy and Performance: Implications for Asian Emerging Market Multinational Enterprises.’’ Juichuan Chang, the author, presents additional discussion of the issues and turns our attention to a different part of the world. Chang provides an integrated framework that conceptualizes the antecedents of international expansion by emerging market businesses in relation to ﬁrm performance. The data set uses 115 MNEs from four Asian countries: Hong Kong, Taiwan, Korea, and Singapore. The methodology develops multiple-item measures of multiple dimensions to clarify ownership structure and categorizes three diversiﬁcation strategies. Chang then tests how ownership structure and diversiﬁcation strategy affect the ﬁnancial performance of emerging market MNEs using return on assets and return on sales. The results indicate that the relationship between ownership structure and ﬁrm performance is nonlinear (in fact, an S shape). The results also indicate that excessive international diversiﬁcation, product diversiﬁcation, or geographic scope of the expansion process negatively moderate the impact of Asian MNEs’ performance. The third paper in this section moves the discussion to a different part of the world. Roxana Wright examines the ‘‘Successful Adaptation Strategies of Multinational Enterprises in Central and Eastern Europe.’’ The paper explores strategies of adaptation to the environment as employed by multinational corporations and empirically investigates determinants of success using a sample of 100 MNE subsidiaries operating in 19 Central and Eastern European countries. Organizations are treated as adaptive systems that have to match the complexity of their environments. The research recognizes the complex nature of the market institutions emerging from transition, which emphasizes the need for new managerial frameworks. Adaptive approaches such as vertical integration and/or value chain development, leveraging autonomy and integration, local knowledge acquisition, and embedding in the social and political environment are explored in their relationship to success in the region.

8

J. JAY CHOI AND REID W. CLICK

The last paper in this section on strategy and performance moves the discussion back to the U.S. in an empirical investigation of the Fortune 500 companies. ‘‘The Relationship Between Organizational Structures and Performance: The Case of the Fortune 500,’’ by Nicole Avdelidou-Fischer, focuses on Fortune 500 companies because they have generally been among the most proﬁtable and admired in the world. After a discussion of whether companies should organize regionally, nationally, or globally, the important assumption is made that each structural type utilizes resources differently in generating proﬁt. Performance is conceptualized as return on capital employed and return per employee. Statistical tests reveal that structural types are positively related to ﬁnancial performance, calculated as return on capital employed, with multidivisional-structured companies outperforming functional-structured ones. Tests also suggest that structural types are not related to human resource performance, calculated as return per employee.

MERGERS AND ACQUISITIONS Part IV considers a speciﬁc aspect of the investigation of value creation and MNEs: mergers and acquisitions. Three papers in this section demonstrate that we can learn a lot about value creation in MNEs by studying large, perhaps extreme, events. The papers, in fact, consider mergers and acquisitions in three very different industries: telecommunications, ﬁnancial services, and steel. Alireza Tourani-Rad and Zoltan Toth take a case study approach and examine a single acquisition that proved to be generally unsuccessful. Their paper, ‘‘Difﬁculties in Value Creation: Telecom New Zealand’s Acquisition of AAPT Ltd.,’’ begins with an overview of the Australian and New Zealand telecommunications markets and Telecom New Zealand’s (TCNZ) NZ$2 billion acquisition of AAPT Ltd., Australia’s third largest telecommunications ﬁrm, in 2000. The paper documents that a few years later, after writing off approximately NZ$1 billion, TCNZ is considering a sell-off at a considerable loss. Tourani-Rad and Toth discuss the strategic reasons behind the acquisition and explain how smaller telecom companies are struggling to compete with the incumbent telecom in Australia. They further conduct an event study to assess the impact of the acquisition on both TCNZ’s and AAPT’s share prices and look at some of the post-acquisition issues. In a second case study, ‘‘Analysis of Global Competitors’ Reaction to Mega Merger Announcements by an MNC: The Case of the Citicorp

Introduction to Value Creation in Multinational Enterprise

9

Travelers Merger,’’ Isaac Otchere and Suhadi Mutopo ﬁnd that global competitors, especially banks in Europe and the U.S., reacted positively to the Citicorp and Travelers merger announcements in 1998. Uncertainties created by investigations into the merger proposal had a signiﬁcant impact on competitors’ stock price. The announcement that the merger had been consummated also elicited a signiﬁcantly positive reaction from the rivals following the resolution of uncertainties emanating from the regulatory challenges. The positive reaction by competitors suggests that the merger was a wealth-creating event for the large ﬁrms in the ﬁnancial services industry. The expected beneﬁts outweighed any competitive effects resulting from the merger. The competitors’ reaction was, however, not homogenous. Cross-sectional analysis in the paper shows that the abnormal returns earned by the competitors were higher the larger the competitor. In addition, the abnormal returns were greater for the U.S. rivals. Otchere and Mutopo conclude that the fact that global competitors reacted positively to the Citicorp–Travelers mega merger announcement is consistent with their assertion that the merger had ramiﬁcations that go beyond regulatory concerns in the U.S. Huaichuan Rui’s paper, ‘‘A Combined Cascade Model to Explain Industrial Consolidation: Theory and an Application to Steel,’’ offers a rigorous theoretical examination of mergers and acquisitions along with a test of the theory using data from the global steel industry and, particularly, the steel industry in China. Rui points out that expansion through mergers and acquisitions continues to be a viable international strategy utilized by industrial ﬁrms. A striking feature of this is that global giant ﬁrms lead the wave and generate an unimaginable impact on relatively small and weak ﬁrms across sectors and even nations. Rui observes that there seems to be a ‘‘cascade effect’’ between the industrial consolidations in these areas. A combined cascade model developed in the paper explains that the power imbalance caused by the degree of consolidation of the players within a ﬁrm’s value system determines the movement and direction of the cascade effect. Rui concludes that with the existence of such an effect, mergers and acquisitions will be a mutually interdependent, dynamic, reversible, and endless process among industries.

FINANCE AND GOVERNANCE Following on the heels of our detailed discussions of mergers and acquisitions, Part V offers detailed discussion of ﬁnance and governance in

10

J. JAY CHOI AND REID W. CLICK

MNEs. There is one theoretical paper, and the others offer empirical investigations of China, the U.K., and Russia. Patrick J. Schena examines the traditional topic of foreign exchange exposure for a new set of companies in ‘‘Measuring and Managing the Foreign Exchange Exposure of Chinese Companies.’’ Speciﬁcally, the paper explores the sensitivity of Chinese stock returns to changes in trade-weighted indices of the Chinese currency (the RMB) and the currencies of China’s trading partners from 1999 to 2003. The paper analyzes an exposure elasticity crosssectionally using accounting variables to proxy for size and costs of ﬁnancial distress. Schena ﬁnds that internationally oriented Chinese companies have experienced exchange exposure particularly against the Japanese yen. He also ﬁnds that, against a trade-weighted index, there is no empirical evidence that Chinese ﬁrms are engaged in hedging activities. However, when exposures are measured in yen terms, Schena ﬁnds that Chinese ﬁrms, particularly exporters, engage in active currency hedging. A second type of exposure, interest rate exposure, is examined in ‘‘U.K. Measures of Firm-Lived Equity Duration’’ by Richard Lewin, Marc Sardy, and Stephen Satchell. Investors often have much of their portfolios invested in equities that are exposed to interest rate risk. Hedging underlying equity exposures is not easy; in contrast, ﬁxed-income investors have duration to immunize bond portfolios from small ﬂuctuations in interest rates. U.S. equity duration estimates from a dividend discount model result in long durations – often in excess of 50 years. Based on U.K. data, Lewin, Sardy, and Satchell develop an alternative approach to generate equity duration as a by-product of asset pricing. Their analysis suggests that the equity premium puzzle may comprise an important element in reconciling this approach to equity duration with traditional dividend discount model alternatives. Alexandra Lai and Oana Secrieru study another traditional topic in international ﬁnance that is closely related to foreign exchange exposure: exchange rate pass-through. Their paper, ‘‘Multinationals and Exchange Rate Pass-Through,’’ offers a rigorous theoretical approach to the topic. The paper examines the impact an MNE has on the degree of exchange rate pass-through when it engages in Cournot competition with domestic and foreign rivals. The MNE can locate its production for the foreign market domestically – resulting in intra-ﬁrm trade – or in the foreign country – resulting in international production. In addition to incomplete exchange rate pass-through, Lai and Secrieru show that an MNE increases the sensitivity of domestic market prices and reduces the sensitivity of foreign market prices to exchange rate movements. In addition, Lai and Secrieru

Introduction to Value Creation in Multinational Enterprise

11

show that intra-ﬁrm trade prices are more sensitive to exchange rate movements than their international production counterparts and react in the opposite direction. The ﬁnal paper in this section considers the role of ﬁnancial information within the context of corporate governance in a transitional country. In ‘‘Corporate Governance in Russia: A Case Study of Timeliness of Financial Reporting in the Telecom Industry,’’ Robert McGee address timeliness of ﬁnancial reporting as an important element of transparency. Speciﬁcally, McGee looks at the telecommunications industry in Russia and computes the number of days it takes companies to receive an audit opinion, then compares the time lag to the number of days it takes non-Russian companies in the telecommunications industry to receive an audit opinion. McGee ﬁnds that Russian companies take longer to report ﬁnancial results than do nonRussian companies. In addition, larger Russian companies take less time to report their ﬁnancial condition than do small Russian ﬁrms, but the difference is not signiﬁcant. Companies using Russian Accounting Standards take signiﬁcantly less time to report ﬁnancial results than companies using either International Financial Reporting Standards (IFRS) or U.S. Generally Accepted Accounting Principles (GAAP). In addition, companies using IFRS take signiﬁcantly longer to report ﬁnancial results than companies using U.S. GAAP. McGee also ﬁnds that the dominant auditor in the Russian telecommunications industry did not complete audits in signiﬁcantly less time than did non-dominant auditors. Finally, in discussion, McGee concludes that although Russian companies take far less time to issue ﬁnancial statements now than they did a few years ago, it is premature to deﬁnitively conclude that the improvement is signiﬁcant due to the limited data set.

THE FINANCIAL SERVICES SECTOR The ﬁnancial services sector receives particular attention in Part VI. The two papers in this section offer a detailed discussion of many issues already introduced in this volume, including mergers and acquisitions, and macrolevel data on capital ﬂows. Edward Boyer and Jay Choi take up the issue of rapid consolidation globally in the ﬁnancial services industry in ‘‘Mergers and Consolidation of Financial Service Firms: Global Trends and Strategies for Value Creation.’’ Consolidation has proceeded not only in the same market but also across different market segments and across national boundaries. Boyer and Choi

12

J. JAY CHOI AND REID W. CLICK

begin by outlining the general trend of the mergers and acquisitions and consolidation of the ﬁnancial service industry in both the U.S. and the global economy. They next identify and analyze the reasons that contribute to the consolidation of the ﬁnancial service industry, and examine some cases of successful and unsuccessful ﬁnancial service mergers and acquisitions. Finally, they draw some strategic implications. The second paper builds on the ﬁrst paper in this part by looking at the same phenomenon in a speciﬁc area of the world. Wendy Jeffus, Jesu´s Arteaga-Ortiz, and Harvey Arbela´ez look at ‘‘Cross-Border Investment in the Latin American Banking Sector.’’ They point out that cross-border investment in the banking sector has been a large part of overall merger and acquisition activity in Latin America. Spain and the U.S. have been the largest investors, participating in almost 70% of the total transaction value. Speciﬁcally, the total value of Spanish investment within that region is over 50 billion dollars. After an introduction and an explanation of the importance of FDI and the implications for cross-border investment in banking, the paper focuses on the largest investor in the Latin American banking sector and attempts to ﬁnd an explanation for the increasing participation of Spanish banks. Jeffus, Arteaga-Ortiz, and Arbela´ez discuss a potential new reality: Latin America could be the geographic location where major contenders in the banking industry worldwide will be engaged in battles for global dominance. The data for this analysis consist of worldwide investment in Latin America from 1985 to 2002.

THE INFLUENCE OF THE STATE The ﬁnal section of the volume, Part VII, more closely addresses the inﬂuence of the state, although the range of discussion is quite broad. The topic has come up in several other papers. For example, the paper by Otchere and Mutopo in Part IV demonstrated the inﬂuence of the U.S. government in the Citicorp–Travelers merger through regulation. The papers in this section offer more detail and depth on the inﬂuence of the state and consider countries other than the U.S. The ﬁrst paper, Arvind Jain’s ‘‘Governance and Political Risk,’’ looks at corruption. Jain sees political risk as arising from renegotiation of implicit or explicit contracts under which foreign investors enter a host country. Governments will legitimately enter into renegotiation to increase the share of rents earned by society. Corrupt political leaders, however, will use their powers to extract rents from foreign investors for personal gains rather than

Introduction to Value Creation in Multinational Enterprise

13

for the good of society. Jain, therefore, recommends that political risk assessment should assess the intentions of government as well as the strengths of political and social institutions that keep leaders under control. In addition, he points out that ﬁrms should also understand that their own actions may contribute to creating political risk. In the second paper, Jaleel Ahmad addresses ‘‘Strategic Trade Policy for Small States: A Political Economy Perspective.’’ Ahmad points out that there is a literature on strategic trade policy dealing with industries and sectors characterized by international rivalry for market shares and the struggle to capture ‘‘rents’’ over and above normal factor rewards. Ahmad builds on this literature by exploring the validity and implications of strategic trade policy for small states and small ﬁrms that are not major players in international markets. The smallness of the ﬁrms may, in fact, be an advantage rather than a hindrance. The implications of smallness for strategic behavior are examined in a simple game-theoretic framework. The insights become sharper when extended to intra-industry trade in differentiated products. Ahmad concludes that the desirable policy interventions for small countries and ﬁrms are quite different from those for large ﬁrms. The third paper considers a traditional topic in ﬁnance, the underpricing of initial public offerings (IPOs), and demonstrates that in certain circumstances the inﬂuence of the state is very important. Yong Wang and Xiaotian (Tina) Zhang extend the traditional ﬁnance topic to account for the role of the state in ‘‘Strategic IPO Underpricing: The Role of Chinese State Ownership.’’ Wang and Zhang point out that the stock markets in emerging economies are attractive to international investors but their unique characteristics need to be examined. In particular, Chinese stock markets experienced much more signiﬁcant IPO underpricing than most other stock markets in the world. The paper offers a two-period wealth maximum model to explain the strategic IPO underpricing by state owners. Given the fact that the entire IPO procedure (including the IPO price) is regulated and controlled by state owners, Wang and Zhang argue that state owners strategically underprice the IPO because they care less about the IPO proceeds and more about the wealth gain after IPO. Empirically, Wang and Zhang ﬁnd a positive relationship between IPO underpricing and state ownership in the Chinese stock market, and this is consistent with the wealth maximization hypothesis of IPO pricing. Jean-Marc Blanchard considers another approach to the issue of the inﬂuence of the state in ‘‘Multinational versus State Power in an Era of Globalization: The Case of Microsoft in China, 1987–2004’’ the fourth paper in this section. Blanchard examines relative power by comparing and

14

J. JAY CHOI AND REID W. CLICK

contrasting two schools of thought: a ‘‘multinationals in command’’ perspective and a ‘‘states in command’’ perspective. The paper offers an alternative analytical framework, a modiﬁed bargaining power model highlighting three factors as playing a decisive role in shaping the power relationship between states and multinationals: the balance of needs, allies/ adversaries, and the institutional environment. To illustrate the model and demonstrate its explanatory value, Blanchard examines the experience of Microsoft in China. In the ﬁfth paper on the inﬂuence of the state, ‘‘Internationalization and Value Creation in the Global Textiles and Apparel Industry: A Comparative Analysis of Lithuania and Moldova,’’ Sanford Moskowitz considers the political afﬁliations of the state as a determinant of the success of local industries. The study examines the internationalization process within the textiles and apparel industry and shows how the evolution of an industry toward greater internationalization is intricately linked to its ability to move up its speciﬁc value chain. The analysis compares and contrasts the ability of this industry within a Western European country that is a member of the European Union (Lithuania) and a non-accession Eastern European country (Moldova) to move up the textiles and apparel value chain and so achieve higher levels of internationalization. In examining and relating the relevant factors, the analysis provides insights into – and suggests important modiﬁcations to – such important concepts and themes as the stage theory of internationalization, the role of ‘‘inward–outward’’ linkages in the value creation process, the mechanism of small- and medium-size ﬁrm internationalization, and the part played by the European Union in the internationalization (and thus globalization) process. Liesl Riddle builds on the discussion of the textile and apparel industry in the sixth and ﬁnal paper in this part, ‘‘Strategy During an Industry Crisis: The Post-Quota Experience of Turkish Apparel Manufacturers.’’ Riddle points out that discussions about the impact of quota elimination in the apparel industry have focused on countries that obviously beneﬁt from their elimination (such as China) and those that clearly are harmed by their demise (such as the sub-Saharan African countries). However, the global production landscape of the apparel trading system also includes numerous countries for which the post-quota environment is uncertain – and vital, but little attention has been paid to the speciﬁc impact of the quota elimination on these countries. At the dawn of quota elimination, uncertainty loomed particularly large for Turkey, the world’s fourth largest apparel exporting nation. Riddle’s investigation draws on secondary data and a survey of Turkish clothing exporters to chronicle Turkey’s quota elimination

Introduction to Value Creation in Multinational Enterprise

15

experience during 2005 and 2006. The results indicate that most Turkish apparel exporters expected their market share in the European Union and the U.S. to remain the same or improve in 2005. The reality was far worse: by 2006 the industry had incurred large losses and production had declined sharply. The case provides fashion managers and scholars with insight into the experience of a country for which the post-quota elimination future was uncertain but for whom the apparel industry was vital in 2005.

CONCLUSION Taken together, the 23 original papers in this volume cover a rich array of topics and capture the excitement of studying value creation in MNEs. They provide lessons that are applicable to a variety of participants, ranging from managers in ﬁrms to public policymakers. The wide spectrum of ﬁelds represented and methodologies implemented suggests that researchers are looking at value creation in MNEs from many different perspectives. Thanks to the generosity of the contributing authors, this volume of International Finance Review captures the essence of current thought on value creation in MNEs, and hopefully will stimulate further investigation of the topic in the future.

This page intentionally left blank

PART II: MULTINATIONALS AND FOREIGN DIRECT INVESTMENT

This page intentionally left blank

STRATEGIC AND FINANCIAL DETERMINANTS OF FOREIGN DIRECT INVESTMENTS Jongmoo Jay Choi and Eric C. Tsai ABSTRACT Conventional foreign direct investment (FDI) theories regard FDIs as strategic moves based on operational or industrial organization considerations. We demonstrate that financial factors are also important in corporate FDI decisions. The financial factors concern internal capital market strength and corporate governance and include exchange rate changes, internal and external financing cost, risk diversification, and agency costs. There is variability in the significance of financial variables depending on industries and destinations. The integrated model with both strategic and financial factors is superior to either component model in explaining FDIs. However, financial factors are no less important in explaining the prevailing FDI phenomena than strategic or operational variables.

1. INTRODUCTION The mainstream theory of foreign direct investments (FDIs) grew out of international trade and industrial organization traditions. International Value Creation in Multinational Enterprise International Finance Review, Volume 7, 19–60 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07002-6

19

20

JONGMOO JAY CHOI AND ERIC C. TSAI

trade theory indicates that international differences in locational endowments lead to international trade. Industrial organization theory suggests ﬁrm-speciﬁc ownership factors as the primary reasons why multinational corporations (MNCs) can overcome entry costs in making investments abroad. These endowment-focused theories (of either location or ﬁrm ownership) can explain FDI ﬂows from developed to less-developed regions of the world. To explain the vast majority of FDI ﬂows between developed countries (DCs), internalization theory was developed.1 It suggests that, in some areas, MNCs as institutions may be more operationally efﬁcient as arbiters of resource allocation than markets. To achieve efﬁcient global resource allocation, i.e., foreign production, an MNC can create its own internal markets to reduce transaction costs and to prevent valuation loss associated with external transactions, especially for information-intensive products in which MNCs specialize. Dunning (1977) suggests an eclectic theory that combines international trade, industrial organization, and internalization theories in an integrated general strategic framework. The advantages for MNCs are due to three strategic sources: ownership, location, and internalization (OLI). The integrated strategic theory is a milestone in the FDI theory, but it leaves some issues unanswered. First of all, the OLI framework does not explain why the most common entry mode of FDIs is international mergers and acquisitions (M&As) rather than greenﬁeld investments. Second, internalization is theoretically beneﬁcial to MNCs regardless of where investment occurs. Internalization therefore does not seem to justify the phenomenon that most FDIs occur between DCs. Third, signiﬁcant bilateral short-term FDI swings are frequently observed, and it is not clear whether these swings correspond to changes in those strategic factors. Finally, as Itaki (1991) notes, Dunning’s OLI framework does not include any ﬁnancial factors at all. This is in contrast to normative decision making in practice that usually includes both strategic and ﬁnancial factors in international investment decisions. Harris and Ravenscraft (1991) suggest three hypotheses for cross-border investments: (a) imperfections in product and factor markets, (b) biases in government and regulatory policies, and (c) imperfections in capital markets. Their empirical work, however, only examined the ﬁrst hypothesis – which is actually the core of the conventional strategic FDI theories. In this paper, we focus on examining the third hypothesis with the intention of relating it to the ﬁrst. We do that by introducing the ﬁnancial variables, as well as the structural ones, as determinants of FDIs. Our ﬁnancial model depicts FDIs as the consequence of the ‘‘interaction’’ between a strong ‘‘internal’’ capital market and ‘‘internal’’ corporate

Strategic and Financial Determinants of Foreign Direct Investments

21

governance. We show that both ﬁnancial and strategic variables are signiﬁcant in explaining corporate international investment behavior, and that these results are fairly robust across industries and for different speciﬁcations. We suggest then that the neglect of ﬁnancial variables in the FDI model may constitute a misspeciﬁcation of the model. However, compared to the strategic model, the ﬁnancial model exhibits greater variability depending on the industry sector and investment destinations. By incorporating ﬁnancial variables, we show that Dunning’s ‘‘eclectic’’ theory would be even more eclectic. Beyond this implication for the academic FDI literature, this paper should demonstrate the usefulness of our integrated model for corporate practitioners who may be examining strategic and ﬁnancial reasons for investing abroad or for policymakers who may be interested in providing incentives to attract foreign investments. The structure of this paper is as follows. Section 2 discusses the related literature. Section 3 presents the estimation framework and Section 4 describes data used in the empirical work. Empirical results are then presented in Section 5. Section 6 includes a summary conclusion.

2. REVIEW OF RELATED WORK 2.1. Strategic Perspectives on FDIs Traditional international trade theory that emphasizes location endowments as determinants of FDIs is insufﬁcient as a theory of FDI because it cannot explain why international investment takes place in addition to international trade. The modern strategic theory of FDI, developed by Hymer (1976), Kindleberger (1969), and Caves (1971), starts with the notion that excess proﬁt projects are created by ‘‘structural imperfections’’ of some kind in goods or factor markets. FDIs take place when MNCs take on those projects as they try to utilize their ﬁrm-speciﬁc ownership endowments, such as proprietary technology and others, to generate greater returns (despite their disadvantage of facing signiﬁcant entry costs) than those of indigenous local ﬁrms. MNCs can produce greater proﬁt by switching from an international exchange business (export) to a foreign production business (FDI). Buckley and Casson (1976), Dunning (1977), and Casson (1979) point out that the internalization of intermediate goods markets gives an advantage to a ﬁrm that can reduce or eliminate the transaction costs caused by ‘‘natural market imperfections.’’ Furthermore, the uncertainty in the valuation of intangible investments or information-intensive products in the open market

22

JONGMOO JAY CHOI AND ERIC C. TSAI

implies that internalization is a way of appropriating the potential value of research and development within the ﬁrm. Thus, the ability of an MNC to generate rents due to its ownership endowments is enhanced by the beneﬁts of an internal market created within a ﬁrm, and when these beneﬁts are important, FDIs make more sense rather than exporting or licensing. Dunning (1977, 1980, 1988) consolidated different strands of mainstream FDI theories in an integrated eclectic framework that combines OLI factors conducive to FDIs. 2.2. Financial Perspectives on FDIs The market imperfection paradigm assumed in the mainstream strategic theory of FDI by Hymer (1976), Kindleberger (1969), and Caves (1971) does not exclude imperfections in ﬁnancial markets. In addition, the internalization theory emphasizes the distinction between external and internal transactions, which could include ﬁnancial transactions. Therefore, a ﬁnancial perspective on FDIs is entirely consistent with the assumptions behind the mainstream strategic FDI theories in the industrial organization framework. From a ﬁnancial standpoint, variables of major interest include valuation, cost, and risk. For instance, why is a given foreign project valued more for an MNC than for an indigenous ﬁrm? Otherwise no FDI will take place. Table 1 indicates that about three-quarters of all direct investments abroad by U.S. ﬁrms between 1992 and 1998 are in the form of M&As rather than greenﬁeld investments. The choice between M&As and greenﬁelds would depend, among other things, on how the ﬁrm is valued in the market place as opposed to the value of the ﬁrm’s assets. If both ﬁnancial and real markets are efﬁcient and in perfect equilibrium, then there is no disparity in valuation and there is little incentive to utilize M&As versus greenﬁelds or, for that matter, little difference between investing at home and abroad. Aliber (1970) was the ﬁrst to suggest that FDIs can take place because of ﬁnancial factors in an environment of partially segmented international capital markets. MNCs generally have superior access to ﬁnancing in the stable currency zone in international capital markets while the indigenous ﬁrm can only raise funds in the relatively weak currency unit. The stability of currency means a lower currency risk premium, and hence a lower cost of capital, for MNCs than for local indigenous ﬁrms. As such, Aliber (1970) suggested the possibility that a multinational ﬁrm can obtain a higher valuation than an indigenous ﬁrm from the same project due to the advantage of lower cost of capital.

Outward and Inward Foreign Direct Investment (FDI) Flows and Stocks of the United States. 1986–1991 Annual Average

1992

1993

1994

1995

1996

1997

1998

1992–1998 Annual Average

Panel A: Outward and Inward FDI Flows of the U.S. Outward ﬂows 26.0 U.S. as % of world 14.4 Inward ﬂows 49.0 U.S. as % of world 30.8

39.0 19.4 18.9 10.8

74.8 31.1 43.5 20.0

73.3 25.8 45.1 18.6

92.1 26.1 58.8 17.8

74.8 19.7 76.5 21.3

110.0 23.2 109.3 23.5

132.8 20.5 193.4 30.0

85.3 23.1 77.9 22.4

1980

1985

1990

1995

1997

1998

Panel B: Outward and Inward FDI Stocks of the U.S. Outﬂow stock 220.2 251.0 U.S. as % of world 42.9 36.6 Inward stock 83.0 184.6 U.S. as % of world 16.4 23.6

435.2 25.4 394.9 22.3

696.1 24.5 535.6 19.2

860.7 25.1 681.7 19.8

993.6 24.1 875.0 21.4

Note: FDI ﬁgures are in billions of U.S. dollars. Data Source: Various issues of World Investment Report, United Nations Conference on Trade and Development.

Strategic and Financial Determinants of Foreign Direct Investments

Table 1.

23

24

JONGMOO JAY CHOI AND ERIC C. TSAI

The differential access to ﬁnancing is further attributed to information asymmetry by Froot and Stein (1991). With the presence of information imperfection in otherwise globally integrated capital markets, using an auction example in which a U.S. target ﬁrm was sold to the highest bidder, they show that a depreciation of the host country currency systematically reduces the relative wealth of host country agents, facilitating acquisitions of local ﬁrms or assets by foreign ﬁrms. In their view, information imperfections in capital markets cause external ﬁnancing to be more expensive because of monitoring costs. No ﬁrm, therefore, is able to borrow, without incurring additional cost, the full amount of the project or the value of the acquisition target. Since complete external ﬁnancing is costly, the relative corporate wealth between foreign and domestic bidders as a source of internal ﬁnancing becomes relevant. A depreciation of the host country currency systematically raises the relative wealth of foreign ﬁrms, who in turn can outbid indigenous local ﬁrms. Nonetheless, given additional operational costs associated with FDIs, their explanations based on ﬁnancing do not extend to why FDIs are sustainable in the long run. Given partially segmented international capital markets and/or the presence of agency costs, the two-country corporate valuation model of Choi (1989) shows that corporate international investment decisions depend on two major factors: (a) the real foreign exchange risk, which affects the cost of capital, and (b) the gains from diversiﬁcation of operational cash ﬂows, which are necessary at the corporate level given agency costs or market imperfections. The models by Black (1978) and Stulz (1983) emphasize the change in relative distribution of wealth globally as a factor inducing foreign investments. In the empirical literature, exchange rates received the most attention but the results on them are somewhat mixed. A signiﬁcant linkage is shown between exchange rates and U.S. FDIs in the empirical work of Froot and Stein (1991), Grosse and Trevino (1996), Caves (1989), Ray (1989), and Swenson (1994). Klein and Rosengren (1994) also ﬁnd real exchange rate and relative wealth to be signiﬁcant determinants of direct investment inﬂows in the U.S., but the effects of relative wage rates are insigniﬁcant. Choi and Jeon (2006) ﬁnd a signiﬁcant cointegrating relationship between real exchange rates and direct investment ﬂows of G-7 countries. Their multivariate cointegration analysis also establishes a linkage between direct investment ﬂows and several ﬁnancial variables including the cost of capital and real wealth in addition to real exchange rates. Healy and Palepu (1993), however, ﬁnd the coefﬁcients of both the contemporaneous and lagged changes in exchange rates to be insigniﬁcant on

Strategic and Financial Determinants of Foreign Direct Investments

25

FDIs. Stevens (1998) estimates the Froot and Stein (1991) model but ﬁnds only weak support for the relationship between FDIs and foreign exchange rates. Dewenter (1995), using the survey data of foreign acquisitions of U.S. targets, ﬁnds that, after controlling for relative corporate wealth, the relationship between foreign exchange rate level and foreign investment relative to domestic investment (acquisition activity) is statistically insigniﬁcant. Blonigen (1997) suggests that the mixed support for the relationship between FDIs and exchange rates may be related to ﬁrm-speciﬁc assets. If involved in international acquisitions, ﬁrm-speciﬁc assets can generate returns in currencies other than that used for purchase, and thus subject acquisitions to exchange risk. He reports that Japanese acquisitions in the U.S. are correlated with real dollar depreciation in industries that tend to have more ﬁrm-speciﬁc assets, such as manufacturing industries. Cushman (1985) maintains that the reduced cost of capital is a consequence of riskadjusted expected real currency appreciation of the source country, which in turn will stimulate FDIs. However, the net effect also depends on an increase in output prices versus input costs caused by exchange rate changes. That is, the foreign exchange effect is generally ambiguous depending on their effects through output and input markets. Cushman’s pooled estimation results show that FDIs are negatively related to an expected real appreciation of the source country currency and positively related to exchange rate volatility. This implies that the use of aggregate data may obscure the differential effects of foreign exchange rates on FDIs. The literature on ﬁnancial determinants of FDIs as discussed thus far has centered on the issues of differential valuation or cost. What has rarely been studied is the aspect that FDIs could also be triggered by agency behaviors. Recent studies on the diversiﬁcation discount, however, offer some parallel comparisons to this issue. Although they make no distinction between the types of diversiﬁcation (industrial versus global) in examining why managers diversify their ﬁrms, Aggarwal and Samwick (2003) argue that agents enjoy risk reduction and private beneﬁts from diversiﬁcation. It is the private beneﬁts such as improved career prospects, prestige, power, perquisites, compensation, and managerial entrenchment that in many cases drive managers to diversify. If this is the case, global diversiﬁcation theoretically could offer managers a greater degree of risk reduction and private beneﬁts. Click and Harrison (2000) ﬁnd a negative correlation between multinationality and management-owned equity and conclude that private gains are the reason for globalization and the global diversiﬁcation discount. All these are in line with our proposition that agency issues are one of the driving forces for FDIs.

26

JONGMOO JAY CHOI AND ERIC C. TSAI

In this paper, we intend to establish that ﬁnancial factors are as important as strategic ones in explaining FDIs. In fact, our results support FDIs as being driven by the interaction between internal capital market strength and internal corporate governance. We achieve this by examining ﬁrm-level data, which have not been used in existing literature on ﬁnancial determinants of FDIs. In contrast to existing work that focus on strategic factors only or those that focus on a single ﬁnancial variable, we suggest an integrated ﬁnancial model of FDIs that include agency issues as well as valuation, cost, exchange rates, and risk diversiﬁcation.

3. ESTIMATING FRAMEWORK The above discussion of existing literature suggests two different views regarding the determinants of FDIs. One hypothesis is strategic in that FDIs take place as a result of a ﬁrm’s strategic move to take advantage of market imperfections and differential costs in product and factor markets. This view is represented by the OLI model, which indicates that corporate FDIs increase with favorable location factors and strong ﬁrm-speciﬁc ownership endowments, and out of the efﬁciency of internal markets within an MNC setting. Even though the main focus in this paper is on the ﬁnancial determinants of FDIs, it is imperative to ﬁrst verify whether OLI theory maintains its validity in explaining FDIs in the 1990s. H1 (Strategic Hypothesis). Corporate FDIs are determined by strategic considerations pertinent to location factors, ownership endowments, and efﬁciency of internalization. The other is a ﬁnancial hypothesis, which depends on imperfections and asymmetries in capital markets. Given the imperfect integration of international ﬁnancial markets, MNCs may have an advantage because of their superior access to lower cost of ﬁnancing or in stable currency units. To the extent that there exist deviations from purchasing power parity, uncertain changes in exchange rates would have a real impact on international investments. In addition, given information asymmetry that favors internal over external ﬁnancing, exchange rates may inﬂuence the relative wealth of the ﬁrm, which may inﬂuence the propensity to acquire international ﬁrms. In sum, the existing literature on the ﬁnancial determinants of FDIs primarily pursues this line of inquiry, namely the interaction of foreign exchange rates, internal ﬁnancing, and corporate wealth on FDIs. Theoretically, although its scope is a bit narrow, it is plausible.

Strategic and Financial Determinants of Foreign Direct Investments

27

In light of the mainstream view of FDI as a strategic one, some additional remarks and justiﬁcations on the ﬁnancial perspectives may be in order. Our ﬁnancial model depicts that the combination of internal capital market strength and internal corporate governance factors motivates FDIs. First, to overcome their additional costs, the occurrence of FDIs requires internal capital market strength. A ﬁrm could actively take advantage of a relative corporate wealth increase due to foreign exchange rate movements. Of course, it also could be the result of corporate free cash ﬂow building up and the lack of domestic investment opportunity. Regardless of the reasons, however, a stronger internal capital market does provide MNCs an edge over the indigenous local ﬁrms in the host country. Given imperfect or informationally asymmetric capital markets, a depreciation of the host country currency effectively raises relative wealth of the foreign MNCs. Given the costly external ﬁnancing due to monitoring costs, an increased wealth implies an advantage of using cheaper internal source of capital. This ﬁnancing advantage puts MNCs in a better position than indigenous local ﬁrms during bidding competitions. It is a plausible explanation for the dominance of M&As as a mode of entry for FDIs. In comparison, the ﬁnancing advantage for greenﬁeld investments may be relatively less compelling outside the bidding scenario. Greenﬁeld investments would have a greater operational risk than M&As as well as the uncertainty in valuation. Nevertheless, with or without the inﬂuence from foreign exchange, imperfectly integrated international capital markets imply a differential access to ﬁnancing, which gives an advantage to one group of ﬁrms (i.e., MNCs) over the other (i.e., local indigenous ﬁrms). Second, internal capital market strength, however, does not necessarily lead to FDIs. Depending on managerial objectives and agency behaviors, it is up to the management to exploit the internal capital market strength for globalization, so FDI is also a consequence of internal corporate governance. One can argue that MNCs’ ﬁnancial advantages from internal capital market strength are short term in nature. That may well be, although there is no evidence that ﬁnancial effects are necessarily short-lived. It is well documented that, in the presence of agency costs and other ﬁnancial market imperfections, diversiﬁcation can provide a risk-reduction beneﬁt to the ﬁrm. International diversiﬁcation allows overall corporate risk to be lowered through reduction in ﬁnancial or business risk. If so, it may well be the compelling reason for the ﬁrm to initiate and sustain FDIs over the long run despite the added cost and risk associated with operating abroad. Even though risk reduction is a legitimate reason for cross-border investments, it is not at all farfetched that FDIs are conducted for the sake of

28

JONGMOO JAY CHOI AND ERIC C. TSAI

managers’ pursuit of private beneﬁts at the expense of shareholders and lead to unintended global diversiﬁcation. Earlier studies have suggested various agency problems pertinent to investments. For instance, investment can deviate from its optimal level due to agency problems, such as the underinvestment and substitution problems suggested by Myers (1977) and Jensen and Meckling (1976). Regardless of whether any given level of international investment is optimal, agency problems can still inﬂuence certain investment decisions. And it is not obvious that the agency problems are of short-term nature because they depend on long-standing and internationally dissimilar practices of corporate governance and industrial organization. At the higher level, managers’ pursuit of personal beneﬁts, such as gains in prestige, power, human capital, perquisites and compensation, and entrenchment, can be the predominant force driving cross-border investments. This, of course, directs us to believe that agency behavior could lead to FDI. Although FDI driven by agency behavior has not yet been explicitly proposed and tested, recent studies on the diversiﬁcation discount are consistent with this suggestion. Again, the existing studies on ﬁnancial determinants of FDIs extend their research primarily within the scope of factors related to internal capital markets. It is, of course, critical, but not necessarily a sufﬁcient condition. Our ﬁnancial model describes both the necessary and sufﬁcient conditions for FDI occurrence and provides a more comprehensive framework for explaining FDI behaviors from ﬁnance perspectives. In the present estimation, the ﬁnancial factors related to internal capital market strength include foreign exchange rate movement, corporate free cash ﬂow, and cost of capital (relevant if internal funding is insufﬁcient). The other ﬁnancial factors related to internal corporate governance are international diversiﬁcation, ﬁnancial and business risk, and agency cost. In short, FDI outﬂows may be affected positively by currency depreciation of the host country, internal ﬁnancing capacity of the ﬁrm, and beneﬁts from international diversiﬁcation, and negatively by agency cost and the cost of external ﬁnancing. H2 (Financial Hypothesis). Corporate FDIs are determined by ﬁnancial factors such as foreign exchange rate movements, ﬁnancing methods, corporate risks, international diversiﬁcation, and agency costs. We do not think that either hypothesis, standing alone, is sufﬁcient to fully explain actual direct investment decisions by the ﬁrm. We think that an exclusive emphasis on either hypothesis may be indicative of the ﬁeld of specialization of a researcher more than a complete depiction of reality. After all, it is undisputable that no major corporate decisions are being

Strategic and Financial Determinants of Foreign Direct Investments

29

made regarding international investments without considering both strategic and ﬁnancial dimensions of the project. More importantly, we consider both theories to be conceptually consistent. The mainstream strategic models are based on imperfections in goods and factor markets, which create an advantage for one class of ﬁrms as opposed to another. Similarly, if imperfections and information asymmetry exist in international capital markets, one class of ﬁrms may have a ﬁnancial advantage over another. Recourse to market imperfections of some kind is the same either way. In fact, Oxelheim, Stonehill, and Randoy (2001) classify ﬁnancial factors within the OLI framework. Thus, we state an integrated hypothesis that includes both strategic and ﬁnancial factors. H3 (Integrated Hypothesis). Corporate FDIs, in general, are determined by both strategic and ﬁnancial considerations jointly. The relative importance of these factors, however, may vary depending on industry sectors or the context in which investments are made. To summarize, we can characterize FDI in the following general functional form: FDI ¼ f ðStrategicfactors; Financialfactors; ControlvariablesÞ

(1)

where Strategic factors ¼ ½location factors; ownership endowments, internalization variables, Financial factors ¼ ½internal capital market factors, internal corporate governance factors or more specifically, Financial factors ¼ ½exchange rate; internal and external financial profile,

risk and diversification; agency cost.

Eq. (1) is estimated for a subset as well as for the entire sample. The speciﬁc list of variables used in the empirical work is discussed in the next section. Whether a strategic model is sufﬁcient to explain FDIs or whether ﬁnancial variables are needed in addition is an empirical issue. We use the ﬁrm-level panel data to investigate different hypotheses in various ways. Panel data analysis could reveal more information over time and crosssectionally. However, a pooled regression of all panel data across all ﬁrms and all time periods horizontally, FDI it ¼ a þ bDit þ it (where D is a vector

30

JONGMOO JAY CHOI AND ERIC C. TSAI

of ﬁnancial determinants, i indicates a ﬁrm, and t time), shows various degrees of autocorrelation. Hence, the results of pooled estimations are not reported in the interest of space. The identiﬁcation of correct functional form, however, might resolve this problem. Based on the panel data analysis methodologies developed, we also tried alternative functional forms, which include ﬁxed effect and random effect models, FDI it ¼ ai þ bDit þ it : These estimations assume a common slope but can take different assumptions on the distributions of the intercepts. However, heteroskedasticity is found to be present along with autocorrelation in some cases. To reduce these problems, we assume the stable relationship between the means of each explanatory variable for that is, we estimate FDI i ¼ a þ bDi þ i ; P individual ﬁrms,P where FDI i ¼ ðFDIit =TÞ; Di ¼ ðDit =TÞ; and T is the number of estimation periods.

4. DATA SOURCES AND DESCRIPTION OF VARIABLES Perhaps due to data convenience, the existing empirical work on the ﬁnancial determinants of FDIs is typically based on aggregate national or industry data, such as common macroeconomic measures for the cost of capital or exchange rates (e.g., U.S. dollar per SDR for exchange rate and three-month T-Bill rate for cost of capital). To test our proposed ﬁnancial hypothesis, ﬁrm-speciﬁc data are required, because the aggregate data can mask important differences that may exist at the level of ﬁrms. For instance, the effect of foreign exchange volatility on relative corporate wealth and cost of capital cannot be precisely measured, given the diversity of corporate international investment activities and dissimilarity of ﬁrm characteristics. This may help explain why the existing studies have mixed results. In our analysis, we use disaggregated ﬁrm-level data in corporate FDI, and explanatory variables such as ﬁrm-speciﬁc cost of capital as well as the exchange rate computed as a weighted average that reﬂects the extent and diversity of international investments by individual ﬁrms. We retrieve most of the data for sample ﬁrms from the Standard and Poor’s CompuStat Research Insight dataset. The criterion for inclusion in the sample is U.S. incorporations showing an ownership of foreign assets between 1992 and 1998.2 Table 2 shows descriptive characteristics for these ﬁrms. A total of 1,838 ﬁrms are selected, of which 1,131 (62%) are in manufacturing industries, and 1,551 (84%) and 1,014 (55%) had direct investments in DCs and less developed countries (LDCs), respectively. The

Strategic and Financial Determinants of Foreign Direct Investments

Table 2.

Descriptive Characteristics for the Sample of U.S. Corporations.

Total assets Foreign assets Foreign asset ratio (%) Total sales Foreign sales Foreign operating proﬁts EBIT Earnings volatility (%) Beta Debt ratio (%) Free cash ﬂow Cash reinvestment ratio (%) Agency cost (%) WACC (%)

31

Full Sample (N ¼ 1,838)

Manufacturing Industries (N ¼ 1,131)

Other Industries (N ¼ 621)

3307.00 (16360.00) 936.64 (5372.00) 24.43 (20.68) 2086.00 (8123.00) 659.60 (3704.00) 65.15 (287.00) 257.02 (1062.00) 9.45 (35.45) 0.64 (0.12) 21.55 (31.86) 130.09 (873.00) 4.03 (1.98) 38.79 (22.55) 11.32 (16.44)

2447.00 (13959.00) 810.72 (4548.00) 24.79 (19.27) 2114.00 (8768.00) 831.41 (4596.00) 79.99 (337.00) 228.77 (946.00) 9.25 (39.39) 0.57 (0.17) 20.49 (31.34) 142.26 (1026.00) 3.38 (1.48) 40.95 (20.21) 11.65 (18.89)

2232.00 (15741.00) 514.48 (5225.00) 24.21 (23.29) 1845.00 (6925.00) 333.17 (1039.00) 31.95 (153.00) 187.40 (1021.00) 9.61 (23.13) 1.14 (0.69) 21.64 (33.03) 79.57 (441.00) 5.14 (3.05) 36.85 (26.01) 10.99 (23.00)

Note: This descriptive statistics table shows means of each sample groups of U.S. corporations (standard deviations are in parentheses). Total assets, foreign assets, total sales, foreign sales, earnings before interest and taxes (EBIT), and free cash ﬂow are in millions of U.S. dollars. Free cash ﬂow comprises income before extraordinary items, depreciation, amortization, deferred taxes, equity in earnings of unconsolidated subsidiaries, extraordinary items and discontinued operations, and minority interest. Standard & Poor’s CompuStat Research Insight Database is the data source for all raw data that is the basis of computing other ratios. Foreign asset ratio is foreign assets divided by total assets. Earnings volatility is the standard deviation of EBIT scaled by the mean of total assets. Betas are computed by CompuStat. Debt ratio is long-term debts over total assets. Agency cost is measured by the ratio of operating expense to total sales. The weighted average cost of capital (WACC) is a weighted cost of equity (as measured by capital asset pricing model) and tax-adjusted cost of debt. Manufacturing industries comprise ﬁrms with SIC between 2,000 and 3,999. Other industries exclude these manufacturing ﬁrms as well as ﬁnancial ﬁrms with SIC between 6,000 and 6,999.

32

JONGMOO JAY CHOI AND ERIC C. TSAI

computation of several explanatory variables requires certain macroeconomic data, such as exchange rates, risk-free rate, and gross domestic products. They are obtained from the International Monetary Fund’s International Financial Statistics dataset. Finally, geographic distance is obtained from World Almanac and World Atlas. Table 3 summarizes the deﬁnitions of all variables, their speciﬁc data sources and expected signs, along with a correlation matrix for ﬁnancial variables. Beyond that, a brief discussion and some justiﬁcation for the variables selected and their anticipated signs are in order. 4.1. Dependent Variable To analyze ﬁnancial determinants of FDI, ﬁrm-level data representing FDI activities are essential. Another relevant issue is the use of FDI ﬂows or FDI stocks. The distinction is crucial but the choice of FDI stocks is obvious: FDI ﬂows reﬂect the sum invested in afﬁliates by foreign companies during a speciﬁc time period, which afﬁliates may spend to accumulate ﬁxed and current assets, to repay past borrowings, or for some other objectives. However, the FDI stock represents the total value of assets attributable to the foreign investor. Thus, the FDI stock is an indicator for the value of assets engaged in international production. (Stephan & Pfaffmann, 2001).

Conventional FDI data collected by government agencies in each country (e.g., U.S. Department of Commerce) or international institutions (e.g., IMF, OECD, or UNCTAD) are available for sectors and regions with some degree of disaggregation, but not to the ﬁrm level. Fortunately, the availability of corporate foreign assets in the CompuStat dataset provides a direct measure of FDI stock. In this study, the dependent variable, FDIi, is foreign assets of ﬁrm i in natural logarithm form, which is consistent with existing FDI studies such as in Miller and Reuer (1998), where foreign assets are used as an explanatory variable in examining the effect of FDI on exchange exposure of the ﬁrm, and FDI in natural logarithm form serves its intended purposes in Buckley, Clegg, Forsans, and Reilly (2001). Independent variables include strategic and ﬁnancial variables as well as control variables. 4.2. Strategic Variables The strategic framework for FDI tested in this paper follows the mainstream OLI model. As such, the selection of these variables is in accordance with empirical OLI studies. An obvious location factor is production cost (PC).

Deﬁnitions and Expected Signs of Variables and Correlation Matrix for Financial Variables.

Variable Name

Description

Expected Sign

Panel A: Summary of Variable Definitions and Expected Signs Dependent variable FDIi (foreign direct Foreign assets of ﬁrm i associated with foreign operations identiﬁed as ‘‘Identifiable investment) Assets’’ in the Geographic Segment of CompuStat, in natural logarithm Strategic variables PCi (production cost) TCi (transportation cost)

XARi (fixed asset ratio) RDRi (R&D ratio) FSRi (foreign sales ratio) IARi (intangible asset ratio) SERi (sales expense ratio) Financial variables WFXi,t (weighted foreign exchange composite)

Cost of goods sold divided by total sales of ﬁrm i Weighted average of geographic distance in miles (in natural logarithm): X FI j TC i ¼ ln Dj ; TF I i j

+ +

where FIj is foreign investment (foreign assets) in country j; TFIi,t is total foreign direct investment (total foreign assets) of ﬁrm i in a particular year; Dj is the distance in miles between U.S. and country j where the ﬁrm operates in the speciﬁc year Fixed assets scaled by total assets of ﬁrm i Research and development expense as a percent of total sales of ﬁrm i Foreign sales divided by total sales of ﬁrm i Intangible assets divided by total assets of ﬁrm i

+ + + +

Sales expense divided by total expense of ﬁrm i

+

Weighted multilateral foreign exchange rate change for ﬁrm i in year t: X FI j;t ej;t WF X i;t ¼ ln ; TF I t ej;t1 j

33

where FIj is foreign assets in country j; TFIt is total foreign direct investment stock of ﬁrm i in year t; ej,t (home currency per foreign currency) is the annual bilateral foreign exchange rate between U.S. and country j in which the ﬁrm operates in year t

Strategic and Financial Determinants of Foreign Direct Investments

Table 3.

34

Table 3. (Continued ) Variable Name IFi (internal financing)

ACEi (agency cost expense ratio) ACAi (agency cost asset utilization ratio) Control variables Si (firm size) CONi (concentration ratio) USW (U.S. wealth) HCWi (host country wealth)

Free cash ﬂow of ﬁrm i including income before extraordinary items, depreciation, amortization, deferred taxes, equity in earnings of unconsolidated subsidiaries, extraordinary items and discontinued operations, and minority interest Weighted average cost of equity and tax-adjusted cost of debt of ﬁrm i The number of geographic segments in which ﬁrm i operates Long-term debt ratio representing ﬁrm i’s capital structure The standard deviation of ﬁrm i’s earnings before interests and taxes (EBIT) scaled by the mean of its total assets to avoid size bias Operating expense over total sales of ﬁrm i

Expected Sign +

+ +/ /+ +

Total sales divided by total assets of ﬁrm i

Total sales of ﬁrm i in natural logarithm Total sales of four largest ﬁrms in the industry of ﬁrm i divided by total sales of the industry The gross domestic products (GDP) of the U.S. in natural logarithm Weighted average GDP (in natural logarithm) of all host countries in which the ﬁrm operates: X FI j HCW i ¼ ln GDPj ; TF I i j

+

where FIj is the foreign assets in country j. TFIt is the level of total foreign direct investment of ﬁrm i in the particular year

JONGMOO JAY CHOI AND ERIC C. TSAI

EFi (external financing) IDi (international diversification) FRi (financial risk) BRi (business risk)

Description

IFi

EFi

IDi

FRi

BRi

ACEi

ACAi

Panel B: Correlation Matrix for Financial and Control Variables Financial Variables IFi 0.06 1.00 (o.01) EFi 0.02 0.01 1.00 (0.09) (0.42) IDi 0.06 0.20 0.01 1.00 (o.01) (o.01) (0.21) FRi 0.03 0.21 0.04 0.06 1.00 (0.02) (o.01) (o.01) (o.01) BRi o.01 0.25 0.01 0.01 0.04 (0.91) (o.01) (0.90) (0.42) (o.01) ACEi o.01 0.02 0.01 0.02 0.03 (0.99) (0.13) (0.90) (0.06) (o.01) ACAi 0.02 0.19 0.04 0.04 0.11 (0.09) (o.01) (o.01) (0.01) (o.01)

0.03 (o.01) 0.01 (0.43)

0.05 (o.01)

1.00

Control Variables Si 0.09 (o.01) CONi 0.03 (0.02)

0.09 (o.01) 0.01 (0.47)

0.13 (o.01) 0.02 (0.05)

0.10 (o.01) 0.04 (o.01)

0.75 (o.01) 0.06 (o.01)

0.02 (0.02) 0.01 (0.88)

0.31 (o.01) 0.06 (o.01)

0.27 (o.01) 0.03 (o.01)

Si

CONi

1.00 1.00

1.00 0.05 (o.01)

1.00

USW

Strategic and Financial Determinants of Foreign Direct Investments

WFXi

35

USW HCWi

36

Table 3. (Continued ) WFXi

IFi

EFi

IDi

FRi

BRi

ACEi

ACAi

Si

CONi

USW

0.15 (o.01) 0.30 (o.01)

0.12 (o.01) 0.04 (o.01)

0.01 (0.70) 0.01 (0.29)

0.11 (o.01) 0.03 (0.02)

0.01 (0.36) 0.06 (o.01)

0.03 (o.01) 0.01 (0.84)

0.01 (0.23) 0.04 (o.01)

0.03 (o.01) 0.05 (o.01)

0.07 (o.01) 0.05 (o.01)

0.26 (o.01) 0.10 (o.01)

1.00 0.11 (o.01)

JONGMOO JAY CHOI AND ERIC C. TSAI

Note: Panel B shows Pearson correlation coefﬁcients between variables. In parentheses is the p value under H0: r ¼ 0, and o.01 indicates the p value is less than 0.01. Data sources: FDI is available from the CompuStat Geographic Segment database. CompuStat deﬁnes Identifiable Assets in a geographic segment or country as the tangible and intangible assets that are used by, or directly associated with, each geographic segment or country. The seven geographic segments as classiﬁed by CompuStat are Africa, Asia, Europe, Paciﬁc, South America, North America, and Other Foreign Operations. CompuStat only identiﬁes 10 countries with heaviest U.S. foreign investments, South Africa, Japan, Philippines, Great Britain, France, Germany, Australia, Brazil, Canada, and Mexico. To compute TC, geographic distance in miles is obtained from World Almanac and World Atlas. However, if some of a ﬁrm’s foreign operation is identiﬁed only in a geographic segment or only as foreign, for that portion of investment, the average distance in miles to that region or the total average miles is used. ID is available from the CompuStat Geographic Segment database. CompuStat reports the actual number of geographic segments disclosed by the ﬁrms. To compute WFX, bilateral exchange rates are obtained from International Financial Statistics (IFS) dataset. However, if some of a ﬁrm’s foreign operation is identiﬁed only in a geographic segment or only as foreign, for that portion of investment, the average exchange rate change in that segment or the rate of USD per SDR is used, respectively. To compute USW and HCW, GDPs are obtained from IFS dataset by IMF. In calculating HCW, however, if some of a ﬁrm’s foreign operation is identiﬁed only in a geographic segment or only as foreign, for that portion of investment, the average GDP of the countries in that region or the average world GDP is used. To calculate all other strategic, ﬁnancial, and control variables, the accounting data are retrieved from CompuStat dataset.

Strategic and Financial Determinants of Foreign Direct Investments

37

Some empirical work in location analysis (e.g., Culem, 1988; Kravis & Lipsey, 1982; Maki & Meredith, 1986) uses wage rates or unit labor cost. However, several researchers, such as Huang (1997), argue that the labor cost may not be an accurate measure of location cost since the proportion of labor cost in the total production cost has been decreasing. Another location factor, transportation cost (TC), is frequently measured in empirical work (e.g., Grosse & Trevino, 1996) by geographic distance. As for ownership factors, the ﬁxed asset ratio (XAR) is in line with the notion that corporations with higher operating leverage tend to have a greater ownership advantage in expanding their foreign production due to an oligopolistic edge such as economies of scale. Quite common in related work (e.g., Dunning, 1980; Pugel, Kragas, & Kimura, 1996), research and development expense ratio (RDR) measures a ﬁrm’s ownership advantage through its investment in knowledge production and technology know-how, which give the ﬁrm a competitive edge in the global market. Internalization variables comprise foreign sales ratio (FSR), measuring the degree of foreign involvement; intangible asset ratio (IAR), measuring the level of information-based assets; and sales expense ratio (SER), measuring operating efﬁciency. Internalization gains through FDIs derive from a higher level of informationbased assets, a greater degree of foreign involvement, and improving operational efﬁciency. IAR gauges the need for internalization, because internalization reduces loss of efﬁciency and uncertainty in the valuation of intangible or information-based assets given the market imperfection for such assets in the external markets.3 FSR represents a corporation’s degree of foreign involvement in many empirical studies (e.g., Click & Harrison, 2000; Duru & Reeb, 2001).4 Overall, conventional FDI theories, OLI in particular, would suggest positive effects on FDI and thus positive expected signs for these strategic variables. More favorable location, ownership, and internalization conditions motivate ﬁrms to step up their cross-border direct investments. 4.3. Financial Variables ‘‘Internal’’ capital market factors include foreign exchange rate, free cash ﬂow, and cost of capital. In particular, the weighted foreign exchange composite (WFX) takes into account different FDI compositions of ﬁrms, and a negative sign for its coefﬁcient is anticipated. Host country currency depreciation encourages FDI due to stronger internal capital markets for MNCs. Financing method variables include those that capture the proﬁle of internal ﬁnancing (IF) and external ﬁnancing (EF). Consistent with existing

38

JONGMOO JAY CHOI AND ERIC C. TSAI

literature in corporate ﬁnance, free cash ﬂow represents corporate wealth or internal ﬁnancing capacity. We expect a positive sign for it, because the buildup of free cash ﬂow (stronger internal capital market) offers ﬁrms an edge in the international bidding scenario. We estimate cost of capital for each ﬁrm and a negative sign is anticipated since, if internal funding is insufﬁcient, a higher cost of external ﬁnancing deters foreign investments.5 ‘‘Internal’’ corporate governance factors are related to managerial objectives or agency behaviors that lead to cross-border expansion and diversiﬁcation. They include international diversiﬁcation (ID), an overall measure for global diversiﬁcation; ﬁnancial risk (FR) and business risk (BR), measures of intended diversiﬁcation for risk reduction; and expense ratio (ACE) and asset utilization ratio (ACA), measures for agency costs. International diversiﬁcation (ID) measures a ﬁrm’s intention to spread its operations globally and thereby achieve risk-reduction beneﬁts of a more diverse operating cash ﬂow. It should be noted that even when diversiﬁcation is not a primary objective, any international investment—whether intended or not—has a diversiﬁcation dimension. We measure the extent of international diversiﬁcation by the number of geographic segments in which a ﬁrm has FDIs (foreign assets). Other than country count, it is most frequently used in geographic diversiﬁcation studies (e.g., Bodnar, Tang, & Weintrop, 1999; Duru & Reeb, 2001). Overall, there should exist a positive relationship between international diversiﬁcation and FDI. More speciﬁcally, the intent of global diversiﬁcation is to reduce risk in either business or ﬁnancial dimension. As is common in the ﬁnance literature, we use longterm debt ratio, representing the ﬁrm’s capital structure, as a measure of ﬁnancial risk. Business risk is measured by earnings volatility, the standard deviation of a ﬁrm’s earnings before interests and taxes but scaled by the mean of its total assets to avoid size bias. Given a certain level of total risk a ﬁrm can or will assume, it is unclear whether ﬁrms are more interested in diversifying ﬁnancial or business risk. However, it appears that the goal of reducing a particular type of risk through FDI commands a positive sign for its coefﬁcient. Furthermore, since they work as substitutes, the other one would have a negative coefﬁcient. Regardless of its signiﬁcance, the type of risk with a negative sign may be simply irrelevant in a ﬁrm’s diversiﬁcation decision because the negativity is a consequence of the ﬁrm assuming a certain level of total risk. As for agency cost, a commonly used proxy for agency cost in the ﬁnance literature (e.g., Ang, Cole, & Lin, 2000) is the operating expense ratio (ACE), measuring how well the management controls operating costs including expenditure on excessive perquisites and

Strategic and Financial Determinants of Foreign Direct Investments

39

other direct agency costs.6 Another agency cost measure is the asset utilization ratio (ACA) measuring how well ﬁrms deploy and utilize their assets. Generally, low asset utilization ratio, or high expense ratio, indicates the presence of a more severe agency problem. If private beneﬁts encourage managers to diversify their ﬁrms globally, ACE is expected to be positive and ACA negative. In the interest of parsimony, however, ACE is our primary measure of agency cost in estimation. These ﬁnancial variables are selected in accordance with existing empirical literature as well as econometric considerations. However, the effect of agency behavior on FDI has not been tested at all in existing literature. Furthermore, a comprehensive analysis of ﬁnancial determinants of FDIs based on ﬁrm-level data has not yet been done. It is imperative to verify that the collection of ﬁnancial variables in our estimating framework is appropriate in a certain statistical sense. Panel B of Table 3 shows the correlation matrix for ﬁnancial variables. The null hypothesis of no correlation cannot be rejected at any signiﬁcance level for most of the pairs. For those few rejected, the magnitude of correlation is only about 0.02 or less. 4.4. Control Variables We use control variables at different levels: ﬁrm size (S) at the ﬁrm level, concentration ratio (CON) at the industry level, and the U.S. (USW) and host country wealth (HCW) at the country level. These variables take into account a ﬁrm’s resources, competition in the industry, and wealth of home and host countries. Given different patterns of global asset deployment among ﬁrms, the computation of certain variables, including WFX, TC, and HCW, is necessary to reﬂect each ﬁrm’s distinctive foreign investment activities. Their weighting schemes are identical as detailed in Table 3.

5. EMPIRICAL RESULTS 5.1. Strategic Factors Table 4 shows the estimation results of the strategic model. Most of the strategic variables in the OLI model are statistically signiﬁcant and have the expected signs. For the full sample, except for sales expense ratio, all variables are signiﬁcant at the 1% level and have the expected signs.7 Thus, overall the strategic motives for FDI, as indicated by the OLI model, appear

40

Table 4.

Strategic Determinants of Foreign Direct Investments FDIi ¼ a þ b1 PCi þ b2 TCi þ b3 XARi þ b4 RDRi þ b5 FSRi þ b6 IARi þ b7 SERi þ b8 Si þ b9 CONi þ b10 USW þ b11 HCWi þ i

Full Sample (N ¼ 1,838)

C (intercept) Strategic variables Location variables PC (production cost) TC (transportation cost) Ownership variables XAR (ﬁxed asset ratio) RDR (R&D ratio)

5.63 (2.39)

Manufacturing Industries (N ¼ 1,131) 2.20 (0.88)

Other Industries (N ¼ 621) 17.67 (2.94)

DCs (N ¼ 1,551) 10.02 (3.70)

LDCs (N ¼ 1,014)

2.63 (0.55)

0.18 (9.43) 0.34 (5.74)

0.17 (8.67) 0.27 (4.17)

0.53 (2.90) 0.55 (4.07)

0.21 (9.85) 0.00004 (2.57)

0.01 (4.20) 0.0001 (2.53)

0.73 (4.36) 0.07 (3.92)

0.71 (3.80) 0.15 (2.91)

0.66 (1.74) 0.04 (1.55)

1.03 (7.04)

1.27 (5.22)

JONGMOO JAY CHOI AND ERIC C. TSAI

The dependent variable, foreign direct investment (FDI), is corporate foreign asset in natural logarithm. Corporate foreign asset is deﬁned as identiﬁable asset (including both tangible and intangible assets) used by or directly associated with foreign operations outside the U.S.

IAR (intangible asset ratio) SER (sales expense ratio) Control variables S (ﬁrm size) CON (four-ﬁrm concentration ratio) USW (U.S. GDP) HCW (weighted host country GDPs) Adjusted R2 LM heteroskedasticity test (H0: homoskedasticity)

2.82 (27.76) 0.96 (5.33) 0.29 (1.66)

3.06 (26.59) 1.11 (5.82) 0.21 (1.09)

2.21 (10.07) 0.14 (0.30) 0.23 (0.56)

2.52 (21.19) 0.96 (4.99)

1.88 (9.46) 0.16 (0.44)

0.98 (81.19) 0.002

0.99 (71.24) 0.005

0.94 (35.24) 0.008

0.99 (76.32) 0.004

0.93 (44.77) 0.001

(0.98) 0.04 (0.17) 0.04 (1.71)

(1.72) 0.28 (1.04) 0.05 (1.74)

(1.95) 1.14 (1.71) 0.008 (0.13)

(2.18) 0.70 (2.31) 0.11 (2.99)

(0.35) 0.32 (0.06) 0.17 (3.96)

0.91 4.94a

0.92 5.20a

0.87 0.41c

0.83 2.52c

0.73 3.82b

Note: LM heteroskedasticity test follows w2 distribution with one degree of freedom. Manufacturing industries comprise ﬁrms with SIC between 2,000 and 3,999. Other industries exclude these manufacturing ﬁrms as well as ﬁnancial ﬁrms with SIC between 6,000 and 6,999. DCs and LDCs refer to foreign direct investments in developed and less developed countries, respectively. The classiﬁcation of developed and developing countries is according to United Nations Conference on Trade and Development. Signiﬁcant at 1% level. Signiﬁcant at 5% level. Signiﬁcant at 10% level. The t-statistics are in parentheses. a The null hypothesis of homoskedasticity is rejected at the 5% level, but not rejected at the 1% level. b The null hypothesis of homoskedasticity is rejected at the l0% level, but not rejected at the 5% level. c The null hypothesis of homoskedasticity is not rejected at any of the conventional 10%, 5%, or 1% level.

Strategic and Financial Determinants of Foreign Direct Investments

Internalization variables FSR (foreign sales ratio)

41

42

JONGMOO JAY CHOI AND ERIC C. TSAI

to be strong during the sample period of 1992–1998. Similarly, all OLI variables are signiﬁcant for manufacturing industries at the 1% level except for sales expense ratio, with additional signiﬁcance shown for three control variables (ﬁrm size, concentration ratio, and host country GDP). In contrast, other industries (non-manufacturing/non-ﬁnancing) have their FDI most directly related to location variables. Two out of three internalization variables (intangible asset ratio and sales expense ratio) and one ownership variable (R&D ratio) are insigniﬁcant for the other industry category. In addition, the other ownership variable is only marginally signiﬁcant. This makes sense because compared to manufacturing ﬁrms, non-manufacturing ﬁrms do not produce products that are heavily embedded with ﬁrm-speciﬁc ownership assets. They possess less ownership advantage and thus have a lesser need for internalization. It appears that location is the major concern in their FDI decisions. Table 4 further shows the behavioral pattern of FDI by U.S. ﬁrms depending on destinations, DCs versus LDCs. There does not exist distinctive patterns between FDI in DCs and in LDCs.8 Again, conventional FDI theories, internalization included, fail to explain the phenomenon that most FDIs occur between DCs. 5.2. Financial Factors Table 5 on the ﬁnancial model of FDI indicates the general validity of the model. The results for the full sample show all ﬁnancial variables having the expected signs at a high signiﬁcance level except for foreign exchange, which is signiﬁcant for the manufacturing industries and investments in DCs. However, compared to the strategic model, there is a greater variation in the results depending on industry sector and host country. The overall results, nevertheless, support our proposition that FDI is the consequence of the interaction between a stronger internal capital market and internal corporate governance. In sum, home currency appreciation, accumulation of corporate free cash ﬂow, managerial objectives for risk reduction, and agency behavior, collectively, are the driving forces for FDI. Models with various combinations of variables, as shown in Table 6, are also estimated on the full sample and two conclusions may be drawn. First, except for foreign exchange, the results including signs and signiﬁcance for the variables vary little across different equations. Clearly, as indicated in the previous correlation matrix, these variables do not interact with each other much and our results are relatively robust. Second, the foreign exchange variable has a much greater variability in its signiﬁcance across the equations but always maintains the correct sign. It seems that the effect of

Financial Determinants of Foreign Direct Investments FDIi ¼ a þ b1 WFXi þ b2 IFi þ b3 EFi þ b4 IDi þ b5 FRi þ b6 BRi þ b7 ACEi þ b8 ACAi þ b9 Si þ b10 CONi þ b11 USW þ b12 HCWi þ i

The dependent variable, foreign direct investment (FDI), is corporate foreign asset in natural logarithm form. Corporate foreign asset is deﬁned as identiﬁable asset used by or directly associated with foreign operations outside the U.S. Full Sample (N ¼ 1,838)

C (intercept)

0.82 (0.28)

Other Industries (N ¼ 621)

DCs (N ¼ 1,551)

LDCs (N ¼ 1,014)

2.53 (0.39)

7.47 (2.18)

1.91 (0.35)

0.47 (1.61) 0.13 (5.89) 0.17 (2.40)

0.52 (1.92) 0.07 (2.83) 0.14 (2.04)

0.93 (0.96) 0.19 (3.86) 0.05 (0.22)

2.02 (3.75) 0.12 (4.62) 0.02 (0.14)

0.09 (0.28) 0.13 (3.32) 0.30 (2.60)

0.43 (10.39)

0.42 (9.41)

0.50 (5.80)

0.18 (3.61)

0.32 (4.25)

43

Financial variables ‘‘Internal’’ capital market WFX (FDI-weighted foreign exchange composite) IF (internal ﬁnancing: corporate free cash ﬂow) EF (external ﬁnancing: cost of capital) ‘‘Internal’’ corporate governance ID (international diversiﬁcation)

3.27 (1.16)

Manufacturing Industries (N ¼ 1,131)

Strategic and Financial Determinants of Foreign Direct Investments

Table 5.

44

Table 5. (Continued )

FR (ﬁnancial risk: capital structure) BR (business risk: earnings volatility) ACE (agency cost: expense ratio)

CON (four-ﬁrm concentration ratio) USW (U.S. GDP) HCW (weighted host country GDP) Adjusted R2 LM heteroskedasticity test (H0: homoskedasticity)

Manufacturing Industries (N ¼ 1,131)

Other Industries (N ¼ 621)

DCs (N ¼ 1,551)

LDCs (N ¼ 1,014)

0.55 (4.81) 0.92 (2.14) 0.07 (4.67)

0.35 (2.99) 0.66 (1.47) 0.09 (6.37)

0.53 (2.02) 0.63 (0.71) 0.29 (1.28)

0.69 (5.10) 1.03 (2.12) 0.10 (5.73)

0.24 (1.12) 1.65 (1.93) 0.19 (0.84)

0.85 (33.28) 0.0004 (0.16) 0.08 (0.27) 0.08 (2.28)

0.94 (34.31) 0.01 (3.76) 0.37 (1.13) 0.06 (1.44)

0.68 (12.14) 0.0004 (0.09) 0.07 (0.09) 0.08 (1.00)

0.85 (27.64) 0.003 (0.94) 0.37 (0.97) 0.32 (6.04)

0.96 (20.49) 0.006 (1.29) 0.22 (0.36) 0.26 (5.19)

0.79 5.32a

0.84 5.32a

0.68 4.45a

0.73 2.30b

0.69 5.29a

Note: LM heteroskedasticity test follows w2 distribution with one degree of freedom. Manufacturing industries comprise ﬁrms with SIC between 2,000 and 3,999. Other industries exclude these manufacturing ﬁrms as well as ﬁnancial ﬁrms with SIC between 6,000 and 6,999. DCs and LDCs refer to foreign direct investments in developed and less developed countries, respectively. The classiﬁcation of developed and developing countries is according to United Nations Conference on Trade and Development. Signiﬁcant at 1% level. Signiﬁcant at 5% level. Signiﬁcant at 10% level. The t-statistics are in parentheses. a The null hypothesis of homoskedasticity is rejected at the 5% level, but not rejected at the 1% level. b The null hypothesis of homoskedasticity is not rejected at any of the 10%, 5%, or 1% conventional level.

JONGMOO JAY CHOI AND ERIC C. TSAI

Control variables S (ﬁrm size)

Full Sample (N ¼ 1,838)

Further Testing on Financial Determinants of Foreign Direct Investments FDIi ¼ a þ b1 WFXi þ b2 IFi þ b3 EFi þ b4 IDi þ b5 FRi þ b6 BRi þ b7 ACEi þ b8 ACAi þ b9 Si þ b10 CONi þ b11 USW þ b12 HCWi þ i

The dependent variable, foreign direct investment (FDI), is corporate foreign asset in natural logarithm form. Corporate foreign asset is deﬁned as identiﬁable asset used by or directly associated with foreign operations outside the U.S. Full Sample (N ¼ 1,838) (1) C (intercept)

0.43

3.27 (1.16)

(1.83)

0.47 (1.61) 0.13 (5.89) 0.17 (2.40)

1.30 (3.45) 0.62 (28.94) 0.24 (2.43)

0.43 (10.39)

0.73 (13.09)

(3)

(4)

(5)

(6)

(7)

(9.13)

2.83 (0.96)

6.11 (2.21)

0.03 (1.67) 0.03 (0.51)

0.69 (2.52) 0.13 (5.89) 0.16 (2.20)

0.46 (1.59) 0.13 (5.92) 0.17 (2.40)

0.58 (1.90) 0.12 (5.38) 0.20 (2.61)

0.66 (2.32) 0.13 (6.09) 0.06 (0.93)

0.39 (10.83)

0.43 (10.27)

0.43 (10.40)

0.77 (0.31)

2.99 (17.37)

2.50

0.43 (10.88)

45

Financial variables ‘‘Internal’’ capital market WFX (FDI-weighted foreign exchange composite) IF (internal ﬁnancing: corporate free cash ﬂow) EF (external ﬁnancing: cost of capital) ‘‘Internal’’ corporate governance ID (international diversiﬁcation)

(2)

Strategic and Financial Determinants of Foreign Direct Investments

Table 6.

46

Table 6. (Continued ) Full Sample (N ¼ 1,838) (1) FR (ﬁnancial risk: capital structure) BR (business risk: earnings volatility) ACE (agency cost: expense ratio) ACA (agency cost: asset utilization)

CON (four-ﬁrm concentration ratio) USW (U.S. GDP) HCW (weighted host country GDP) Adjusted R2 LM heteroskedasticity test (H0: homoskedasticity)

1.08

(4.81) 0.92 (2.14) 0.07 (4.67)

(6.95) 3.91 (6.91) 0.24 (3.79)

0.85 (33.28) 0.0004 (0.16) 0.08 (0.27) 0.08 (2.28) 0.79 5.32a

0.59 0.51b

(3)

(4)

0.22

0.56

0.55

0.50

(4.95) 1.00 (2.34) 0.07 (4.69)

(4.85) 0.92 (2.15) 0.07 (4.67)

(4.16) 0.77 (1.72) 0.09 (5.40)

0.96 (43.17) 0.001 (0.53) 0.11 (0.42) 0.09 (3.26)

0.85 (33.36)

0.85 (33.39)

0.08 (2.29)

0.91 (34.99) 0.0003 (0.11) 0.16 (0.49) 0.06 (1.61)

0.87 (36.09) 0.001 (0.59) 0.38 (1.25) 0.09 (2.72)

0.83 6.56a

0.79 6.31a

0.79 5.36a

0.77 1.28b

0.79 4.24a

(2.34) 0.06 (4.40) 0.69 (17.68)

(5)

Note: LM heteroskedasticity test follows w2 distribution with one degree of freedom.

Signiﬁcant at 1% level. Signiﬁcant at 5% level. Signiﬁcant at 10% level. The t-statistics are in parentheses. a

The null hypothesis of homoskedasticity is rejected at the 5% level, but not rejected at the 1% level. The null hypothesis of homoskedasticity is not rejected at any of the 10%, 5%, or 1% conventional level.

b

(6)

(7) 0.49 (4.80) 0.07 (4.34)

JONGMOO JAY CHOI AND ERIC C. TSAI

Control variables S (ﬁrm size)

(2)

0.55

Strategic and Financial Determinants of Foreign Direct Investments

47

foreign exchange on FDI is consistent but its importance is lessened with the presence of some factors. To have a more comprehensive discussion on the proposition in this paper, we shall discuss the results of each ﬁnancial factor individually both on the full sample and across different industry sectors and destinations. 5.2.1. Foreign Exchange The impact of the foreign exchange effect on FDI is via home currency appreciation, creating a stronger internal capital market for MNCs. In full sample estimations, the FDI-weighted foreign exchange composite variable has statistically signiﬁcant coefﬁcients with negative signs in most cases, meaning that a depreciation in the foreign currency or an appreciation in the U.S. dollar results in an increase in outward FDI by U.S. ﬁrms. This is also true for ﬁrms in the manufacturing industry, but the effect is insigniﬁcant for the ‘‘other’’ industries. These results are consistent with the view of Blonigen (1997), who argues that the exchange rate effect depends on ﬁrm-speciﬁc assets, which are more pronounced in manufacturing ﬁrms than in nonmanufacturing ﬁrms. In the context of acquisition, ﬁrm-speciﬁc assets can generate returns in currencies other than that used for purchase, and thus subject the acquiring ﬁrm to exchange risk. Hence, the exchange rate effect on FDI is more signiﬁcant for manufacturing than for non-manufacturing ﬁrms. It is also interesting that there are signiﬁcant differences in exchange rate impacts depending on destinations. The results for FDI in DCs are the same as those for manufacturing industry, i.e., outward FDI by U.S. ﬁrms is positively impacted by a strong dollar. However, the exchange rate is not a signiﬁcant factor for FDI in LDCs. This is attributable to the fact that many currencies of the LDCs are effectively ﬁxed or managed by the government. That is, there is little variability in observed data and hence little observed exchange rate impact. However, it does not mean that there is no risk or no risk management problem in developing countries, only that the risk is postponed and resolved ultimately by an eruption of crises or crashes as happened in the Asian ﬁnancial crisis of 1997. In addition, DCs have more ﬁrms with advanced technologies or information-based assets. In the case of international acquisitions, ﬁrm-speciﬁc assets can be relevant to the effect of foreign exchange. Another point worth mentioning for the full sample estimations in Table 6 is that some factors seem to overshadow foreign exchange. It appears that, without controlling for ﬁrm size, foreign exchange is a very important factor (Eq. (2)). However, after controlling for ﬁrm size and host

48

JONGMOO JAY CHOI AND ERIC C. TSAI

country GDP and with the presence of international diversiﬁcation and business risk, the exchange rate is not a factor for FDI (Eqs. (1) and (5)). One possible explanation for its lesser importance for large ﬁrms is that they tend to have more resource and expertise in managing foreign exchange risk and may need less help from home currency appreciation to improve their internal capital markets. Furthermore, when large ﬁrms consider various factors simultaneously (e.g., international diversiﬁcation, business risk, and host country wealth), foreign exchange seems to take a back seat in the FDI decision. These two reasons, foreign exchange being overshadowed by some other factors and the link between ﬁrm-speciﬁc assets and foreign exchange, may explain why the results of foreign exchange in the existing FDI studies are mixed. 5.2.2. Internal and External Financing Internal ﬁnancing capacity shows the strength of a corporation’s internal capital market, but external ﬁnancing becomes relevant when internal funding is insufﬁcient. Internal ﬁnancing or corporate wealth is usually measured in empirical work by corporate free cash ﬂow. With an increase in free cash ﬂow, a ﬁrm is likely to conduct more FDI, especially if it runs out of domestic investment opportunities. If a ﬁrm requires external ﬁnancing, typically measured by its cost of capital, the higher rate undoubtedly hinders its ability to make cross-border investments. Both these ﬁnancing effects, positive for internal and negative for external ﬁnancing, are conﬁrmed for the full sample, but internal ﬁnancing tends to be more signiﬁcant than external ﬁnancing. These results are consistent with the view of Froot and Stein (1991) that internal ﬁnancing is more important than external ﬁnancing in international acquisitions. The internal ﬁnancing variable is signiﬁcantly positive both across all industry sectors and regardless of destinations. The effects of external ﬁnancing are more selective. The coefﬁcients are signiﬁcant for manufacturing industries but insigniﬁcant for ‘‘other’’ industries. Utility ﬁrms (which are included in the other category) would raise equity capital to support initial investments, and given relatively stable income, their need for external ﬁnancing might be relatively small on an ongoing basis. Manufacturing ﬁrms, in contrast, emphasize growth and the requirement of a higher degree of operating leverage could mean less free cash ﬂow during the expansion stage. As a result, external ﬁnancing is relatively more important.9 In addition, the internal ﬁnancing variable is signiﬁcant for FDI in both DCs and LDCs. However, the coefﬁcient of external ﬁnancing is signiﬁcant

Strategic and Financial Determinants of Foreign Direct Investments

49

only in LDCs. This appears to be related to country risk. Given the complexity of FDI in LDCs, more funding, management expertise, experience in international business, and resource base for risk management may be required.10 It is entirely likely that ﬁrms with FDIs in LDCs must also rely on external ﬁnancing. Another possibility, although further investigation is necessary, is the corporate concern for country risk or mere government regulation. As a means of minimizing their country risk exposure, multinational ﬁrms may prefer partial local ﬁnancing for their investments in LDCs. In addition, some host country governments may limit the percentage of ownership by foreign investors. 5.2.3. International Diversification and Risk Reduction International diversiﬁcation has both intended and unintended dimensions. The intended global diversiﬁcation, in general, is regarded as the consequence of risk reduction in that operations are spread out over various countries or regions. The international diversiﬁcation variable is highly signiﬁcant and has the expected positive sign across the board – for the full sample, for all industry sectors, and for both DCs and LDCs. Most international investments, as it seems, are motivated by the desire to diversify globally for ﬁnancial or business risk-reduction purposes. Of course, it is also possible that, for some investments, such a result is unintended. That is, even if diversiﬁcation does not ﬁgure prominently in terms of motivation, these international investments, nonetheless, achieve a degree of geographical diversiﬁcation. The underlying goal, however, could have a serious effect if managers are set to pursue their own personal gains at the expense of shareholders, which is to be discussed in the next section. In Table 6, ﬁnancial risk and business risk variables are both signiﬁcant and have exactly opposite signs as expected in all the equations in the full sample estimation. This is due to the fact that, to an extent, both risks work as substitutes. For a given amount of total risk a ﬁrm is willing or able to take, a ﬁrm with lower business risk can afford to increase its ﬁnancial risk, or vice versa. However, since ﬁnancial risk has a positive coefﬁcient with greater signiﬁcance, it appears that the target for risk reduction, if intended, is ﬁnancial risk. Furthermore, for both manufacturing and other industries, ﬁnancial risk is signiﬁcant but business risk is not. Another plausible interpretation of this result is that risk diversiﬁcation may apply to ﬁnancial risk rather than business risk. This interpretation is consistent with the ﬁnding of Burgman (1996), who shows that international diversiﬁcation does not necessarily lower earnings volatility – a measure of business risk – for MNCs.

50

JONGMOO JAY CHOI AND ERIC C. TSAI

Furthermore, ﬁnancial risk is relevant only for FDIs in DCs but not for those in LDCs. Apparently, ﬁrms consider DCs the better destination for their ﬁnancial risk diversiﬁcation. This is not surprising because most international acquisitions are ﬁnanced by cash or equity swaps. DCs have well-developed capital markets that facilitate fair equity valuation. The notion that diversifying ﬁnancial risk through investments in DCs would lower MNCs’ debt ratio over time is consistent with results of Lee and Kwok (1988) and Fatemi (1988), who suggest that MNCs have a lower debt ratio than domestic ﬁrms. 5.2.4. Agency Cost Agency problem could lead to unintended global diversiﬁcation as well as potential ﬁrm value destruction. Of the two measures for agency cost, the expense ratio is the primary proxy in the interest of parsimony. However, in the full sample estimation, positive expense ratio and negative asset utilization ratio coefﬁcients point in the same direction, a positive link between agency costs and FDI – more severe agency problems leading to higher FDI. Given the strong signiﬁcant and consistent results on the full sample, it is not at all difﬁcult to conclude that FDI is also driven by agency behavior – the pursuit of managers’ private beneﬁts. This is obviously consistent with the ﬁndings in recent diversiﬁcation discount studies. For instance, Click and Harrison (2000) establish the inverse relationship between multinationality and managerial equity ownership, and Duru and Reeb (2001) ﬁnd managers diversify their ﬁrms in part due to their private beneﬁts. It is also in line with the ﬁnding of Lee and Kwok (1988) that MNCs tend to have higher agency cost.11 The grouping results in Table 5 indicate that, after controlling for ﬁrm size, the agency problem exists only for manufacturing industries and for FDI in DCs. One plausible explanation is that the agency problem is more severe for corporations with more ﬁrm-speciﬁc or information-based assets. Given the relative abundance of these kinds of assets in manufacturing ﬁrms, in particular, such ﬁrms face more complex potential agency conﬂicts. These assets also make corporations less transparent and create more difﬁculty for corporate governance and monitoring. This explanation is in line with the ﬁnding in Harris and Ravenscraft (1991) that cross-border takeovers are more likely to take place in R&D intensive industries (more likely manufacturing). The result on DCs can be similarly understood. DCs have more advanced manufacturing ﬁrms (with greater ﬁrm-speciﬁc and information-based assets) than LDCs. Firms with such assets are preferable targets for managers of acquiring ﬁrms who pursue their own personal gains.

Strategic and Financial Determinants of Foreign Direct Investments

51

Overall, the ﬁnancial factors appear to be important in FDI decisions in addition to strategic factors. Given imperfections and information asymmetry in international capital markets, corporate FDIs are inﬂuenced by such ﬁnancial variables related to internal capital markets and internal corporate governance. However, relative to strategic variables, there appears to be greater variability in the level of signiﬁcance across industry sectors and destinations of investments for ﬁnancial variables. In addition, all these ﬁnancial variables are signiﬁcant in the full sample estimations, but ﬁnancial considerations on FDI appear to be more critical for manufacturing than for other industries and more essential for FDI in developed than in LDCs.12 5.3. An Integrated Model To put the relative importance of strategic and ﬁnancial factors into perspective, we now combine these two sets of variables, in addition to control variables, in an integrated empirical model. Estimation results of an integrated model generally preserve the previous results obtained in each of the two models, especially for the full sample and FDI in DCs and LDCs.13 On the full sample, Table 7 shows that ﬁnancial factors are as important as strategic factors for FDI by U.S. ﬁrms, with virtually all the variables signiﬁcant. The table shows the same result – an equal importance of ﬁnancial and strategic factors – for manufacturing. The performance of the integrated model declines a bit for other industries, but here strategic variables appear to be much more important than ﬁnancial variables. The integrated model performs better for U.S. FDIs in DCs than for those in LDCs. However, there is no clear evidence in terms of the relative importance of strategic and ﬁnancial variables in either region. We also calculated Schwartz’s Bayesian criterion for the three representative models: strategic, ﬁnancial, and integrated.14 The Bayesian criterion should be as small as possible because the addition of an independent variable would raise this score if there is no incremental explanatory power. By this criterion, the strategic model performs better than the ﬁnancial model for other industries. For the manufacturing industries, both models perform just about equally. The integrated model works best for the full sample. We did not pursue this line of inquiry in depth in this paper because, technically, it is contingent on the selection of a representative strategic or ﬁnancial model. We are content, in this paper, with establishing the basic point that ﬁnancial variables are necessary in addition to strategic variables in understanding FDI decisions fully, and leave the exact determination of their relative contribution for future research.

52

Table 7.

Strategic and Financial Determinants of Foreign Direct Investments

FDIi ¼ a þ b1 WFXi þ b2 IFi þ b3 EFi þ b4 IDi þ b5 FRi þ b6 BRi þ b7 ACEi þ b8 PCi þ b9 TGCDi þ b10 XARi þ b11 RDRi þ b12 FSRi þ b13 IARi þ b14 SERi þ b15 ACAUi þ b16 Si þ b17 CONi þ b18 USW þ b19 HCWi þ i

Full Sample (N ¼ 1,838)

C (intercept) Financial variables WFX (weighted foreign exchange) internal capital market variable IF (internal ﬁnancing) internal capital market variable EF (external ﬁnancing) – internal capital market variable ID (international diversiﬁcation) – internal corp. governance var. FR (ﬁnancial risk) – internal corporate governance variable

4.32 (1.88) 0.19 (0.63) 0.07 (3.85) 0.24 (3.99) 0.14 (3.90) 0.24 (2.49)

Manufacturing Industries (N ¼ 1,131) 0.83 (0.35) 0.09 (0.31) 0.07 (3.52) 0.33 (5.68) 0.16 (4.15) 0.24 (2.48)

Other Industries (N ¼ 621)

DCs (N ¼ 1,551)

LDCs (N ¼ 1,014)

13.16 (2.40)

1.71 (0.61)

4.37 (0.82)

0.96 (1.16) 0.08 (1.82) 0.02 (0.14) 0.08 (1.07) 0.25 (1.06)

0.23 (0.48) 0.05 (2.22) 0.12 (1.37) 0.12 (2.75) 0.24 (2.02)

0.10 (0.28) 0.04 (0.86) 0.45 (3.50) 0.60 (7.72) 0.08 (0.40)

JONGMOO JAY CHOI AND ERIC C. TSAI

The dependent variable, FDI, is corporate foreign asset in natural logarithm. Corporate foreign asset is identiﬁable asset, tangible or intangible, used by or directly associated with foreign operations occurring outside the U.S.

0.80 (2.35) 0.12 (4.16)

0.62 (1.75) 0.03 (0.95)

0.91 (1.23) 0.29 (1.06)

1.65 (4.13) 0.23 (6.40)

1.78 (1.92) 0.02 (0.04)

Strategic variables PC (production cost) – location variable TC (transportation cost) – location variable XAR (ﬁxed asset ratio) – ownership variable FSR (foreign sales ratio) – internalization variable IAR (intangible asset ratio) – internalization variable SER (sales expense ratio) – internalization variable

0.27 (2.74) 0.20 (3.71) 1.59 (10.28) 2.90 (24.90) 0.83 (4.24) 0.22 (1.34)

0.06 (0.57) 0.19 (3.36) 0.97 (5.44) 2.96 (23.29) 0.97 (4.66) 0.38 (2.00)

1.17 (2.94) 0.31 (2.54) 2.02 (5.90) 2.89 (12.22) 0.47 (1.10) 0.15 (0.43)

0.57 (4.98) 0.12 (1.85) 1.35 (6.99) 2.91 (20.70) 1.33 (5.85) 0.45 (2.19)

1.34 (2.61) 0.32 (2.21) 2.14 (5.64) 2.59 (9.87) 0.14 (0.30) 0.48 (1.20)

0.89 (39.93) 0.002 (1.00) 0.02 (0.06) 0.02 (0.71)

0.91 (38.44) 0.01 (2.72) 0.53 (2.07) 0.07 (2.07)

0.89 (16.97) 0.003 (0.92) 0.82 (1.37) 0.09 (1.34)

0.90 (33.57) 0.003 (1.06) 0.12 (0.39) 0.01 (0.21)

0.99 (20.72) 0.002 (0.36) 0.28 (0.48) 0.21 (4.00)

0.88

0.92

0.84

0.77

Control variables S (ﬁrm size) CON (four-ﬁrm concentration ratio) USW (U.S. GDP) HCW (weighted host country GDP) Adjusted R2

0.82

Strategic and Financial Determinants of Foreign Direct Investments

BR (business risk) – internal corporate governance variable ACE (agency cost: expense ratio) – internal corp. governance var.

53

54

Table 7. (Continued )

LM heteroskedasticity test (H0: homoskedasticity)

Full Sample (N ¼ 1,838)

Manufacturing Industries (N ¼ 1,131)

Other Industries (N ¼ 621)

DCs (N ¼ 1,551)

LDCs (N ¼ 1,014)

5.10a

1.12c

3.82b

0.41c

6.40a

JONGMOO JAY CHOI AND ERIC C. TSAI

Note: LM heteroskedasticity test follows w2 distribution with one degree of freedom. Other industries exclude manufacturing and ﬁnancial ﬁrms. DCs and LDCs refer to FDIs in developed and less developed countries, respectively. Signiﬁcant at 1% level. Signiﬁcant at 5% level. Signiﬁcant at 10% level. The t-statistics are in parentheses. a The null hypothesis of homoskedasticity is rejected at the 5% level, but not rejected at the 1% level. b The null hypothesis of homoskedasticity is rejected at the l0% level, but not rejected at the 5% or 1% level. c The null hypothesis of homoskedasticity is not rejected at any of the conventional 10%, 5%, or 1% level.

Strategic and Financial Determinants of Foreign Direct Investments

55

6. CONCLUDING REMARKS The mainstream FDI literature emphasizes strategic factors. This almost exclusive emphasis on strategic factors does not seem to be justiﬁed in practice. We introduce ﬁnancial factors in addition to strategic factors as determinants of FDI. The estimation based on 1,838 U.S. ﬁrms for the period from1992 to 1998 indicates that the traditional OLI strategic factors are in fact signiﬁcant in explaining FDI behavior in full sample estimations. The results do not differ much across destinations, although evidence suggests that strategic considerations have been more important for manufacturing than other industries. At the same time, evidence strongly indicates the importance of ﬁnancial factors as well. The relevance of ﬁnancial variables in FDI decisions is based on the assumption of imperfections in ﬁnancial markets. Hence, the ﬁnancial model of FDI shares the same assumption – market imperfections of some kind – with the mainstream FDI literature in the industrial organization framework. Estimations show that almost all ﬁnancial variables identiﬁed – foreign exchange rate, proﬁle of internal and external ﬁnancing, risk and diversiﬁcation variables, and agency costs – are signiﬁcant for the full sample, for ﬁrms in manufacturing industries, and for investments in DCs. In general, these results support the basic notion underlying our ﬁnancial model that FDI is the consequence of interaction between a stronger internal capital market and internal corporate governance. International diversiﬁcation and internal ﬁnancing variables, in particular, are important in all estimations regardless of destinations or industries. FDI in other industries is dependent only on the ﬁrm’s ﬁnancial risk in addition to these two variables. The fact that FDI made by manufacturing industries (but not other industries) is signiﬁcantly affected by foreign exchange appears to be related to the presence of ﬁrm-speciﬁc assets. Beyond the two common ﬁnancial determinants – international diversiﬁcation and internal ﬁnancing – an additional list of ﬁnancial variables is signiﬁcant for FDI in DCs: foreign exchange, ﬁnancial risk, and agency cost. For LDCs, only external ﬁnancing is relevant in addition to internal ﬁnancing and diversiﬁcation. This reﬂects a different degree of development of capital markets and country risk in developed and developing countries. If one has to choose, there is a sense in which, overall, the strategic factors are the slightly more important considerations in FDI decisions. However, we also show that ﬁnancial factors are important in FDI decisions in addition to strategic factors. Empirically, our ﬁnancial model works quite well

56

JONGMOO JAY CHOI AND ERIC C. TSAI

by itself or in conjunction with the OLI factors. Given imperfections and information asymmetry in international capital markets, generally, FDIs are inﬂuenced by such ﬁnancial variables as foreign exchange, internal and external ﬁnancing, risk and diversiﬁcation, and agency problems of the concerned parties. However, relative to strategic variables, there appears to be greater variability in the level of signiﬁcance of ﬁnancial variables across industry sectors and across investment destinations. Most importantly, however, our ﬁnancial model is better than the mainstream OLI framework in one sense – in that it is able to explain the prevailing FDI phenomena when OLI factors fail. First of all, why is the most popular FDI entry mode consistently and predominantly M&As rather than greenﬁeld investments? It could be explained by the strength of the internal capital market and the consequence of internal corporate governance. Host country currency depreciation or corporate free cash ﬂow buildup leads to a stronger internal capital market. It favors foreign acquisitions because the edge from currency movement could be short-lived and there is no reason not to use up the excess capacity from free cash ﬂow quick, thus eliminating the inefﬁciency. Typically, greenﬁeld investment is a longer-term project to plan, deploy, and complete. Agency behavior also favors foreign acquisitions simply because it is easier and quicker for managers to gauge and materialize their potential gains from acquiring an established ﬁrm than by building one from scratch. Second, why most FDIs occur between DCs can similarly be explained from ﬁnancial perspectives. A stronger internal capital market facilitated by host country currency devaluation is more likely to happen in DCs than in LDCs because the foreign exchange rate oftentimes is ﬁxed or managed by LDC governments. International acquisitions driven by agency behavior also favor DCs, where there resides a larger set of ideal targets for maximizing managers’ personal beneﬁts. Another relevant point is that the popularity of M&A as the primary entry mode implies that DCs are the ideal destinations, as M&A is more likely to occur in a country having well-developed capital markets and legal systems. In addition, our evidence also shows that DCs are the destinations for ﬁrms to diversify their ﬁnancial risk through foreign investments. Finally, ﬁnancial factors are logical candidates to explain the often-observed short-term bilateral FDI swings between two countries. After all, ﬁnancial variables such as foreign exchange, free cash ﬂow, and cost of capital could vary signiﬁcantly in a short period of time. Ever changing managerial objectives regarding global diversiﬁcation and agency behavior in pursuing private beneﬁts also assure that FDI is not likely to be a constant toward a particular country for an extended period of time.

Strategic and Financial Determinants of Foreign Direct Investments

57

NOTES 1. For instance, in 1998, 91.6% of all FDI outﬂows were from developed countries and 71.5% of all FDI inﬂows were to developed countries. Source: World Investment Report, 1999, United Nations Conference on Trade and Development (UNCTAD). 2. Foreign assets are deﬁned as identiﬁable assets, which include both tangible and intangible assets, used by or directly associated with operations abroad by U.S. ﬁrms. 3. Intangible assets in CompuStat dataset include a variety of patents, rights and agreements, licenses, goodwill, designs and drawings, franchises, quotas, subscription and client lists, and trademark. 4. One legitimate concern here is the potential problem of simultaneity. Granted, foreign sales are generated through both domestic and foreign productions. However, internalization focuses on the need for MNCs to switch from export (domestic production) to FDI (foreign production) to improve efﬁciency. If the sole purpose of FDI is to boost foreign sales, total sales go up too. FSR does not necessarily change signiﬁcantly, especially if the ratio is relatively large. Therefore, simultaneity may be of little concern because the variable is measured by ratio rather than level. 5. The cost of equity is estimated by the capital asset pricing model, where the market risk beta is estimated for each ﬁrm based on the monthly data for the previous 60-month period, which is reported by CompuStat. The weights are based on the market value of equity and the book value of debt. 6. In comparison, the agency cost in Myers (1977) is advertising and R&D expenses divided by total sales. We did not use this because it blurs the distinction with internalization or ownership advantage. R&D expenses are often used as a measure of internalization or ownership factor. 7. Although the intercept is signiﬁcant, it is negative, which is simply an indication that the starting point of FDI is relatively small given the dependent variable in natural logarithm form. The same goes for other equations with signiﬁcant and negative intercepts. 8. Choi, Kim, and Chandran (1995) provide a summary of literature regarding ‘‘conventional’’ and ‘‘non-conventional’’ FDIs by multinationals from developed and developing countries, respectively. 9. As Table 2 shows, manufacturing industry has a lower cash reinvestment ratio – roughly 60% of that for ‘‘other’’ industry. 10. Not surprisingly, the signiﬁcance of external ﬁnancing on FDIs made by manufacturing ﬁrms and in LDCs is consistent with ﬁrm size effect (S). The size effects are greater for both groups than their counterparts. 11. To entertain classic agency theories, the agency problem related to FDIs is likely to be a substitution problem due to foreign exchange risk and country risk involved in international investments. If a foreign project is riskier than a comparable domestic project, the agency problem is more of the substitution problem as illustrated in Jensen and Meckling (1977) rather than underinvestment problem as posited in Myers (1978). The substitution problem describes the tendency of the management to substitute a riskier for a safer project for the interest of certain parties (in this case, their own).

58

JONGMOO JAY CHOI AND ERIC C. TSAI

12. Incidentally, the ﬁrm size has a positive impact on FDI for the full sample and across all industry sectors and destinations. However, there is evidence that the absolute magnitudes of the size coefﬁcients are somewhat larger for manufacturing than for other industries, indicating the scale of resources necessary for FDI. The size effects are greater for LDCs than for DCs, suggesting a notion that operations in LDCs generally require a larger resource base for risk management and experience in international business. 13. A couple of different variables does lose their signiﬁcance for manufacturing and other industries and weighted foreign exchange also becomes insigniﬁcant for manufacturing ﬁrms and FDIs in DCs. However, this is not considered severe, given that all other results are consistent with previous estimations and that foreign exchange is overshadowed by ﬁrm size and few other variables. 14. Schwartz’s Bayesian criterion is believed to be superior to Akaike’s information criterion because it is asymptotically consistent for large sample (see Wei, 1990).

REFERENCES Aggarwal, R. K., & Samwick, A. A. (2003). Why do managers diversify their ﬁrms? Agency reconsidered. Journal of Finance, 58, 1613–1649. Aliber, R. A. (1970). A theory of foreign direct investment. In: C. Kindleberger (Ed.), The international corporation. Cambridge: MIT Press. Ang, J. S., Cole, R. A., & Lin, J. W. (2000). Agency costs and ownership structure. Journal of Finance, 55, 81–106. Black, F. (1978). The ins and outs of foreign investment. Financial Analysts Journal, 34, 1–7. Blonigen, B. A. (1997). Firm-speciﬁc assets and the link between exchange rates and foreign direct investment. American Economic Review, 87, 447–465. Bodnar, G. M., Tang, C., & Weintrop, J. (1999). Both sides of corporate diversification: The value impacts of geographic and industrial diversification. Working Paper. Buckley, P. J., & Casson, M. C. (1976). The future of multinational enterprise. London: Macmillan. Buckley, P. J., Clegg, J., Forsans, N., & Reilly, K. T. (2001). Increasing the size of the ‘‘country’’: Regional economic integration and foreign direct investment in a globalised world economy. Management International Review, 41, 251–274. Burgman, T. A. (1996). An empirical examination of multinational corporate capital structure. Journal of International Business Studies, 27, 553–570. Casson, M. (1979). Alternatives to the multinational enterprise. London: Macmillan. Caves, R. E. (1971). International corporations: The industrial economics of foreign investment. Economica, 38, 1–27. Caves, R. E. (1989). Exchange-rate movements and foreign direct investment in the United States. In: D. B. Audretsch & M. P. Claudon (Eds), The internationalization of U.S. markets. New York: New York University Press. Choi, J. J. (1989). Diversiﬁcation, exchange risk, and corporate international investment. Journal of International Business Studies, 20, 145–155. Choi, J. J., & Jeon, B. N. (2006). Financial factors in foreign direct investments: A dynamic analysis of international data. Research in International Business and Finance, (In Press, Corrected Proof).

Strategic and Financial Determinants of Foreign Direct Investments

59

Choi, J. J., Kim, K., & Chandran, R. (1995). Foreign direct investments by ﬁrms from developing and developed countries: Stylized facts and theoretical interpretations. In: Y. Kim & K. Oh (Eds), The US–Korea economic partnership. London: Avebury. Click, R. W., & Harrison, P. (2000). Does multinationality matter? Evidence of value destruction in U.S. multinational corporations. Working Paper. Culem, C. G. (1988). The locational determinants of direct investments among industrialized countries. European Economic Review, 32, 885–904. Cushman, D. O. (1985). Real exchange rate risk, expectations, and the level of direct investment. Review of Economics and Statistics, 67, 297–308. Dewenter, K. (1995). Do exchange rates drive foreign direct investment? Journal of Business, 68, 405–433. Dunning, J. H. (1977). Trade, location of economic activity and the MNE: A search for an eclectic approach. In: B. Ohlin (Ed.), The international allocation of economic activity. London: Macmillan. Dunning, J. H. (1980). Toward an eclectic theory of international production: Some empirical tests. Journal of International Business Studies, 11, 9–31. Dunning, J. H. (1988). The eclectic paradigm of international production: A restatement and some possible extensions. Journal of International Business Studies, 19, 1–31. Duru, A., & Reeb, D. M. (2001). Evidence on the relation between geographic diversification and firm value. Working Paper. Fatemi, A. M. (1988). The effect of international diversiﬁcation on corporate ﬁnancing policy. Journal of Business Research, 16, 17–30. Froot, K. A., & Stein, J. C. (1991). Exchange rates and foreign direct investment. Quarterly Journal of Economics, 106, 1191–1217. Grosse, R., & Trevino, L. J. (1996). Foreign direct investment in the United States: An analysis by country of origin. Journal of International Business Studies, 27, 139–155. Harris, R. S., & Ravenscraft, D. (1991). The role of acquisitions in foreign direct investment: Evidence from the U.S. stock market. Journal of Finance,, 46, 825–844. Healy, P. M., & Palepu, K. G. (1993). International corporate equity associations: Who, where, and why? In: K. A. Froot (Ed.), Foreign direct investment. Chicago, IL: University of Chicago Press. Huang, G. (1997). Determinants of United States–Japanese foreign direct investment: A comparison across countries and industries. New York: Garland. Hymer, S. H. (1976). The international operations of national firms: A study of direct foreign investment. Cambridge, MA: MIT Press. Itaki, M. (1991). A critical assessment of the eclectic theory of the multinational enterprise. Journal of International Business Studies, 22, 445–460. Jensen, M. C., & Meckling, W. H. (1976). Theory of ﬁrm managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. Kindleberger, C. P. (1969). American business abroad; six lectures on direct investment. New Haven, CT: Yale University Press. Klein, M. W., & Rosegren, E. S. (1994). The real exchange rate and foreign direct investment in the United States: Relative wealth vs. relative wage effects. Journal of International Economics,, 36, 373–389. Kravis, I. B., & Lipsey, R. E. (1982). The location of overseas production and production for export by US multinational ﬁrms. Journal of International Economics, 12, 201–223.

60

JONGMOO JAY CHOI AND ERIC C. TSAI

Lee, K., & Kwok, C. (1988). Multinational corporations vs. domestic corporations’ international environmental factors and determinants of capital structure. Journal of International Business Studies,, 19, 195–217. Maki, D. R., & Meredith, L. N. (1986). Production cost differentials and foreign direct investment: A test of two models. Applied Economics, 18, 1127–1134. Miller, K. D., & Reuer, J. J. (1998). Firm strategy and economic exposure to foreign exchange rate movements. Journal of International Business Studies, 29, 493–514. Myers, S. C. (1977). Determinants of corporate borrowing. Journal of Financial Economics, 5, 147–175. Oxelheim, L., Stonehill, A., & Randoy, T. (2001). On the treatment of ﬁnance speciﬁc factors within the OLI paradigm. International Business Review, 10, 381–398. Pugel, T. A., Kragas, E. S., & Kimura, Y. (1996). Further evidence on Japanese direct investment in U.S. manufacturing. Review of Economics and Statistics, 78, 208–213. Ray, E. J. (1989). The determinants of foreign direct investment in the United States, 1979–85. In: R. C. Feenstra (Ed.), Trade policies for international competitiveness. Chicago, IL: University of Chicago Press. Stephan, M., & Pfaffmann, E. (2001). Detecting the pitfalls of data on foreign direct investment: Scope and limit s of FDI data. Management International Review, 41, 189–218. Stevens, G. V. G. (1998). Exchange rates and foreign direct investment: A note. Journal of Policy Modeling, 20, 393–401. Stulz, R. (1983). On the determinants of net foreign investment. Journal of Finance, 38, 459–468. Swenson, D. L. (1994). Impact of U.S. tax reform on foreign direct investment in the United States. Journal of Public Economics, 54, 243–266. Wei, W. W. S. (1990). Time series analysis: Univariate and multivariate methods. Redwood City, CA: Addison-Wesley.

MODELING THE EVOLUTIONARY SEQUENCE OF INTERNATIONAL JOINT VENTURES Elmar Lukas ABSTRACT This paper contributes to the literature on foreign direct investment under uncertainty and underscores the importance of modeling the evolutionary sequence pattern of foreign market entry. The model presented helps to refine the application of real options in the international joint venture context by providing a closed-form solution in continuous time to value their overall strategic flexibility. Moreover, the analysis provides a novel perspective on existing empirical results and generates a number of new testable predictions.

1. INTRODUCTION As a result of an increasingly volatile economic environment, one of the main changes in the nature of multinational enterprises’ (MNEs) operations is related to the institutional choice of achieving access to foreign markets. Unlike identifying situations where a MNE can achieve majority control,

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 61–73 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07003-8

61

62

ELMAR LUKAS

globalization forces have affected the need for ﬁrms to cooperate. Due to increased competition and the rapid pace of technological development, inter-company collaboration is now crucial to the pursuit of competitive advantage. As a consequence, the number of international joint ventures (IJVs), both horizontal (between competing companies) and vertical (between companies within a given value chain), has risen dramatically during the past two decades, especially when ﬁrms operate in highly uncertain industries (e.g., Broll & Marjit, 2005; Dunning & Narula, 2004; Todeva & Knoke, 2005). The increase is, however, not only driven by achieving access to markets but by seeking foreign tangible or intangible assets as well. Hence, sourcing of complementary assets internationally favors the choice of IJVs and acts as an ever-growing driving force for international collaboration in general (e.g., Hagedoorn, 2002). A joint venture (JV) is a legal organization formed by two or more parties to undertake jointly economic activity for mutual beneﬁt. One can broadly distinguish between two basic organizational modes: equity JV and nonequity JV. The former is created when each partner has an equity share in the new venture. Non-equity JVs, in contrast, are agreements to cooperate in some way, but they do not involve the creation of new ﬁrms.1 A JV becomes international if at least one partner has its headquarter outside the venture’s country of operation or if the venture has a signiﬁcant level of operation in more than one country (Geringer & Hebert, 1991). The search for determinants that drive the use and success of IJVs has a long tradition in the business and economics literature.2 The research conducted was primarily empirically in nature (e.g., Blodgett, 1992; Gatignon & Anderson, 1988; Hennart, Kim, & Zeng, 1998; Kogut & Chang, 1996; Leung, 1997). While being extensively analyzed empirically, less effort, however, has been made in scrutinizing the properties of IJVs through rigorous theoretical modeling (e.g., Buckley & Casson, 1996; Marjit, Broll, & Mallick, 1995). Further, all attempts do not account for the fact that a ﬁrm’s commitment to invest into a new market is associated with sunk costs that cannot be recovered once the project is initiated. Moreover, foreign direct investment (FDI) decisions are to a large portion investment decisions under uncertainty and are only the ﬁrst commitment of subsequent expansion. Hence, with respect to the initial switching decision, i.e., whether to abandon export or not, one has additionally to consider the impetus of subsequent expansion (Gilroy & Lukas, 2006). Another criticism stems results from the fact that these models only consider the unidirectional case. Thus, they lack explanation of divestment or strategic reorientation (e.g., Buckley & Tse, 1996; Rivoli & Salorio, 1996).

Modeling the Evolutionary Sequence of International Joint Ventures

63

In the last decade, researchers have highlighted the importance of a more dynamic perspective in FDI theory. Consequently, real options theory has gained signiﬁcant attention in the international business ﬁeld because it not only captures dynamic aspects of decision theory but also represents an advance in conceptualizing and measuring the value of strategy under uncertainty (e.g., Chi & Seth, 2002). In brief, real options theory suggests viewing real investments as options buying the ﬁrm the rights such as to make investments later, the right to defer or alter scale or to initiate subsequent investments.3 Besides others, Buckley and Casson (1998) have drawn attention to this by arguing that the existing models do value FDI decisions only with respect to their immediate effects rather than in terms of possible new investment opportunities. In other words, in most cases initial foreign production serves as a platform for expansion abroad, indicating that the initial investments carry a high option value due to possible new investment opportunities, e.g., regional technology platforms (e.g., Chang & Rosenzweig, 2001; Howells & Wood, 1993; Kogut & Chang, 1996). It is clear that this fact is most obvious for IJVs and it is Kogut (1991) who puts this thought further. Possible project interdependencies within the IJV allow for strategic ﬂexibility, calling for an interpretation of IJVs as platform investments. Thus, although unproﬁtable from a stand-alone perspective, the value of a JV can be much higher due to the ﬂexibility to acquire later stakes of the venture in the future. Consequently, the termination of an IJV does not indicate its failure but the exploitation of its ﬂexibility. Based on a two-stage binomial model, Chi and McGuire (1996) model a situation in which the MNE has the option to acquire or sell out the partner’s stake in an equity JV.4 They depict that both options create economic value for the partners, especially if the partners foresee different valuation expectations of the venture ex post and in the presence of transactions costs. By treating two sources of uncertainty explicitly in their model, the authors were also able to show how the options serve to diminish the risk of misappropriation and thus alleviate the difﬁculty of JV contracting under information asymmetry. Besides the theoretical analysis, the authors also present a number of testable hypotheses related to their work. In a more advanced model setting, Pennings and Sleuwaegen (2004) design an option model where both the timing of market entry and the entry mode are determined simultaneously. The switch from export, whether to a wholly owned subsidiary, a JV, or to licensing, is dependent on uncertainty of payoffs, cost structure, competitive stance of incumbents, tax differences, and the degree of cooperation between the JV partners. In particular, the timing of a JV is related to transfer prices, amount of equity share, market

64

ELMAR LUKAS

structure, and to the degree of governmental regulation. However, possible subsequent actions were neglected. While the option idea has become a building block for empirical research on IJVs lately (e.g., Folta & Miller, 2002; Reuer & Leiblein, 2000; Reuer & Tong, 2005), in-depth research is still lacking in the international business literature with respect to the consequent modeling of the evolutionary sequence pattern of an IJV. Thus, the goal of this paper is to model a JV induced market entry under uncertainty in a continuous time setting. The rest of the paper is structured as follows. In Section 2, we will present the model: a two-phase market entry sequence. The main results are presented in Section 3, which also provides a synopsis of major comparative-static results. Finally, Section 4 summarizes the main ﬁndings and provides suggestions for further research.

2. THE MODEL This paper focuses on a representative equity-based JV between two private ﬁrms. Both ﬁrms combine complementary resources comprising ﬁrmspeciﬁc knowledge, which is either related to technology or marketing expertise or both. For the sake of simplicity, it is assumed that only a subset of overall knowledge is shared, however, in an amount that secures the agreed objective, such as R&D collaboration. Moreover, it is assumed that only one ﬁrm is a foreigner to the new market, namely the MNE, which has chosen a local partner in the host country. The choice of which entry strategy an enterprise chooses has no inﬂuence on the proﬁt rates of other enterprises in the foreign market. Moreover, the value of the chosen FDI mode v(t) is ex ante unknown and follows a geometric Brownian motion. Assuming a perfect capital market, the existence of a unique martingale measure Q can be used to modify the stochastic differential equation, which results in dv=v ¼ ðr dÞ dt þ s dZ Q

(1) 2

where rd is the growth rate of the project value, s designates the variance of dv/v, r is the risk-free interest rate, d represents the opportunity cost of waiting, and dZQ indicates a Wiener process with non-zero drift. Let e refer to the initial equity stake the MNE investor has invested in and ¯ be the country-speciﬁc upper boundary for owning equity stakes held by foreign investors in a given country.5 Consequently, this initial strategy generates two exclusive strategic options. However, it is worthwhile to

Modeling the Evolutionary Sequence of International Joint Ventures

65

consider a time span in which the partners become acquainted and can check if joint work is possible for the sake of the venture. We account for this by assuming that the MNE has a certain time period [0,t1] in which it can decide how to continue with its market entry strategy. At the end of this period of length T, the MNE can decide whether it prefers to continue collaboration with the host partner by receiving the right to convert the IJV into a cross-border merger and acquisition (M&A), i.e., by acquiring the remaining shares ð¯ Þ at a later date. Counterbalancing this, the MNE might prefer the right to dissolve the IJV over the growth option by selling its own interest e to the local partner at a later date.6 Let us denote the optimal threshold separating both strategies with c. The option value F, the optimal trigger points vU, vL (representing the actual timing of the subsequent investment/divestment), and c may be solved recursively. From Dixit and Pindyck (1994) as well as Merton (1973) the results for a perpetual investment option and a perpetual divestment option, respectively, are commonly known. Thus, they are just summarized brieﬂy. Under the assumption of a perpetual time to maturity and aforementioned boundary conditions the solution for a perpetual call option results in ( Avb1 vovU CðvÞ ¼ (2) ð¯ Þv I v vU with I as the corresponding cost for acquiring the rest of the equity stake ð¯ Þ; ð1b1 Þ ð¯ Þ 1 b1 I A¼ and b1 ð¯ Þ b1 1 sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 ðr dÞ ðr dÞ 1 2 2r þ þ 2 b1 ¼ 2 s2 s2 2 s as constants.7 From this, the optimal trigger value vU for the M&A strategy can be deduced, which results in vU ¼

1 b1 I ð¯ Þ b1 1

(3)

On the other hand, if the MNE decides to dissolve the IJV it will receive a perpetual put option. To be more precise, on exercising the second stage the MNE gives up an existing project with value ev and receives its abandonment value k (see, e.g., Chi, 2000). The value of this strategic option is thus

66

ELMAR LUKAS

given by

( PðvÞ ¼

Bvb2 k v

v4vL v vL

(4)

whereby s ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 b2 k 1b2 1 ðr dÞ ðr dÞ 1 2 2r and b2 ¼ þ 2 B¼ b2 ðb2 1Þ 2 s2 s2 2 s are again constants. The corresponding optimal threshold value vL for initiating a divestment strategy results in vL ¼

b2 k b2 1

(5)

Consequently, the value of the chooser option is determined by F ¼ erT E Q ½maxfPðvÞ; CðvÞg

(6)

Q

with E [y] as the expectations operator under the martingale measure Q. This results in solving the following integral: 2 v ZL Zc rT 4 F ¼e ðk vÞ dPðvÞ þ Bvb2 dPðvÞ vL

1

Zvu þ

Avb1 dPðvÞ þ

c

Z1

3

_ 7 ðð Þv IÞ dPðvÞ5

ð7Þ

vu

where dP(v) denotes the implied probability measure. To derive a closedform solution for the complex chooser option one has to determine the aforementioned optimal threshold c. Thus, c is determined by the intersection of P(c) and C(c). From Acb1 ¼ Bcb2 ; we get !1=g 1b b1 vL 2 c¼ (8) 1 ð¯ Þ b2 v1b U with g ¼ b1 b2 : Solving Eq. (7) results in F ¼ kerT Nðd 3 Þ vedT Nðd 4 Þ þ Bvb2 Nðd 7 Þ Bvb2 Nðd 8 Þ þ Avb1 Nðd 5 Þ Avb1 Nðd 6 Þ þ ð¯ ÞvedT Nðd 1 Þ IerT Nðd 2 Þ

ð9Þ

Modeling the Evolutionary Sequence of International Joint Ventures

67

with v as the value of the overall IJV at time 0, N(y) as the cumulative normal distribution and v 1 v 1 ln ln þ r d þ s2 T þ r d s2 T vU 2 vU 2 pﬃﬃﬃﬃ pﬃﬃﬃﬃ d1 ¼ ; d2 ¼ s T s T ln d3 ¼

v 1 L r d s2 T 2 v pﬃﬃﬃﬃ ; s T

b1 ln d5 ¼

v U

v

1 r þ b21 s2 T 2 pﬃﬃﬃﬃ ; sb1 T

c 1 2 2 b2 ln r þ b2 s T v 2 pﬃﬃﬃﬃ d7 ¼ ; sb2 T

v 1 L ln r d þ s2 T 2 v pﬃﬃﬃﬃ d4 ¼ s T b1 ln d6 ¼

c 1 r þ b21 s2 T v 2 pﬃﬃﬃﬃ sb1 T

v 1 2 2 L b2 ln r þ b2 s T 2 v pﬃﬃﬃﬃ d8 ¼ sb2 T

Given the derived results, we can state that the MNE will initiate its foreign expansion strategy via the establishment of an IJV if the expanded net present value, i.e., the net present value rule adjusted by the value of subsequent ﬂexibility exceeds the associated costs of entry. Thus, the MNE’s timing decision results in v þ F ðvÞ K 0

(10)

where K comprises the costs for setting up such a market entry strategy and ev designates the initial value of the IJV.

3. RESULTS This section presents a summarization of the results and the comparativestatic analysis. If not noted speciﬁcally, we will assume the following values: r ¼ 0.055, s ¼ 0.4, d ¼ 0.01, and k ¼ 0.8. The value of the IJV’s ﬂexibility F is composed of the option value to dissolve the IJV (i.e., the ﬁrst to fourth terms) and the growth option value, which reﬂects the value of the subsequent cross-border M&A strategy (i.e., the remaining terms). As the result indicates, F increases with uncertainty, abandonment value, and time to

68

ELMAR LUKAS

0.85 0.80 0.75

F(v)

0.70 0.65 0.60 0.55 0.0 0.1 0.2 0.3 0.4 T 0.5 0.6 0.7 0.8 0.9 1.0

Fig. 1.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

v(t)

Value of Combined Acquisition/Divestment Option F With Respect to v(t) and Time of Joint Collaboration T.

maturity T ¼ t1. In addition, the value of the IJV’s ﬂexibility decreases with high initial equity shares, i.e., the lower the initial equity share the more valuable is the option, and higher acquisition costs for the remaining shares. Fig. 1 summarizes some of the results graphically. The comparative-static results for the trigger values vL, and vU are wellknown from the standard literature.8 The threshold value vU becomes larger and so does the propensity to wait when turning the IJV into a merger; the higher the costs of acquiring the remaining shares I are, the smaller b1 is. Moreover, an increase in involved aggregate investment uncertainty leads to an increase in vU. In addition, the trigger value is also dependent on the size of equity share e. If the MNE already holds a majority in the IJV, 1=ð¯ Þ becomes signiﬁcantly large, thus indicating an increased propensity to wait before acquiring the remaining shares (i.e., higher threshold value). Due to the concavity of vU(s), the effect of e and I is more signiﬁcant when the aggregate uncertainty of the overall project is higher. The opposite can be observed for the trigger value of the divestment stage. Low uncertainties correlate with a high threshold value. Due to the dependence of b2 on s, vL decreases as uncertainty increases, indicating that the propensity to wait also increases. This effect is further ampliﬁed, the lower the initial equity share e

Modeling the Evolutionary Sequence of International Joint Ventures

69

or the higher the recovery value k is. However, due to the convexity of vL(s), the effect of k and e is more signiﬁcant the lower the aggregate uncertainty of the overall project is. The chooser option is a path-dependent derivative. Hence, implications about the kind of termination the MNE chooses at time t1 can only be made in conjunction with the threshold c. As noted earlier, at t1 the MNE chooses that strategy, giving maximum return option value according to maxfCðvÞ; PðvÞg: Consequently, if vt1 is greater than c, the MNE will stick to its current strategy and further collaborate until the above-mentioned threshold vU is reached, turning it into a merger. If vt1 is lower than c, the MNE will further collaborate while at the same time preferring to dissolve the IJV. From the results derived, c shows now two different trends with respect to its dependence on project uncertainty. If the MNE holds a majority in the IJV, the threshold increases the higher the aggregate uncertainty is. Consequently, with increased uncertainty there is a perceived trend toward sell out because the MNE demands a higher project value for compensating the associated risks accompanied with a merger strategy. For minority IJVs, however, c is inversely dependent on project uncertainty. Thus, the chance for a subsequent merger is even greater the higher the project uncertainty becomes. Furthermore, given the fact that c is (not) reached, the propensity to initiate the investment (divestment) is even faster the lower the uncertainty s is (i.e., because only small upward (downward) movements of v(t) are needed to hit the corresponding threshold value). Both trends are dampened by a decrease in recovery value k. Fig. 2 depicts graphically the dependence of c on uncertainty and equity stake. In most cases, host governments are concerned about the actual amount invested in their country. While waiting for new information generally improves the overall risk exposure of the MNE in such a setting, i.e., given sunk costs and uncertainty, value for the host country is only generated if the ﬁrm exercises its real options right. Treating the initial decision of whether to start the IJV or not as an option right, the ﬁrm will invest as soon as Eq. (10) is satisﬁed. Because F(v) is always positive, the generated ﬂexibility can offset the transaction cost comprised by K, which is, e.g., the case for highly uncertain projects (see, e.g., Sanchez, 1993).

4. SUMMARY The expansion of MNEs is a path-dependent process that is reﬂected in the fact that the observed internationalization processes of MNEs are not only a

70

ELMAR LUKAS

2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.0

1.0 0.9 0.8 0.1

0.7 0.6

0.2 0.3

0.5

0.4

0.4

0.5 0.6 0.7

Fig. 2.

0.3

Critical Threshold Value c(r,d,s,e,I,k) for the IJV Termination Strategy With Respect to Uncertainty s and Equity Stake ð¯ ¼ 1Þ:

unidirectional paths. Therefore, strategic reorientation, divestment, or withdrawal must also be considered as serious strategies. The results show the new complementary insight that the evolutionary sequence of market entry via IJVs is not only driven by the growth option, as commonly modeled in the literature, but also by the ﬂexibility to dissolve the JV. While it has been commonly agreed that IJVs are a transitional form of foreign market expansion, less emphasis has been placed on what triggers the choice of termination form. Consequently, the model provides a solution that allows revealing which kind of termination is chosen by the MNE given the initial equity stake in the venture and uncertainty. Moreover, implications for governmental policies to attract FDI can be deduced from the model. The study presented provides new opportunities for further empirical research under an option framework or to extend the model by introducing, e.g., tax differentials.

NOTES 1. See Glaister, Husan, and Buckley (1998). However, in non-equity JVs sharing or exchange of equity may occur between partners (see ibid, p. 170).

Modeling the Evolutionary Sequence of International Joint Ventures

71

2. See, e.g., Gilroy (1993). 3. A detailed introduction to real options is given by Dixit and Pindyck (1994) and Trigeorgis (1998). 4. Although not explicitly stated, this model can be interpreted as a chooser option, albeit only two discrete states are considered. 5. Thus, the venture will be an equity joint venture JV if 0:05oo¯ (see, e.g., Gomes-Casseres, 1987). 6. The last step may be justiﬁed because a subsequent innovation renders an existing partner’s technology obsolete or due to misappropriation risk. Consequently, the venture is abandoned for the sake of a new venture or for withdrawal from the foreign market. 7. It is assumed that the acquisition price is ﬁxed right from the start. For a justiﬁcation of this assumption refer to, e.g., Chi and McGuire (1996). 8. See, e.g., Dixit and Pindyck (1994).

ACKNOWLEDGMENTS The author would like to thank Udo Broll, B. Michael Gilroy, and participants of 4th Korea and the World Economy Conference in Seattle for helpful comments and discussions. Any remaining errors are the sole responsibility of the author.

REFERENCES Blodgett, L. L. (1992). Factors in the instability of international joint ventures: An event history analysis. Strategic Management Journal, 13(6), 475–481. Broll, U., & Marjit, S. (2005). Foreign investment and the role of joint ventures. South African Journal of Economics, 73(3), 474–481. Buckley, A., & Tse, K. (1996). Real operating options and foreign direct investment: A synthetic approach. European Management Journal, 14(3), 304–314. Buckley, P. J., & Casson, M. C. (1996). An economic model of international joint venture strategy. Journal of International Business Studies, 27, 849–876. Buckley, P. J., & Casson, M. C. (1998). Models of the multinational enterprise. Journal of International Business Studies, 29(1), 21–44. Chang, S. J., & Rosenzweig, P. (2001). The choice of entry mode in sequential foreign direct investment. Strategic Management Journal, 22(8), 747–776. Chi, T. (2000). Option to acquire or divest a joint venture. Strategic Management Journal, 21(6), 665–687. Chi, T., & McGuire, D. J. (1996). Collaborative ventures and value of learning: Integrating the transaction cost and strategic option perspectives on the choice of market entry modes. Journal of International Business Studies, 27(2), 285–307.

72

ELMAR LUKAS

Chi, T., & Seth, A. (2002). Joint ventures through a real options lens. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategies and alliances: What we know 15 years later (pp. 71–87). Amsterdam: Elsevier. Dixit, A. K., & Pindyck, R. S. (1994). Investment under uncertainty. Princeton, NJ: Princeton University Press. Dunning, J. H., & Narula, R. (2004). Multinationals and industrial competitiveness: A new agenda. Cheltenham, UK: Edward Elgar. Folta, T. B., & Miller, K. D. (2002). Real options in equity partnerships. Strategic Management Journal, 23(1), 77–88. Gatignon, H., & Anderson, E. (1988). The multinational corporation’s degree of control over foreign subsidiaries: An empirical test of a transaction cost explanation. Journal of Law, Economics and Organization, 4, 305–336. Geringer, M., & Hebert, L. (1991). Measuring performance in international joint ventures. Journal of International Business Studies, 22(2), 249–263. Gilroy, B. M. (1993). Networking in multinational enterprises: The importance of strategic alliances. Columbia: University of South Carolina Press. Gilroy, B. M., & Lukas, E. (2006). The choice between greenﬁeld investment and cross-border acquisition: A real option approach. Quarterly Review of Economics and Finance, 46(3), 447–465. Glaister, K. W., Husan, R., & Buckley, P. J. (1998). UK international joint ventures with the triad: Evidence for the 1990s. British Journal of Management, 9(3), 169–180. Gomes-Casseres, B. (1987). Joint venture instability: Is it a problem? Columbia Journal of World Business, 22(2), 97–102. Hagedoorn, J. (2002). Inter-ﬁrm R&D partnerships: An overview of major trends and patterns since the 1960. Research Policy, 31(4), 477–492. Hennart, J.-F., Kim, D.-J., & Zeng, M. (1998). The impact of joint ventures on the longevity of Japanese stakes in the U.S. manufacturing afﬁliates. Organization Science, 9(3), 382–395. Howells, J., & Wood, M. (1993). The globalisation of production and technology. London: Belhaven Press. Kogut, B. (1991). Joint ventures and the option to expand and acquire. Management Science, 37(1), 19–33. Kogut, B., & Chang, S. J. (1996). Platform investments and volatile exchange rates: Direct investment in the U.S. by Japanese electronic companies. Review of Economics and Statistics, 78(2), 221–231. Leung, W. F. (1997). The duration of international joint ventures and foreign wholly-owned subsidiaries. Applied Economics, 29(10), 1255–1269. Marjit, S., Broll, U., & Mallick, I. (1995). A theory of overseas joint ventures. Economics Letters, 47(3–4), 367–370. Merton, R. C. (1973). Theory of rational option pricing. Bell Journal of Economics and Management Science, 4, 141–183. Pennings, E., & Sleuwaegen, L. (2004). The choice and timing of foreign market entry under uncertainty. Economic Modelling, 21(6), 1101–1115. Reuer, J. J., & Leiblein, M. J. (2000). Downside risk implications of multinationality and international joint ventures. Academy of Management Journal, 43(2), 203–214. Reuer, J. J., & Tong, T. (2005). Real options in international joint ventures. Journal of Management, 31(3), 403–423.

Modeling the Evolutionary Sequence of International Joint Ventures

73

Rivoli, P., & Salorio, E. (1996). Foreign direct investment and investment under uncertainty. Journal of International Business Studies, 27, 335–357. Sanchez, R. (1993). Strategic ﬂexibility, ﬁrm organization, and managerial work in dynamic markets: A strategic options perspective. Advances in Strategic Management, 9, 251–291. Todeva, E., & Knoke, D. (2005). Strategic alliances and models of collaboration. Management Decision, 43(1), 123–148. Trigeorgis, L. (1998). Real options: Managerial flexibility and strategy in resource allocation. Cambridge: MIT Press.

APPENDIX A While the ﬁrst two and the last two integrals of Eq. (7) are similar to the Black–Scholes integral and are solved for accordingly, one has to apply Itoˆ’s lemma to dvb and vb, respectively, to solve the remaining integrals. Thus, neglecting dividends this time we get dvb ¼ rvb dt þ sbvb dZ Q

(A.1)

b TþsbZQ T

(A.2)

vbT ¼ vb erT1=2s

2 2

The last two terms of the exponential function can be substituted into a stochastic process XV Nð1=2s2 b2 T; s2 b2 TÞ: Exemplarily, the solution is drafted for only one integral. The resulting integral e

rT

ZvU

Avb1 dPðvÞ

(A.3)

c

can be transformed by substituting Eq. (A.2) into ¼e

rT

Zb

Avb1 erTþs f Nð1=2s2 b2 T; s2 b2 TÞ ðsÞ ds

(A.4)

a

with a ¼ b1 lnðc=vÞ rT and b ¼ b1 lnðvU =vÞ rT as lower and upper boundaries. Utilizing the symmetry features of the normal distribution, i.e., Rb Rb Ra a f ðxÞ dx ¼ 1 f ðxÞ dx 1 f ðxÞ dx; the integral can easily be solved.

This page intentionally left blank

NEW TRENDS AND PERFORMANCES OF KOREAN OUTWARD FDI AFTER THE FINANCIAL CRISIS Seong-Bong Lee ABSTRACT This paper analyzes new outward foreign direct investment (OFDI) patterns, OFDI performance after the financial crisis in Korea, and the anticipated impact of these changes in OFDI on the Korean economy. This paper examines current trends in Korean OFDI activities from various viewpoints, including the geographical distribution of investments, industry, size, and investors’ equity share. This paper also verifies the relationship between OFDI and the Korean economy through an in-depth analysis of motives, performance, and prospects of Korean OFDI. Finally, two emerging issues regarding Korean OFDI are discussed.

1. INTRODUCTION Korean outward foreign direct investment (OFDI) increased rapidly during the 1990s. The Korean chaebols were main players in this OFDI increase (Bae & Hwang, 1997; Dent & Randerson, 1997; Park, 2000). However, it Value Creation in Multinational Enterprise International Finance Review, Volume 7, 75–96 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07004-X

75

76

SEONG-BONG LEE

decreased signiﬁcantly during the three to four years after the ﬁnancial crisis (Lee, 2000), and then reversed in 2002, increasing once again. One feature of the current rise in OFDI is that not only are large enterprises engaging in OFDI but also small- and medium-size companies. Therefore, Korean OFDI now shows a mixed investment pattern in which two different types of investment coexist: one performed by big companies, including Samsung, LG, and Hyundai Motors, to secure large local markets such as the United States and the European Union; the other by small- and medium-size companies to ensure low labor costs. This new pattern of OFDI reﬂects structural changes in the Korean economy as well as the growing trend of reciprocal inﬂuences. This paper analyzes new OFDI patterns, OFDI performance after the ﬁnancial crisis in Korea, and the anticipated impact of these changes in OFDI on the Korean economy. The ﬁrst part of this paper gives a brief historical overview and talks about recent developments in Korean OFDI policy. The following part examines current trends in Korean OFDI activities from various viewpoints, including the geographical distribution of investments, industry, size, and investors’ equity share. This paper also veriﬁes the relationship between OFDI and the Korean economy through an in-depth analysis of motives, performance, and prospects of Korean OFDI. Finally, two emerging issues regarding Korean OFDI are discussed.

2. TREND OF KOREAN OUTWARD FDI 2.1. OFDI Policy in Korea During the late 1980s, the Korean economy was booming, emerging from the lack of foreign currency that characterized the previous period, and Korean local ﬁrms, which matured rapidly under an export-led economic system, started to engage in active market-seeking activities, which directly led to a large increase in OFDI. From the late 1980s until the mid-1990s, Korean OFDI activities showed constant progress. Korea became a member of the OECD in 1996 and, from an institutional perspective, developed a freer market system and executed the switchover from an approval system to a notiﬁcation system regarding its OFDI rules in August 1997 (Park, 2000). From the 1997 ﬁnancial crisis until 2001, however, Korea showed an overall decline or stagnation in its OFDI activities. A large number of Korean ﬁrms closed or sold off their overseas subsidiaries. A particular

OFDI Performance after the Financial Crisis in Korea

77

characteristic of that time is that many overseas subsidiaries had a heavily indebted capital structure, which required parent companies to feed them in the form of loan guarantees. As a result, the Korean government introduced a system limiting the amount of loan guarantees from Korean parent companies in overseas capital investment to achieve a higher degree of soundness in OFDI activities. The limit for a loan guarantee was set at 95 percent of the outstanding level of guarantee, which only applied to afﬁliates of the 30 largest chaebols. Since 2002, Korean OFDI has been increasing, with a foreign exchange reserve of over $200 billion. The Korean government has also changed its OFDI policy to be less limiting to promote more OFDI. The limit on loan guarantees will be abolished in 2006. The current Korean policy for supporting OFDI can be categorized into the following three policies: loan programs, information provisions, and the reduction of non-commercial risks. The ﬁrst policy, loan programs, consists of a loan program of the Export-Import Bank of Korea for ﬁnancing of OFDI (up to 80 percent of the OFDI amount) and a loan program from the international economic cooperation fund of the Korean government, which is applicable to OFDI in developing countries with a long-term return period (The Ministry of Finance and Economy, 2002). The Export-Import Bank of Korea is a representative of the institutions that provide Korean overseas investors with useful information on the economic, business, and living conditions in host countries. The Korea Trade and Investment Promotion Agency provides, through its branch ofﬁces in each country, information to Korean ﬁrms that are investing in a particular region. In addition, the Korean Export Insurance Corporation offers export insurance to investors to reduce the non-commercial risks of OFDI, such as losses from war, nationalization, and so forth. The Korean government signed the ‘‘Investment Protection Agreement’’ with more than 70 countries to provide a more secure legal shelter for Korean OFDI. It also signed tax treaties with 62 countries to avoid possible international double taxation on the incomes of Korean subsidiaries abroad. 2.2. Overall OFDI Trends Clear differences between pre- and post-crisis trends in Korean OFDI over the past decade can be observed. Before the crisis of 1997, OFDI, led by Korean ﬁrms, was consistently rising. In 1994, it started increasing by nearly $1 billion per year due to the fact that large Korean ﬁrms, especially chaebols, began actively engaging in OFDI in accordance with global

78

SEONG-BONG LEE

business strategies. However, a radical restructuring of Korean ﬁrms during and after the crisis resulted in a sharp decline in OFDI performance, a trend that would reverse after 2002. The net outstanding invested amount of Korean OFDI is now $43.5 billion (October 31, 2005). Another important characteristic of post-crisis Korean OFDI is the sharp increase in OFDI withdrawal from 1998 to 2001. The annual average amount of OFDI withdrawal during the 1993–1997 period equals about $350 million (Table 1). During the post-crisis period, it exceeded $1 billion in 1998 and $3.3 billion in 2001. Then the amount began to decline in 2002, and eventually fell to $760 million in 2004. The total amount of OFDI withdrawal during 1998–2002, when the ﬁrm restructuring process was most extensive, equals $8 billion, which amounts to no less than 64.5 percent of $12.3 billion, the total OFDI withdrawal during the whole period from 1968 to the end of October 2005. This statistic shows that many Korean ﬁrms closed their foreign subsidiaries during the restructuring period. The net outstanding invested amount, which is the amount invested subtracted by the amount withdrawn, has demonstrated a recovery since 2002. This can be interpreted as Korean ﬁrms resuming active OFDI with the near-completion of the restructuring process. Table 1. Year Until 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005a Total

Korean OFDI Trends (Thousands of Dollars). Invested Amount

Withdrawn Amount

Net Amount

5,221,230 1,264,637 2,304,511 3,139,923 4,450,375 3,677,605 4,792,703 3,329,688 5,051,696 5,139,834 3,681,759 4,019,355 5,932,860 3,821,552

802,300 245,226 273,031 313,458 652,512 267,082 1,066,117 1,057,325 1,450,319 3,292,283 1,091,841 693,268 766,537 364,932

4,418,930 1,019,411 2,031,480 2,826,465 3,797,863 3,410,523 3,726,586 2,272,363 3,601,377 1,847,551 2,589,918 3,326,087 5,166,323 3,456,620

55,827,728

12,336,231

43,491,497

Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each year. a Until October 31, 2005.

Millions of Dollars

OFDI Performance after the Financial Crisis in Korea

6,000 5,500 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0

Others Latin America Europe North America Asia

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Fig. 1.

79

2004

Korean OFDI by Region (In Years). Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each Year.

2.3. OFDI by Region By 2004, Korean OFDI in Asian countries was $16.8 billion, which accounted for 43 percent of the total net outstanding invested amount of OFDI. North America ranked second with $10.8 and accounted for 27 percent. European countries made up 17 percent with $6.7 billion and Korean OFDI in Latin American countries was $3.4, 9 percent of Korea’s total OFDI. Fig. 1 shows OFDI trends by location. Most interesting is the fact that Korean OFDI in Asia has increased in recent years. This increased investment is closely related to rapidly increased investment in China. Fig. 2 shows ﬁgures for Korean investment in China. According to this graph, there has been rapid growth, or 54 percent of annual growth, since 2001: $0.6 billion in 2001, $1.2 billion in 2002, $1.63 billion in 2003, and $2.2 billion in 2004. In 2002, China became the primary recipient of Korean investment, occupying the United States’ former position. Investments in the United States were $0.57 billion in 2002, $1.05 billion in 2003 and $1.34 billion in 2004, exhibiting a widening investment gap between the United States and China. 2.4. OFDI by Industry If we look at the outstanding net amount invested by industry, the manufacturing sector occupies the largest piece of the pie at 53 percent ($21 billion). The second largest belongs to wholesale and retail, which takes up 22 percent ($8.7 billion). The mining industry follows, holding 6 percent

2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0

04

03

20

01

02

20

20

00

20

99

20

98

19

97

19

96

19

94

95

19

19

93

Total invested projects

19

Fig. 2.

2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0

Total invested amount

Projects

SEONG-BONG LEE

19

Millions of Dollars

80

Korean OFDI Trends in China. Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each Year.

Fig. 3. Korean OFDI in the Manufacturing Sector by Type and Region. Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each Year.

($2.1 billion), and other sectors including the service industry take up 19 percent of the total. This shows that ﬁrms in the manufacturing sector have been the driving force behind Korean OFDI. In other words, Korean investment is concentrated in ensuring overseas production facilities (53 percent) or market expansion to increase sales (22 percent). However, a low degree of investment in the service sector compared to that of the manufacturing sector shows that Korea has relatively weak competitiveness in the world services market. Analyzing investment in ‘‘manufacturing’’ and ‘‘trade and retail’’ by subcategorizing and observing them by region (Figs. 3 and 4) produces

OFDI Performance after the Financial Crisis in Korea

81

Fig. 4. Korean OFDI in the Trade and Retail Sector by Type and Region. Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each Year.

interesting ﬁgures. ‘‘Electronics and telecommunication equipment’’ comprises 33 percent of the total net outstanding invested amount in manufacturing sector, with $6.8 billion, and ‘‘motors and equipment’’ follows it taking up 12 percent with $2.5 billion. It becomes clear that Korean OFDI in manufacturing is primarily led by electronics and automobile ﬁrms. When divided by region, Asia leads with $11.5 billion, or 55 percent of the total net outstanding invested amount in the manufacturing sector, North America with $4.4 billion or 21 percent, and Europe with $4 billion or 19 percent of total manufacturing investment. Such regional disparity demonstrates that Korean manufacturing ﬁrms aim for cost minimization when engaging in OFDI. In other words, the fact that more than half of investments in Asia go to China and South-East Asian countries makes clear the importance of cost minimization through the capitalization of the relatively low wages in those countries. Investment in the North American and European regions can be accounted for by Korean ﬁrms’ desire for costeffective measures in those markets, such as tariff and non-tariff trade barrier evasions and the reduction of transportation costs. In the case of ‘‘trade and retail’’ investments by Korean ﬁrms, 34 percent of the total net outstanding invested amount ($7.1 billion) was achieved by electronics and equipment ﬁrms at the end of 2004, and 21 percent ($4.4 billion) by automobile ﬁrms. This means that the aggregate amount invested in foreign sales subsidiaries by electronics and automobile ﬁrms takes up 55 percent of the total overseas investment. Therefore, this implies that Korean ﬁrms, particularly exporters of electronics and automobile industries, have a clear motive behind developing overseas markets. The geographical dispersal of trade and retail investment shows that 43 percent of the total amount invested ($9 billion) goes to North America,

82

SEONG-BONG LEE

while Europe and Asia receive 25 percent each. This implies that investment in this sector can be fully characterized as a market-seeking investment with the purpose of increasing exports to the target market. 2.5. OFDI by Investing Enterprise

Millions of Dollars

The investing ﬁrms can be classiﬁed as large enterprises and small and medium enterprises (SMEs). By 2004, the net outstanding invested amount by large enterprises was $27.37 billion, which accounted for 69 percent of the total amount of investments. In comparison, SMEs accounted for 27 percent with $10.7 billion, and other individual investors accounted for 4 percent with $1.5 billion. However, the project basis distribution is different: 61 percent for SMEs, 29 percent for other individual investors, and 10 percent for large enterprises. Large enterprises invested a large amount of money in a few projects, which accounts for the disparity in distribution between amounts and project bases (Fig. 5). This can be observed more clearly when looking at OFDI by invested amount per project. By the end of 2004, the sum of investments over $50 million per project was $14.25 billion, which accounted for 36 percent of the total net outstanding invested amount, and the sum of investments from $10 million to $50 million was $10.29 billion, making up 26 percent of the total. The sum of these two investment types was $24.44 billion, which accounted for 65 percent of total net outstanding invested amount, similar to the 69 percent of the amount invested by large enterprises. The interesting thing is that the share of OFDI by SMEs continues to increase from year to year. From 2000 to 2004, the share of investments by SMEs increased by 33.7 6,000 5,500 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0

Others SME Large enterprise

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Fig. 5. Korean OFDI by Investing Enterprise and Year. Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each Year.

OFDI Performance after the Financial Crisis in Korea

83

percent of annual average growth. This growth rate was two times higher than that of 17.5 percent during 1993–1999. It means that SMEs have been actively participating in FDI in the past ﬁve years. The amount of investments per project by SMEs is also increasing. From 1993 to 1999, the average investment amount per project by SMEs was $701,000, whereas from 2000 to 2004, it increased to $1.1 million. Increasing SME investment, both in project numbers and monetary value, implies that more and more Korean SMEs are not just trying to secure efﬁcient manufacturing facilities at low costs, but rather struggling to survive by moving most of their value chains to foreign countries. 2.6. Management Control over Foreign Subsidiaries of Korean Firms One of the most used indicators for management control over foreign subsidiaries of Korean ﬁrms (Guillen, 2003) is statistics on OFDI by the investor’s equity ratio. On a project basis, wholly owned foreign subsidiaries account for 62 percent, and subsidiaries with 50–100 percent shares by Korean parents were 15 percent of the total number of OFDI projects (Fig. 6). The sum of these two cases is 75 percent and it implies that Korean enterprises are pursuing management control in most OFDI by majority shareholding. On a net outstanding invested amount basis, wholly owned subsidiaries accounted for 66 percent, and subsidiaries with 50–100 percent shareholding were 20 percent. The sum of these two ratios is 86 percent. It is noticeable that the share of majority shareholdings on an amount basis is 11 percent higher than that calculated on a project basis. It implies that there could be more majority shareholding investments from larger enterprises than from SMEs. It implies that large enterprises prefer to have more control over their foreign subsidiaries than SMEs. This tendency could be interpreted to mean

Fig. 6.

Korean OFDI by Investors’ Equity Ratio. Source: The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, each Year.

84

SEONG-BONG LEE

that large Korean ﬁrms have their own signiﬁcant competitive advantages over competitors in foreign markets, and therefore prefer carrying out business themselves. Meanwhile, SMEs are pursuing a cooperative strategy with local partners in foreign markets to a greater degree than large enterprises (Tallman & Shenkar, 1990).

3. MOTIVES AND PERFORMANCES OF KOREAN OFDI 3.1. Motives for Korean OFDI According to an analysis done by The Export-Import Bank of Korea (2004) on motivating factors behind OFDI, written by the investors in the form of a notice to the Korean government, the primary factor was to secure or develop local or third-country markets (about 50 percent of total notices), and a secondary motivator was to take advantage of low labor costs in the host country. Table 2 shows other motives for Korean OFDI. Until 1993, the development of natural resources was the main objective of investment. This statistic is based on amount invested, and a large part of investments during that time was for the development of natural resources. Meanwhile, we can see that the importance of the motive of utilizing low labor costs increased when comparing motives in the period 1997–2001 and after 2002. This corresponds to the fact that OFDI by SMEs has been increasing rapidly since 2002, with intensive concentration in China and other Asian countries. If we look at motives for OFDI by destination and industry (Table 3) from 2002 to November 2004, the trend mentioned above is more clearly Table 2.

Motivation of Korean OFDI by Period (Percent).

Securing or developing local or third-country markets Utilizing local labor costs Avoiding trade barriers Securing raw materials Acquiring advanced technology or management know-how Developing natural resources

1968–1993

1994–1996

1997–2001

After 2002

28.9

50.2

52.4

47.1

14.7 1.7 3.8 1.1

37.2 2.5 4.8 2.6

30.3 2.3 3.9 7.7

38.5 3.1 4.4 4.1

49.9

2.7

3.3

2.8

Source: The Export-Import Bank of Korea (2004).

Motivation of Korean OFDI by Destination and Industry (Percent).

Motive

Destination

Industry

China United States Europe Manufacturing

Securing local or third-country markets Utilizing local labor costs Avoiding trade barriers Securing raw materials Acquiring advanced technology/ management know-how Developing natural resources

Electronics and Telecom Automobiles Equipment

Textiles and Clothes

41.2

65.1

66.7

43.5

50.9

50.7

41.0

47.9 3.2 4.6 1.1

4.7 2.3 3.5 21.2

8.8 0.7 4.8 11.6

45.1 3.4 4.2 2.1

41.7 2.9 0.6 2.5

38.7 4.4 0.9 4.4

52.1 3.2 2.1 0.8

2.1

3.2

7.5

1.8

1.4

0.9

0.7

OFDI Performance after the Financial Crisis in Korea

Table 3.

Source: The Export-Import Bank of Korea (2004).

85

86

SEONG-BONG LEE

identiﬁable. According to this table, the primary motive for Korean OFDI in China was to take advantage of low labor costs (47.9 percent of total). In contrast, the most important motive for Korean OFDI in the United States and Europe was to secure the local market, weighing in at 65.1 and 66.7 percent, respectively. Meanwhile, acquiring advanced technology or management know-how was the second most important reason for Korean OFDI in the United States (21.2 percent) and Europe (11.6 percent). If we look into OFDI motives by industry, the manufacturing industry’s most important investment motive was to make use of low labor costs (45.1 percent of the total). However, in the case of OFDI in globally competitive industries such as the electronic communications equipment industry and the automobile industry, securing or developing local or third-country markets was the most important motive for investment. Meanwhile, in less competitive industries, like the textile and clothing industries, utilizing low local labor costs was the most important motive for investment. While this analysis of OFDI motives does not include the differences between investments by large companies and those of SMEs, another analysis (The Export-Import Bank of Korea, 2004) of 318 cases of OFDI that amount to more than $10 million in outstanding invested amounts at the end of 2003 – mostly carried out by large companies – revealed that securing or developing local or third-country markets was the primary motive in all destinations, creating a great gap in the motive of utilizing low labor costs. If we reverse these results, it indirectly conﬁrms that most OFDI by Korean SMEs were aimed at utilizing local labor costs. 3.2. Performances of Korean OFDI 3.2.1. Performances of Large Foreign Subsidiaries of Korean Firms For all foreign subsidiaries with an invested amount of more than $10 million, the Korean foreign exchange regulation obliges investors to submit an annual report, including ﬁnancial statements of foreign subsidiaries, to the Korean government. The Export-Import Bank of Korea has done annual analyses of these reports and ﬁnancial statements of foreign subsidiaries since 1999 (in other words, for the subsidiaries’ ﬁnancial performance for the accounting year 1998). In 2005, the Bank’s analysis report included an annual comparative analysis of ﬁnancial performances from 1998 to 2004. In this paper, we explain some meaningful results of this analysis, including ﬁnancial performance trends and trade effects. Among the foreign subsidiaries whose invested amount was over $10 million at the end of each year, those with reliable ﬁnancial statements were analyzed.

OFDI Performance after the Financial Crisis in Korea

87

Therefore, for this seven-year period, sample foreign subsidiaries may differ from year to year. 3.2.1.1. Overall Features of Sample Subsidiaries for Performance Analysis. Table 4 shows overall features of sample subsidiaries for each year. The average amount invested for sample subsidiaries has decreased, but average sales and average net income have clearly increased over the past three years. Among the sample foreign subsidiaries during this period, the portion of wholly owned subsidiaries increased from 60 percent in 2002 to over 70 percent in 2003 and 2004. This ﬁgure is based on the number of subsidiaries and amount of equity balance. This shows that large Korean ﬁrms have come to prefer controlling all resources for investment and business in foreign markets to collaborating with local partners. 3.2.1.2. Annual Comparison of Financial Performance. Table 5 shows some selected indicators for the ﬁnancial performances of sample foreign subsidiaries. First of all, the average of the capital stock maintenance ratio, which is calculated by subtracting 1 from the ratio of total equity capital to the capital stock of each subsidiary, was recorded as –38.7 percent in 1998 and – 51.1 percent in 2000. Then the ratio increased to –17.3 percent in 2004. These ﬁgures reveal a trend of increasing net income of foreign subsidiaries, while cumulatively remaining in a condition of capital depletion. The ﬁnancial stability of foreign subsidiaries seems to have improved: the debt ratio was 522.4 percent in 2000 but decreased to 293.8 percent in 2004. Proﬁtability has also been improving, based on the ratio of net income to stockholders’ equity, which reversed to surplus ﬁgures and reached 13.3 percent in 2004. Since 2002, subsidiaries’ annual growth (the growth rate of sales and the growth rate of total assets) has remained over two digits Table 4. Year

Overall Features on Sample Subsidiaries for Performance Analysis (Millions of Dollars). 1998

1999

2000

2001

2002

2003

2004

Cases of foreign subsidiaries 290 318 276 311 318 319 377 Investment balance account 11,657 16,395 16,482 17,844 16,941 17,692 18,232 Average investment amount 40 52 60 57 53 55 48 Average sales 174 222 290 226 262 327 383 Average net income 10 6 1 3 0 1 6 Percent of wholly owned OFDI (%) – 61.4 62.6 65.0 61.9 72.4 75.6 Source: Export-Import Bank of Korea (2005).

88

Table 5. Annual Comparison of Financial Performances of Foreign Subsidiaries with over $10 Million Invested (Percent). Year

Source: Export-Import Bank of Korea (2005).

1999

2000

2001

2002

2003

2004

38.7

38.3

51.1

48.5

42.2

35.7

17.3

400.1 98.8

492.4 109.4

522.4 115.5

453.2 121.5

358.7 101.3

329.2 99.5

293.8 91.2

2.8 28.9

0.7 20.4

2.0 5.1

0.6 12.0

1.2 6.2

1.9 4.8

1.8 13.3

1.2 6.5

24.4 7.9

24.4 8.2

9.8 7.4

10.5 4.3

27.5 17.3

33.8 23.8

1.0 8.6 –

1.3 11.2 21.4

1.6 12.0 13.5

1.5 10.2 24.0

1.8 11.0 7.6

2.3 12.2 10.0

2.6 14.2 16.7

SEONG-BONG LEE

Maintenance ratio of capital stock Indicators concerning ﬁnancial stability Debt ratio Fixed assets to stockholders equity and long-term liabilities Indicators concerning proﬁtability Operating income to sales Net income to stockholders equity Indicators concerning growth Growth rate of sales Growth rate of total assets Turnover of assets (times) Total assets turnover Inventory turnover Return on investment

1998

OFDI Performance after the Financial Crisis in Korea

89

and their sales growth rate is larger than their total asset growth rate. Total assets and inventories turnover, which indicate the utilization of assets, have also been constantly increasing since 2001. From 1999 to 2001, the average annual return on investment was –19.6 percent. In 2002, return on investment ﬁgures reversed to positive ratios and has since been rapidly increasing. 3.2.1.3. Annual Comparison of Trade Surplus Effects. The positive effect of foreign subsidiaries in stimulating Korean exports has been well proved by previous research (Lim & Moon, 2001). The effect, calculated by the ExportImport Bank of Korea as the ratio of exports to foreign subsidiaries to the net outstanding invested amount, was recorded as 134.5 percent in 1998 and then continued to increase to over 200 percent after 2003 (Table 6). Foreign subsidiaries’ import-arousing effect, which is calculated as the ratio of imported amount from foreign subsidiaries to Korea to the net outstanding invested amount of them, reached 34.3 percent in 1998 and continued to increase to reach 125.2 percent in 2004. The improving effect of trade balance during the last seven years, calculated by subtracting the import-arousing effect from the export-arousing effect, was 130 percent, on average. This ﬁgure reached 141 percent in 2004, perhaps due to foreign investment. 3.2.2. Comparison of Foreign Subsidiary Performance by Size At the end of 2004, The Export-Import Bank of Korea conducted a comparative study of the ﬁnancial statuses (The Export-Import Bank of Korea, 2005, pp. 99–106) of 646 small- and medium-size subsidiaries that invested less than $10 million and a group of 377 large-size subsidiaries that invested more than $10 million. Those 1,023 foreign subsidiaries were selected from 3,265 foreign subsidiaries whose investment balance was over $1 million, according to ﬁnancial statements submitted to the Korean government. Table 6.

Annual Comparison of the Trade Surplus Effectsa of Foreign Subsidiaries (Percent).

Year

1998

1999

2000

2001

2002

2003

2004

Export-arousing effect Import-arousing effect Trade surplus effect

134.5 34.3 100.2

158.5 35.6 122.9

188.6 61.0 127.6

154.8 44.0 110.8

156.1 48.8 107.3

229.2 65.5 163.7

266.1 125.2 141.0

Source: The Export-Import Bank of Korea (2005). a Trade surplus effect is the ratio of exports to foreign subsidiaries to the net outstanding invested amount.

90

SEONG-BONG LEE

The statistical overview of this sample (Table 7) accurately represents the overall trends of Korean OFDI as reviewed in Chapter 2. When looking at the regional distribution of sample small- and medium-size subsidiaries in terms of net outstanding invested amount, 74.6 percent of them are located in Asian regions, especially in China (48.2 percent) and in ASEAN countries (13.3 percent). That ratio in the European and North American regions together is just 19.8 percent. For large subsidiaries, locations in Europe and North America together account for 51.5 percent, and 41.3 percent in Asia, mainly in China (20.5 percent) and ASEAN countries (10.3 percent). Table 8 compares selected ﬁnancial ratios between both groups. According to this comparison, the small- and medium-size subsidiaries had more favorable ﬁgures in stability and growth, while in terms of proﬁtability and operation the large-size subsidiaries were identiﬁed as better than small- and medium-size subsidiaries. In comparing ﬁnancial ratios, both groups exhibited negative values in capital stock maintenance ratios. This means that in 2004, subsidiaries of Korean companies were still at a level of capital depletion. The small- and medium-size subsidiaries’ capital stock maintenance ratio was –1 percent, which is relatively high, but the capital stock maintenance ratio of the group of large-size subsidiaries was very low and reached –17.3 percent. In terms of the debt ratio and the stability ratio, the small- and mediumsize subsidiaries showed better values than the group of large-size subsidiaries. This can be interpreted in two ways: ﬁrst, because small- and medium-size enterprises have relatively less experience in foreign investment, this group tends to avoid potential business risks; second, this result reﬂects the fact that small- and medium-size enterprises are less able than large-size enterprises to ﬁnance funds through borrowing. The proﬁtability of the small- and medium-size subsidiaries group is worse than the group of large-size subsidiaries. The reason is that competition among small- and medium-size subsidiaries is very strong in local markets, so proﬁt margins for sales are low. And due to the lack of experience in running a foreign business, sales and administration costs are high. Among the 646 small- and medium-size subsidiaries, 322 subsidiaries resulted in an operating loss. Meanwhile, in terms of return on investment, large-size subsidiaries were almost three times better than the small- and medium-size subsidiaries. The growth rate of the group of small- and medium-size subsidiaries was higher than the group of large-size subsidiaries, whereas the group of largesize subsidiaries showed higher turnover of assets and capital. Since the difference of these two indicators is between the two groups, it is difﬁcult to ﬁnd signiﬁcance in this difference.

Overall Features on Sample Subsidiaries for Comparative Analysis by Size.

Region

Investment of Less Than $10 Million Cases (Number, %)

Investment of More Than $10 Million

Invested Amount (Millions of Dollars, %)

Cases (Number, %)

Invested Amount (Millions of Dollars, %)

Asia China Hong Kong Japan ASEAN North America Europe EU South America Oceania Africa

498 327 25 25 91 85 37 31 13 9 4

(77.1) (50.6) (3.9) (3.9) (14.1) (13.2) (5.7) (4.8) (2.0) (1.4) (0.6)

1,662 1,074 89 95 297 313 129 114 71 34 19

(74.6) (48.2) (4.0) (4.3) (13.3) (14.0) (5.8) (5.1) (3.2) (1.5) (0.9)

215 118 19 10 58 81 50 46 18 11 2

(57.0) (31.3) (5.0) (2.7) (15.4) (21.5) (13.3) (12.2) (4.8) (2.9) (0.5)

7,531 3,729 1,007 338 1,872 6,691 2,703 2,465 586 516 205

(41.3) (20.5) (5.5) (1.9) (10.3) (36.7) (14.8) (13.5) (3.2) (2.8) (1.1)

Total

646

(100.0)

2,227

(100.0)

377

(100.0)

18,232

(100.0)

OFDI Performance after the Financial Crisis in Korea

Table 7.

Source: The Export-Import Bank of Korea (2005, 100).

91

92

SEONG-BONG LEE

Table 8.

Comparison of Financial Performances of Foreign Subsidiaries by Size (Percent).

Average of Selected Indicators

Maintenance ratio of capital stock 100% shareholding subsidiaries Subsidiaries with 50–100% shareholding Subsidiaries with shareholding less than 50% Indicators concerning ﬁnancial stability Debt ratio Fixed assets to stockholders’ equity and long-term liabilities Indicators concerning proﬁtability Operating income to sales Net income to stockholders’ equity Indicators concerning growth Growth rate of sales Growth rate of total assets Turnover of assets Total assets turnover Inventories turnover Return on investment

Investment of Less Than $10 Million

Investment of More Than $10 Million

1.0 1.5 18.6

17.3 19.9 12.0

10.0

12.4

177.9 83.7

293.8 91.2

0.3 1.5

1.8 13.3

56.6 28.5

33.8 23.8

2.0 9.7 6.4

2.6 14.2 16.7

Source: The Export-Import Bank of Korea (2005, pp. 103–104).

Table 9.

Comparison of Trade Effects of Foreign Subsidiaries by Size.

Export-arousing effect Import-arousing effect Trade surplus effect

Investment of Less Than $10 Million

Investment of More Than $10 Million

431.8 133.1 298.6

266.2 125.2 141.0

Source: The Export-Import Bank of Korea (2005, p. 104).

One interesting fact is that the trade surplus effects of the group of smalland medium-size subsidiaries are superior to those of the large-size subsidiaries group. Table 9 compares the trade surplus effects between the two groups. The group of small- and medium-size subsidiaries showed 431.8 percent in export-arousing effect and the large-size subsidiaries group showed 266.2 percent. Also, the export-arousing effects of the small- and

OFDI Performance after the Financial Crisis in Korea

93

medium-size group are much higher and the import-arousing effects of the two groups are similar. From this result, it can be concluded that the contribution to trade surplus per unit investments of small and medium size is higher than for large-size investments. This can be attributed to the fact that small- and medium-size enterprises usually procure a large amount of raw material, by-products, and equipment from Korea, manufacture the products locally at lower labor costs, and sell the products to local Korean largesize subsidiaries or export to other countries.

4. SUMMARY AND EMERGING ISSUES REGARDING KOREAN OUTWARD FDI This paper conducts an in-depth analysis on current Korean OFDI and trends, performance, and impact on the Korean economy. Important characteristics of recent Korean OFDI trends can be summarized as follows. First, Korea has increased the amount of its OFDI activities in Asia. In particular, China has become the primary destination of Korean OFDI, relegating the United States to second place. This reﬂects a global trend in which China is gaining recognition as a favored FDI destination. Second, a large part of FDI is concentrated in the manufacturing sectors, of which some limited industries such as electronics and automobile constitute a large part. This trend reﬂects the fact that Korea is considerably competitive in these two industries in the global market. Third, it is observed that in the past three years OFDI by small- and medium-size companies has largely increased. This can be interpreted in light of the loss of Korean domestic cost competitiveness, especially for SMEs, which resulted in an increase in OFDI of SMEs to other Asian countries, especially China, in order to ensure lower labor costs (Zhan, 2005). Fourth, investments are motivated by both market-seeking purposes and low labor cost seeking purposes, and the importance of each of the purposes can be differentiated by the size of investment: larger investments are for mainly market-seeking purposes, while small and medium types of investments seek low labor costs. In the case of OFDI to developed countries such as the United States and European countries, the second motive after market-seeking was found to be to ‘‘[learn] advanced technology or know-how.’’ Fifth, overseas legal subsidiaries of Korean ﬁrms improved their ﬁnancial performance remarkably in the past seven years. In the case of small and medium types of investments, however, returns on investments are far lower than those of large-scale investments.

94

SEONG-BONG LEE

Sixth, OFDI actually contributes to the improvement of the Korean trade balance, which is accurate when evaluating data limited to each foreign subsidiary-related export and import. However, this analysis is limited in that it did not include any spillover effects of OFDI toward the trade of other Korean ﬁrms. Meanwhile, statistics suggest that small- and mediumsize investments create bigger export-arousing effects. However, as the data include transferred parts of existing production facilities in Korea into export items, it is necessary to evaluate the net export-arousing effect that excludes the transfer of existing facilities. Moreover, there has been no study that clearly and comprehensively evaluates the crowding-out effects of Korean OFDI on domestic investment. Finally, Korean ﬁrms perceive their competitiveness in foreign markets as quite positive. However, with the exception of some large-scale investments, a large number of investment activities do not have good proﬁtability, which actually renders the mid- and long-term competitiveness of Korean ﬁrms rather gloomy. In spite of all this, Korean ﬁrms are expected to consistently expand their overseas investment. This expectation is derived from two facts: ﬁrst, internationally, the Korean economy is now more strongly linked to the global economic system due to increasing international trade and investment; second, domestically, Korea has been undergoing an industrial restructuring process with a rise in costs of production factors such as labor, land, and so forth. Considering this analysis on current trends, it is estimated that Korean OFDI is about to face two issues. The ﬁrst issue is the polarization of OFDI between large ﬁrms and SMEs. Investment by large ﬁrms, which consists of high value-added ﬁelds in value chains such as R&D and design activities, are mainly carried out within Korea, while investments for low value-added activities including assembly are made largely in response to the requirements of cost-effectiveness with the goal of reinforcing global competitiveness from the viewpoint of rationalized production. On the contrary, a remarkable aspect of Korean SMEs is that many of them, as a way of survival, have moved their entire production facilities to a foreign country, giving up the Korean market because of a loss in competitive edge due to an overall rise in Korean production costs. The impact of OFDI by SMEs to long-term enhancement of ﬁrms’ competitiveness is not considerable. In the case of SMEs, therefore, other changes in the management environment such as increased production costs in the local market are likely to affect the very existence of the company. The second issue is the problem of imbalance between OFDI in manufacturing sectors and that in service sectors. So far, Korean OFDI has

OFDI Performance after the Financial Crisis in Korea

95

mainly focused on manufacturing sectors, and even in service sectors it has been ‘‘trade and retail’’ that has been important. Moreover, Korean investment in this ‘‘trade and retail’’ part is actually for the purpose of sales of electronic and automobile products in overseas markets. This trend clearly reﬂects the weakness of Korean OFDI in speciﬁc service industries, such as banking or distribution. Despite a highly complementary relationship between OFDI in the banking and ﬁnancial or distribution sectors and that of the manufacturing sectors, actual foreign investment by Korean banks or logistics companies does not meet the demand of large Korean ﬁrms for ﬁnancial and distribution services. From the fact that growth in the service sectors is a key element for sustainable growth in the Korean economy, it is high time for Korea to improve its competitiveness in the service sectors through more positive OFDI activities. Although these challenges should be overcome on the level of the individual ﬁrm, there should be speciﬁc measures thoroughly examined and implemented on the governmental level. To achieve high quality of OFDI by SMEs, it is crucial to fortify cooperation between large ﬁrms and SMEs, which will consequently minimize the negative effects of SME OFDI in reducing their investment in domestic markets. In addition, the Korean government should ease regulations that hinder investment to pave the way for boosting OFDI activities in the service industries.

REFERENCES Bae, C. S., & Hwang, S. (1997). An empirical analysis of outward FDI of Korea and Japan. Multinational Business Review, 5(2), 71–80. Dent, M. C., & Randerson, C. (1997). Enter the chaebol: The escalation of Korean direct investment in Europe. European Business Journal, 9(4), 31–39. Guillen, F. M. (2003). Experience, imitation and the sequence of foreign entry: Wholly owned and joint-venture manufacturing by South Korean ﬁrms and business groups in China, 1987–1995. Journal of International Business Studies, 34(2), 185–198. Lee, S.-B. (2000). Korea’s overseas direct investment: Evaluation of performances and future challenges. Korea Institute for International Economic Policy Working Paper 00-12. Lim, S.-H., & Moon, H.-C. (2001). Effects of outward FDI on home country exports: The case of Korean ﬁrms. Multinational Business Review, 9(1), 42–49. Park, Y. S. (2000). The Korean chaebols’ dash for globalization: Has it really been driven by government policy? Korea Observer, 31(4), 579–605. Tallman, B. S., & Shenkar, O. (1990). International cooperative venture strategies: Outward investment and small ﬁrms from NICs. Management International Review, 30(4), 299–315. The Export-Import Bank of Korea (2004). A Way to Promote Outward Foreign Direct Investment of Korea (in Korean).

96

SEONG-BONG LEE

The Export-Import Bank of Korea (2005). Analysis of Business Performances of Foreign Subsidiaries of Korean OFDI (in Korean). The Export-Import Bank of Korea, Overseas Direct Investment Statistics Yearbook, Each Year. The Ministry of Finance and Economy (2002). Regulations on the Outward Foreign Direct Investment (in Korean). Zhan, X. (2005). Analysis of South Korea’s direct investment in China. China & World Economy, 1, 94–104.

PART III: STRATEGIES AND FIRM PERFORMANCE

This page intentionally left blank

THE VALUE CREATION PERSPECTIVE OF INTERNATIONAL STRATEGIC MANAGEMENT$ Reid W. Click ABSTRACT This paper applies the concept of value creation to examine the strategic management of multinational enterprises. ‘‘International strategic management’’ is first defined as the process through which value is created by managers operating across a national border. The domain of international strategic management is thus determined by activities that distinguish international management from domestic management in the process of value creation. This perspective on value creation is used to answer three questions pertaining to international strategic management. First, how important is international strategic management? Simple statistics presented here demonstrate that the international component of value creation is important in the U.S. economy. Second, what is the domain of international strategic management? The paper presents a framework in which international strategic management is the aggregation of value

$

An earlier version of this paper, entitled ‘‘Strategic Management of Multinational Enterprises and Value Creation,’’ was presented at the Eighth International Conference on Multinational Enterprises at the Chinese Culture University in Taipei, Taiwan, March 2006. I am grateful for comments from the conference participants.

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 99–123 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07005-1

99

100

REID W. CLICK

created through international production, marketing, and financial activities, and reveals that the domain of international management is vast. Third, does international strategic management make the whole multinational enterprise worth more than the sum of its parts? Empirical evidence suggests that the answer is yes, at least for U.S. multinationals in the early 1990s.

1. INTRODUCTION Management is the process through which value is created by an individual or a group of individuals, and international management is the process through which value is created by an individual or group of individuals operating across a national border. Since the meanings of ‘‘international,’’ ‘‘management,’’ and ‘‘international management’’ are apparently in dispute, or are at least under-understood (see Boddewyn, 1999, and references therein), these simple introductory deﬁnitions require further examination. A dictionary deﬁnition of ‘‘management’’ is ‘‘the act, manner, or practice of managing, handling or controlling something’’ (Morris, 1975, p. 192). In recognition of a raison d’eˆtre for the individuals who manage something, a normative version of the deﬁnition is ‘‘the act, manner, or practice of managing, handling or controlling something in order to create value or wealth.’’ Although the normative appendage is disputable, managers must recognize that the things worth doing are the ones that create value. This is the overarching message conveyed in modern business school curricula. As Harrigan (1992) points out, ‘‘Business schools exist to improve management practice. All scholarly research, development of teaching materials, testing of theories, and other activities carried on by faculties of business schools should help managers develop tools for coping with the problems of international value creation’’ (p. 251). The rest of this paper tackles the essence of value creation, so the deﬁnition being offered is the starting point for discussion rather than a deﬁnitive statement. Since ‘‘international’’ is ‘‘relating to, or involving two or more nations or nationalities’’ (Morris, 1975, p. 685), ‘‘international management’’ is ‘‘the act, manner, or practice of managing, handling, or controlling something involving two or more nations.’’ Hence, incorporating the normative element offered in this paper, international management is the process through which value is created by an individual or group of individuals operating across a national border.

Value Creation Perspective of International Strategic Management

101

In attempting to circumscribe the domain of international management, the deﬁnition of international management provides guidance for two reasons. First, it distinguishes international management from general management with the phrase ‘‘operating across a national border.’’ Second, it distinguishes international management, being focused on the ‘‘process through which value is created’’ internationally, from international business, which encompasses anything related to international commercial activity (such as elements of the operating environment). The domain of international management – as opposed to the domain of management or the domain of international business – should therefore be identiﬁed based on the value created by managerial actions across national borders. Without denying that domestic management is complicated, international management is often distinguished from domestic management by its additional complexities. For example, while domestic management requires a vast knowledge base in accounting, ﬁnance, production, human resources, marketing, economics, law, public policy, behavioral psychology, and so on, international management requires a commensurately larger knowledge base as the traditional areas become internationalized (and thus become part of the domain of international business) to account for additional governments, legal systems, cultural orientations, languages, and the like. In addition, international management requires knowledge of disciplines typically ignored in domestic business, such as international political economy, foreign affairs, and geographic area studies, which are important components of international business. Hence, the complexities that distinguish international management from domestic management result from operating across borders and having to deal with multiple economies, societies, and governments. Mastering international management can be a Herculean task if the domain is too broad, so some priorities must be set to focus on the most germane activities of international value creation. The focus on valuecreating activities is in fact what distinguishes international management from the broader concept of international business. This paper adopts a strategy approach to international management by considering a back-to-basics framework in which international management is the aggregation of value created through international production, marketing, and ﬁnancial activities. The framework is not a simple internationalization of the functional areas of business, but is instead a broadening of the perspective to analyze production, marketing, and ﬁnancial activities in a larger global context. Other elements of international business can be put into this framework if they create value. For example, foreign languages and geographic area

102

REID W. CLICK

studies are often included in the domain of international business. The framework presented here requires speciﬁcation of the value created by having such managerial knowledge to be included in the more exclusive domain of international management. If knowledge of a foreign language facilitates management–labor communication at a foreign plant and thereby increases production efﬁciency, or if the talent improves relations with customers and thereby increases sales revenue abroad, it creates value and is within the domain of international management. However, the importance of the knowledge is directly related to the value created, and foreign languages and area studies are not automatically part of the domain of international management. The concept of value creation is used in this paper to answer three questions pertaining to strategic management of multinational enterprises (MNEs). After this introduction, Section 2 broadly examines value creation and the question, ‘‘How important is international strategic management?’’ Section 3 develops the strategy framework for analyzing international management and asks ‘‘What is the domain of international strategic management?’’ Discussions of international production, marketing, and ﬁnancial activities suggest that international management requires an outward-looking perspective rather than an inward-looking perspective to maximize the value of the MNE, and reveal that the domain of international management is very broad. Section 4 ties the elements of value creation strategy together by asking, ‘‘Does international strategic management make the whole multinational enterprise worth more than the sum of its parts?’’ Empirical evidence from U.S. multinationals in the early 1990s indicates that the answer is yes.

2. VALUE CREATION AND THE RELATIVE IMPORTANCE OF INTERNATIONAL STRATEGIC MANAGEMENT Production and marketing operations represent the two main areas of business activity. Porter (1985) described these as the ‘‘primary activities’’ in the ﬁrm’s ‘‘value chain.’’ The production operations are the ‘‘upstream activities,’’ such as inbound logistics, operations, and the initiation of outbound logistics, which represent the ‘‘cost side’’ or ‘‘supply side’’ of the ﬁrm. The marketing operations are the ‘‘downstream activities,’’ such as the distribution activities of outbound logistics, marketing and sales, and after-sales service, which represent the ‘‘revenue side’’ or ‘‘demand side’’ of the ﬁrm.

Value Creation Perspective of International Strategic Management

103

Often, the objective of managing the supply side is to minimize the costs of producing goods and services, and this requires analysis of competitors and industry structures. A ﬁrm with production advantages, such as a proprietary production technology that utilizes lower quantities of inputs than competitors, is a low-cost producer in the industry and is able to pursue a ‘‘cost-based strategy.’’ The objective of managing the demand side is to maximize the revenues from selling goods and services, which requires analysis of customers and markets. A ﬁrm with a marketing advantage, such as brand recognition or unique products, emerges with market power and is able to pursue a ‘‘differentiation strategy’’ or ‘‘revenue-based strategy.’’ Production and marketing decisions clearly need to be mutually consistent to produce an equilibrium. For additional discussion, see Besanko, Dranove, Shanley, and Schafer (2003). Discussion of value creation in international strategic management must also distinguish international production and marketing from domestic production and marketing. In considering the internationalization of the value chain, Porter (1986a) more speciﬁcally asserts: The distinctive issues in international, as contrasted to domestic, strategy can be summarized in two key dimensions of how a ﬁrm competes internationally. The ﬁrst is what I term the configuration of a ﬁrm’s activities worldwide, or where in the world each activity in the value chain is performed, including how many places. The second dimension is what I term coordination, which refers to how like activities in different countries are coordinated with each other. (p. 17)

Porter then considers various combinations of geographically dispersed or concentrated conﬁgurations versus high or low coordination of activities. By the end of the article, Porter (1986a) suggests that ‘‘today’s game of global strategy seems increasingly to be a game of coordination – getting more and more dispersed production facilities, R&D laboratories, and marketing activities to truly work together’’ (p. 36). These ideas are further developed in Porter (1986b, 1990). In a variation on the issue of coordination, Kogut (1985) asserts that international strategic management is distinguished from domestic management by the importance of ﬂexibility: ‘‘the unique content of a global versus a purely domestic strategy lies less in the methods to design long-term strategic plans than in the construction of ﬂexibility which permits a ﬁrm to exploit the uncertainty over future changes in exchange rates, competitive moves, or government policy’’ (p. 27). Kogut then considers coordination of activities internationally to reap the beneﬁts of ﬂexibility, not just to manage risk and uncertainty, but to proﬁt from them as well. There has subsequently

104

REID W. CLICK

been considerable attention to pre-planned ﬂexibility and international strategic investments; see, for example, de Meza and van der Ploeg (1987), Kogut and Kulatilaka (1994), and Amran and Kulatilaka (1999). The ideas presented in Porter and Kogut relating to conﬁguration, coordination, and ﬂexibility suggest that strategic management of MNEs creates value. Many other articles have also made this assertion in one way or another, including the seminal works of Hamel and Prahalad (1985), Prahalad and Doz (1987), and Ghoshal and Bartlett (1990). The subject begs the question, ‘‘How important is international strategic management?’’ Although difﬁcult to answer deﬁnitively, some simple statistics illuminate the importance. A ﬁrst way to consider the importance of international management relative to domestic management is to examine macroeconomic aggregates. Net domestic product (NDP), a measure of valued created by all factors in a particular country, amounted to $10.3 trillion in the U.S. in 2004 (U.S. Department of Commerce, 2005). Factor income from abroad, which is predominantly earnings on direct investment, was $53.7 billion, representing 0.5% of NDP. Hence, the pure foreign value creation does not appear very large, but further examination suggests that international activities are in fact important. Table 1 shows the breakdown of NDP into its basic components, and corporate proﬁts (before taxes) amount to $1161.5 billion. (Note that this ﬁgure is not a measure of economic value added because part of corporate proﬁts represents the required rate of return to shareholders.) Factor income from abroad thus represents 4.6% of total corporate proﬁts. In addition, these ﬁgures exclude the contributions of exports to domestic proﬁts and the contribution of imported inputs in reducing production costs. Export receipts amounted to 11.4% of NDP, and although the value created from these exports cannot be determined, the gross magnitude suggests that international management is indeed important. Some value is also created by using imported inputs to reduce production costs. Imports amounted to 17.5% of NDP, and although the portion attributable to inputs cannot be determined, this gross magnitude again suggests that international management is important. A second way to evaluate the relative importance of international strategic management is to examine ﬁrm-level data. Data are widely available for the subset of all ﬁrms that are publicly traded, and although this subset likely contains more MNEs than the subset of ﬁrms not publicly traded, the available data are useful in analyzing value creation. In the Compustat database (Standard and Poor’s Corporation, 1997), 6,345 nonﬁnancial U.S. ﬁrms report operating proﬁts for 1994 amounting to $503.1 billion. Of these

Value Creation Perspective of International Strategic Management

105

Table 1. Value Creation in the U.S. Based on Net Domestic Product, 2004. Type of Income Compensation of employees Corporate proﬁts Proprietors’ and rental income Net interest and dividends Net taxes on production and imports, net business current transfer payments, and statistical discrepancy Net domestic product (NDP) Factor income from abroad Exports of goods and services Imports of goods and services

Billions of Dollars

Percent of Total

6693.4 1161.5 1023.8

65.0 11.3 9.9

446.0 974.3

4.3 9.5

10,299.0

100.0

53.7 1173.8

0.05 11.4

1797.8

17.5

Source: U.S. Department of Commerce, Survey of Current Business, August 2005, pp. 36–172.

ﬁrms, 1,215 report operating proﬁts of foreign operations in addition to total operating proﬁt, and operating proﬁts from foreign operations amount to $74.3 billion from total operating proﬁts of $261.8 billion, or 28.4%. The $74.3 billion in proﬁts from foreign operations represents 14.8% of the aggregate proﬁts of $503.1 billion, although this understates the relative proﬁts from foreign operations because not all ﬁrms with foreign operations report proﬁts from foreign operations when they report total proﬁts. In addition, the ﬁgure does not include proﬁts from exports, licensing/franchising agreements, or any form of international business other than direct foreign investment. A third way of capturing the importance of international strategic management is to more broadly consider the sales revenues of the U.S. ﬁrms. A total of 6,345 nonﬁnancial ﬁrms report sales for 1994 amounting to $5 trillion. Table 2 presents some decomposition of this, indicating that 1,634 ﬁrms report export sales and 1,341 ﬁrms report sales from foreign operations. Because only 563 ﬁrms report both export sales and sales from foreign operations, a total of 2,412 ﬁrms report some form of foreign sales. Hence, 38% of the 6,364 ﬁrms are labeled MNEs and 62% appear to be purely domestic corporations. Since MNEs are larger, however, they account for

106

REID W. CLICK

Table 2.

Sales of U.S. Corporations, 1994.

Sales Export sales Sales from foreign operations Total foreign sales Total multinational enterprise sales Total domestic enterprise sales

Number of Firms

Millions of Dollars

6,364 1,634 1,341 2,412 2,412 3,952

5,124,123 149,430 899,993 1,049,423 3,059,468 2,064,655

Source: Compustat and author’s calculations.

60% of total sales and domestic ﬁrms account for 40%. Foreign sales reported by the 6,364 ﬁrms amount to $1 trillion, or 20% of total sales. Among the 2,412 MNEs, the foreign sales account for 34.3% of total sales. Furthermore, the data are again understated because not all ﬁrms with exports or foreign operations report segment sales. Of course, the focus on sales highlights the marketing side of the business and ignores the production side, but comparable data on costs are much harder to get than the data on revenues and so cannot be examined here. Financial activities also create value, although these are a little more suspicious than value creation through production and marketing activities since ﬁnancial management does not obviously contribute either to the sales revenues of the ﬁrm or savings in the production costs of the ﬁrm. Porter (1985, 1986a, 1986b) implicitly considers ﬁnancial management as a ‘‘support activity’’ rather than as a primary activity and hardly addresses the issue. However, some of the most interesting questions in management relate to whether value is created through ﬁnancial advantages, debt capacity, risk management, and so on. In addition, there are ﬁrms that exclusively create value through ﬁnancial activities – such as banks and brokerage houses. Recognizing the difﬁculty of subjecting these ﬁnancial ﬁrms to the same ‘‘supply side’’ and ‘‘demand side’’ analysis as nonﬁnancial ﬁrms, this paper considers the activities of ﬁnancial ﬁrms in the category of ﬁnancial activities. Hence, two categories of value created through ﬁnancial activities are the ﬁnancial activities of nonﬁnancial ﬁrms and the activities of ﬁnancial ﬁrms. One way to assess the importance of international ﬁnancial activities is to examine the performance of U.S. ﬁnancial ﬁrms. In the Compustat database, the operating proﬁts of 714 ﬁnancial ﬁrms amount to $137 billion. Of these, 109 ﬁrms report proﬁts from foreign operations at $4.6 billion from total proﬁts of $61.1 billion, or 7.5%, which also represents 3.4% of total ﬁnancial industry proﬁts. Thus, international ﬁnancial management might

Value Creation Perspective of International Strategic Management

107

be marginally important, but this ﬁgure again understates the importance of proﬁts on foreign involvement because it does not capture the proﬁt on international capital ﬂows booked at home – such as proﬁts on foreign deposits in the U.S. or proﬁts on foreign loans made from the U.S., and similar transactions in the bond and stock markets. International capital ﬂows in the form of intermediated investment and portfolio investment are quite important compared to direct foreign investment, and ﬁrms managing these international capital ﬂows are likely creating value through ﬁnancial activities. Based on these simple statistics for the U.S., we can conclude that strategic management of multinational ﬁrms is indeed important. Although it is not true that ‘‘all management is international’’ in the U.S., international management is important in terms of its contributions at the margin when compared to domestic management, and is therefore well worth investigating.

3. VALUE CREATION AND THE DOMAIN OF INTERNATIONAL STRATEGIC MANAGEMENT Having discussed value creation in terms of production, marketing, and ﬁnancial activities, this section combines these three categories into a basic strategy framework designed to examine the domain of international strategic management. Discussion is particularly focused on identifying the additional complexities resulting from crossing national borders that distinguish international management from domestic management. Each one of these areas could constitute a paper (or book) on its own, so the discussion is general rather than exhaustive, and references to additional literature are provided to ﬁll gaps. A framework for analyzing strategic management of MNEs is presented in Table 3. This summarizes the distinctive issues in international strategic management, elements in the domain of international strategic management, and some representative studies. The table suggests that areas in the domain of international production are international trade theory (comparative advantage), international political economy, foreign affairs, and geographic area studies (political risk), international ﬁnancial theory (exchange rates and risk), and area studies, law, sociology, anthropology, religion, psychology, and language (applied to human resources management). Elements in the domain of international marketing are geographic area studies, sociology, anthropology, religion, and psychology (cultural

108

Table 3.

A Strategy Framework for Analyzing International Management.

Distinctive Issues in International Management

The Domain of International Management

Representative Studies

Production activities

Source inputs domestically or internationally, including supplier relations Produce domestically or internationally International alliance (licensing, franchising, joint venture) versus wholly owned DFI International site selection decisions, including number and size of plants Input mix abroad (capital, unskilled labor, skilled labor, raw materials) Foreign production technology and equipment selection Human resources management in foreign countries Inventory management and crossborder transportation logistics R&D in process technologies

International trade theory, including comparative advantage, analysis of transportation costs, tariffs, and nontariff barriers International political economy and foreign affairs Geographic area studies International ﬁnancial theory, particularly exchange rate behavior and risk analysis International law Sociology/anthropology/religion Psychology Languages

Stobaugh (1969a, 1969b) Flaherty (1986) Flaherty (1996) Dunning (1988) Ferdows (1997) Karrenbrock (1990) Hofstede (1984) Adler (1996) Quelch and Bloom (1999)

REID W. CLICK

International Value Creation Activities

International market selection, including number of foreign markets Market service and distribution methods (choosing exporting, international alliances, or DFI) Development of product attributes in different markets, including R&D into product technologies Product pricing internationally, including currency effects Promotional activities internationally, including language translation

Cultural studies Geographic area studies Sociology/anthropology/religion Psychology International ﬁnancial theory, particularly exchange rate determination and behavior Languages

Levitt (1983) Quelch and Hoff (1986) Douglas and Craig (1989) Hout, Porter, and Rudden (1982) Kashani and Quelch (1990) Ricks (1993) Knetter (1994)

Financial activities

Foreign project evaluation (international capital budgeting) International cost-of-capital analysis and capital structure decisions, including capital structure of foreign subsidiaries Risk management for foreign exchange risk and foreign interest rate/inﬂation rate risk, including diversiﬁcation and hedging strategies International ﬁnancing and the debt denomination decision International tax management

International macroeconomics International ﬁnancial theory, particularly exchange rate and international interest rate behavior and risk analysis International political economy and foreign affairs Geographic area studies International portfolio theory International ﬁnancial markets International tax law

Shapiro (1983) Kogut and Kulatilaka (1994) Adler and Dumas (1984) Lessard and Lightstone (1986) Dufey and Srinivasulu (1983) Aliber (1989)

Value Creation Perspective of International Strategic Management

Marketing activities

109

110

REID W. CLICK

determinants of product demand, product attributes, and promotional programs), international ﬁnancial theory (exchange rates and pricing in different currencies), and language (promotional activities). The domain of international ﬁnancial activities includes international political economy, foreign affairs, and geographic area studies (political risk), international ﬁnancial theory and international macroeconomics (exchange rates and interest rates), international tax law (focusing on ﬁnancing), and international portfolio theory and international ﬁnancial markets (risk management and hedging). Clearly, the domain of international strategic management is quite broad even under the criterion of value creation. Indeed, the diversity of international management is one appealing aspect that draws many managers into the ﬁeld. References to additional literature are provided since space limitations preclude additional discussion here.

4. DOES INTERNATIONAL STRATEGIC MANAGEMENT MAKE THE WHOLE MULTINATIONAL ENTERPRISE WORTH MORE THAN THE SUM OF ITS PARTS? By this point, the notion that international management creates value in ways somewhat different from domestic management is ﬁrmly established. After the discussion of production, marketing, and ﬁnance activities covered in the strategy framework of Section 3, a return to the more general issue of value creation is now warranted. Rather than focus on speciﬁc tasks that distinguish international management from domestic management, recall that Porter and Kogut discuss the distinction in more generic terms focusing on coordination and ﬂexibility. In addition to indicating that international management creates value, both imply that the whole of the MNE is worth more than the sum of its parts; having n international subsidiaries is worth more as an MNE than the summation of the separate values of the n subsidiaries, as long as they are coordinated and are ﬂexible enough to proﬁt from changes in the environment. Several reasons why the whole may be worth more than the sum of its parts have been developed in the international management literature. Kogut (1985) identiﬁes several ‘‘arbitrage’’ opportunities for international management to create value: production shifting, tax minimization, ﬁnancial market imperfections, and information arbitrage. Kogut also identiﬁes a

Value Creation Perspective of International Strategic Management

111

couple of ‘‘leverage’’ opportunities: coordination and political risk management. Proﬁts from arbitrage result from ‘‘moving’’ something from one place to another, but proﬁts from leverage result from using a position in one national market to enhance a position in another market. For example, global coordination can build a coalition of suppliers in one country to improve the position of the ﬁrm with respect to a rival in another, and with regard to political risk the position of a subsidiary in one country is likely to affect the ﬁrm’s bargaining power vis-a`-vis the government in another. A second, more strategic reason for the whole to be worth more than the sum of the parts is due to cross-subsidization of operations. Hamel and Prahalad (1985) provide the example of competition in the tire industry: Michelin attacked Goodyear’s U.S. market by ﬁnancing an aggressive marketing campaign using the proﬁts from its European market, but rather than conﬁne the competition to the U.S. Goodyear attacked Michelin’s European market by ﬁnancing market expansion in Europe using the proﬁts from its U.S. market. International cash ﬂows may thus create value, such that Michelin (or Goodyear) is worth more than a simple summation of its operations in Europe and the U.S. would suggest because one operation can be defended by the other. (Without the second operation, the ﬁrst would have to sink or swim on its own.) Necessarily, the cross-subsidization must not entail throwing good money into an unproﬁtable situation, but must be a temporary response to a situation that retains a positive present value. Economies of scope from global distribution systems and synergies from knowledge created in different environments provide additional reasons for an MNE to be worth more than the sum of its parts; see, for example Ghoshal (1987) and Douglas and Craig (1989). Economies of scope create value due to an ability to share investments across products and markets, as when two products sharing one distribution channel have a lower cost of joint distribution than the total cost of distributing the two products separately. Synergy is the ability to make two operations work better together than they would separately, and can result from learning; a production operation in one country might be improved by knowledge developed in a second country, and simultaneously the production operation in the second country might be improved by knowledge from the ﬁrst country. Given all this theoretical discussion of the value of the whole in relation to the sum of the parts, a natural question to ask is whether there is empirical evidence that the whole of an MNE is worth more than the sum of its subsidiary parts. If so, this is further testimony to the importance of studying international management. Unfortunately, the question has not been

112

REID W. CLICK

extensively researched. In a purely domestic setting, empirical evidence in Berger and Ofek (1995) and Comment and Jarrell (1995) reveals that product diversiﬁcation destroys value because conglomerates lose focus, but literature considering international diversiﬁcation is relatively scarce (however, see Hitt, Hoskisson, & Kim, 1997; Quian & Li, 1998). One way to approach the question is to consider whether an MNE creates more value than otherwise similar domestic ﬁrms. If so, the additional shareholder value is most likely created through international strategic management, the activities that otherwise similar domestic ﬁrms do not undertake. 4.1. Data and Empirical Methodology Data are widely available for the subset of all ﬁrms that are publicly traded. This study utilizes the Compustat database of publicly traded U.S. ﬁrms to compare ﬁrms engaged in international business to their domestic counterparts. In particular, it relies on the geographic segment ﬁles to distinguish exports from sales of foreign operations and to identify the regions in which sales from foreign operations occur. The Compustat data are likely the highest quality available for U.S. ﬁrms. However, they understate the true importance of multinational operations because not all ﬁrms with exports or foreign operations report segment sales. Since there is no way to correct for this, analysis proceeds under the assumption that any understatement is small. Of course, the focus on sales highlights the marketing side of the business and ignores the production side, but comparable data on costs are much harder to obtain than the data on revenues and so cannot be examined here. Although there are many possible measures of ﬁrm performance, the most common measure in the ﬁeld of ﬁnance is the total return to shareholders (capital appreciation and dividend payments). This return is construed as a summary measure of all the information available on the ﬁrm – including all other performance measures such as the return on assets (ROA) frequently studied in the management literature, as in Hitt et al. (1997). In ﬁnance, the total return is typically described as an increasing function of the ﬁrm’s ‘‘beta’’ from the Capital Asset Pricing Model (CAPM). In a series of papers, Fama and French (1992, 1993, 1995) show that for U.S. data, beta in fact does not contain much cross-section information when ﬁrm size and the ratio of book equity to market equity are included as determinants of returns. Financial leverage may have additional explanatory value as well. This paper therefore controls for all of these variables in the investigation of the effects of international business.

Value Creation Perspective of International Strategic Management

113

One way to determine whether an international enterprise has different performance than an otherwise similar domestic ﬁrm is to determine whether measures of international involvement are signiﬁcant determinants of total return, controlling for other characteristics known to affect ﬁrm performance. This paper speciﬁcally considers ﬁve variables measuring international involvement based on exports, sales from foreign operations, and the geographic diversiﬁcation of sales, and considers their contributions to total returns in cross-section regressions controlling for beta, ﬁrm size, the ratio of book equity to market equity, and ﬁnancial leverage. The measures of international involvement are derived from the Compustat data described above. The simplest type of international involvement is to export some domestic output. Bernard and Jensen (1999) revealed that exporting ﬁrms have superior performance when compared to nonexporting ﬁrms. Based on this, the empirical analysis here considers the effect of exporting on ﬁrm returns by using the Compustat data on export sales to form both a dummy variable for ﬁrms reporting exports and the ratio of export sales to the ﬁrm’s total sales (the ‘‘export ratio’’). The dummy variable will capture the differential return for exporting ﬁrms vis-a`-vis nonexporting ﬁrms. If exporting ﬁrms have higher (lower) returns than nonexporting ﬁrms, the coefﬁcient on the dummy variable will be positive (negative). The export ratio will indicate whether the relative magnitude of exports contributes to differential performance. If exporting ﬁrms have higher (lower) returns than nonexporting ﬁrms, returns might also be an increasing (decreasing) function of the proportion of exports in total sales. A more complicated type of international involvement is to operate facilities in foreign countries, and this is what is typically studied to help assess whether ‘‘multinationality’’ affects ﬁrm performance. The literature in ﬁnance is generally inconclusive with respect to the empirical effects of multinationality on total returns. One of the earliest studies, Mikhail and Shawkey (1979), reports that multinational shares outperform domestic shares. However, Brewer (1981) subsequent study reports no signiﬁcant difference in returns between multinational and domestic ﬁrms. After that, Fatemi (1984) reports that returns are identical between multinational and domestic ﬁrms except when the multinational operates in competitive foreign markets, in which case shareholder returns are lower for multinationals than for domestic ﬁrms. Subsequent research is equally inconclusive, so there is no consensus in the profession on the effects of multinationality on shareholder returns. Empirical analysis here uses the Compustat data on sales of foreign operations to form both a dummy variable for ﬁrms with foreign operations

114

REID W. CLICK

and the ratio of sales from foreign operations to the ﬁrm’s total sales (the ‘‘multinational sales ratio’’). If MNEs have higher (lower) returns than otherwise similar purely domestic ﬁrms, the coefﬁcient on the dummy variable will be positive (negative). In addition, the degree of multinationality might be important, such that returns might be an increasing (decreasing) function of the proportion of sales from foreign operations in relation to total sales. The dummy variable and ratio for sales from foreign operations capture basic information as to whether a ﬁrm is an MNE and the overall degree of its multinationality. However, the variables do not capture the degree to which the ﬁrm is diversiﬁed with respect to its multinational activities. The value of multinationality is often associated with an integrated empire of operations, such that the whole is worth more than the sum of the parts. To the extent that international involvement makes the whole of the MNE worth more than the sum of its parts, it is because there are many parts, in many geographic areas throughout the world. Hence, the degree of geographic diversiﬁcation is potentially the single most important determinant of shareholder returns due to international involvement. The most common measure of geographic diversiﬁcation is the entropy measure associated with Theil (1967): ENTROPY ¼

n X

S i ln ð1=S i Þ ¼

i¼1

n X

S i ln ðS i Þ

i¼1

where Si represents the share of operations in the ith region relative to total operations. Sales are typically used as the measure of operations. Compustat reports 1994 sales data broken into seven geographic regions (U.S., non-U.S. North America, South America, Europe, Asia, Oceania, and Africa) for 1,221 ﬁrms. With seven geographic regions, entropy theoretically ranges from 0, which represents sales entirely in one region, to a maximum of 1.95, representing sales equally distributed among the seven regions. For the 1,221 ﬁrms reporting geographic segment data, entropy empirically ranges from 0 to 1.37 and has a mean of 0.567. A central feature of the entropy statistic is that it contains information not contained in the multinational sales ratio, as distinctions can be made among ﬁrms at a particular multinational sales ratio based on the dispersion of the sales from foreign operations. Fig. 1 plots entropy against the multinational sales ratio, revealing of an inverted U shape. A regression estimating entropy as a quadratic function of the foreign sales ratio reveals: ENTROPY ¼ 3:479 RATIO 3:271 RATIO2 ð:260Þ

ð:400Þ

Value Creation Perspective of International Strategic Management

115

1.4

ENTROPY

1.2 1 0.8 0.6 0.4 0.2 0 0

0.2 0.4 0.6 0.8 1 1.2 RATIO OF SALES OF FOREIGN OPERATIONS TO TOTAL SALES

Fig. 1. Entropy as a Function of the Ratio of Sales of Foreign Operations to Total Sales.

with an adjusted R2 ¼ 0.79. This equation implies that on average entropy reaches a maximum at 53% multinational sales. To determine the signiﬁcance of international business the regressions control for other ﬁrm characteristics known to determine shareholder return, at least ex post. The ﬁrst is the CAPM beta, measuring the risk of the stock against the marketwide portfolio. Compustat reports stock betas estimated by Standard & Poor’s Corporation with ﬁve years of monthly data. As an independent variable, beta is expected to have a positive coefﬁcient, such that riskier ﬁrms have higher returns. A lagged value of beta is used in order to reduce endogeneity. Guided by ﬁndings from Fama and French, several additional control variables are included. Market size is the log of the lagged values of total market equity, which is expected to have a negative coefﬁcient when small ﬁrms outperform large ﬁrms. The ratio of the book equity to market equity of the ﬁrm is the log of the lagged value, and is expected to have a positive coefﬁcient since ﬁrms with high book equity in relation to market equity outperform ﬁrms with low book- to market-equity. Financial leverage is the lagged value of the ratio of long-term debt to the sum of long-term debt and the market value of equity. The leverage variable is included in both level and squared form, which are expected to have positive and negative coefﬁcients, respectively, suggesting that increasing leverage raises returns but at a decreasing rate and will even lower returns after some maximum.

116

REID W. CLICK

Regressions also control for industry using a set of 13 industry dummy variables based on SIC code. This setup allows the constant in the regressions to vary by industry, since the industrial composition of MNEs is generally shown to be different from that of domestic corporations. The industries covered are agriculture; mining and construction; food and tobacco; textiles and apparel; lumber, furniture, and paper; chemicals; rubber, leather, stone, and glass; metals; machinery, computers, and electronics; transportation equipment; transportation and communications; wholesale and retail trade; and services. 4.2. Results Tables 4 and 5 present cross-section regression results examining determinants of the return to shareholders of nonﬁnancial ﬁrms. Table 4 contains results using the return for 1994 in a sample of 3,678 ﬁrms for which data are available. Table 5 contains results using the annual average return for the ﬁve-year period 1990–1994 in a sample of 2,891 ﬁrms. The independent variables capturing international involvement are based on data for 1994. The dummy variables and ratios for export sales and foreign sales are as described above. Entropy is based on the formula given above; for any ﬁrms reporting zero sales from foreign operations, entropy is set equal to zero since all their sales therefore occur in the U.S. To conserve space, Tables 4 and 5 report fairly comprehensive equations containing the main combinations of the ﬁve variables representing international involvement. Many other combinations of variables were considered, but are not reported here because the results are consistent with these main combinations. Note that, as is typical in studies of returns, the adjusted R2 of the equations is fairly low – at approximately 11% for 1994 and 6% for 1990–1994. To determine the signiﬁcance of international involvement, the regressions control for other ﬁrm characteristics usually thought to determine shareholder return. Prominent among these variables is beta. For regressions in Table 4 using the 1994 return, beta is based on the period 1990– 1994. For regressions in Table 5 using the 1990–1994 returns, beta is based on the period 1985–1989. The instability of coefﬁcients on beta in crosssection regressions is reﬂected in the results presented in Tables 4 and 5. Using the single-year returns for 1994, beta is shown to have a negative coefﬁcient, which is an ‘‘incorrect’’ sign according to the CAPM, and is statistically signiﬁcant. For average annual returns for 1990–1994, however, the sign is positive and signiﬁcant. According to these results, 1994 was thus

1 Dummy for exports Dummy for multinational sales Export sales ratio Multinational sales ratio Entropy Beta ln(market equity) ln(book/market equity) Leverage Leverage2 Observations Adjusted R2 SEE

8.258 (2.081) 5.899 (2.025)

0.472 (0.133) 4.278 (0.519) 18.182 (1.572) 10.833 (12.815) 25.315 (20.922) 3,678 0.11 50.340

2

3

21.808 (8.417) 10.366 (5.908)

20.870 (8.408)

0.433 (0.135) 4.542 (0.508) 18.531 (1.572) 14.150 (12.932) 29.632 (21.015) 3,676 0.11 50.440

0.412 (0.136) 4.774 (0.477) 18.608 (1.567) 14.822 (12.978) 30.175 (21.074) 3,678 0.11 50.460

Note: Regressions also include industry dummy variables. Signiﬁcant at 10% level. Signiﬁcant at 5% level.

4

9.361 (5.905) 0.424 (0.133) 4.530 (0.508) 18.590 (1.573) 11.068 (12.911) 26.623 (20.996) 3,676 0.11 50.487

5

12.531 (3.199) 0.481 (0.128) 4.440 (0.587) 18.831 (1.758) 11.597 (14.246) 30.294 (22.928) 2,974 0.12 50.552

Value Creation Perspective of International Strategic Management

Table 4. Determinants of Shareholder Returns, 1994 (Heteroscedasticity-Consistent Standard Errors in Parentheses).

117

118

Table 5. Determinants of Average Annual Shareholder Returns, 1990–1994 (Heteroscedasticity-Consistent Standard Errors in Parentheses). 1 Dummy for exports Dummy for multinational sales Export sales ratio Multinational sales ratio Entropy Beta ln(market equity) ln(book/market equity) Leverage Leverage2 Observations Adjusted R2 SEE

2

3

4

5

1.177 (0.873) 1.515 (0.886) 4.033 (3.589) 4.340 (2.408) 1.742 0.439 5.063 13.472 25.519 2,891 0.06 19.749

(0.778) (0.199) (0.608) (5.969) (9.095)

1.755 0.459 5.078 14.426 26.743 2,889 0.06 19.744

1.790 0.564 5.137 14.041 26.320 2,891 0.06 19.759

(0.779) (0.185) (0.608) (5.989) (9.112)

4.146 (2.398) 1.794 0.451 5.088 14.074 26.495 2,889 0.06 19.745

(0.779) (0.195) (0.608) (5.967) (9.100)

2.891 (1.149) 1.825 (0.871) 0.478 (0.230) 5.459 (0.696) 12.565 (5.967) 26.399 (10.447) 2,335 0.07 20.002

REID W. CLICK

Note: Regressions also include industry dummy variables. Signiﬁcant at 10% level. Signiﬁcant at 5% level.

(0.778) (0.195) (0.607) (5.982) (9.105)

3.630 (3.575)

Value Creation Perspective of International Strategic Management

119

a year in which riskier stocks performed poorly, although the period 1990– 1994 was generally one in which riskier stocks outperformed safer ones. The Fama and French variables have consistent coefﬁcients, but not necessarily conforming to earlier ﬁndings. The coefﬁcient on size is positive in both tables, indicating that larger ﬁrms outperformed small ﬁrms during the early 1990s. This sign contradicts the Fama and French ﬁndings, although it is not unusual because the effect of size on ﬁrm returns is the subject of some debate in the ﬁnance literature. The positive sign on the ratio of book equity to market equity is the ‘‘right sign.’’ This ratio is typically regarded as a more robust determinant of returns in the ﬁnance literature, so the result somewhat reafﬁrms this robustness. The effect of leverage is also as expected: there is a positive coefﬁcient on the variable and a negative coefﬁcient on its square, and the variables are jointly signiﬁcant. The maximum return occurs in the range of 19–27% leverage. The signiﬁcance of international involvement as a determinant of shareholder returns is demonstrated by the results provided in the upper rows of Tables 4 and 5. Note that all of the signs are positive, indicating that shareholder returns are consistently positively related to international involvement. Nearly all of the coefﬁcients in Table 4 are statistically signiﬁcant at the 5% level, and only the multinational sales ratio seems to be a relatively weak determinant. The coefﬁcients in Table 5 are more often insigniﬁcant or are only weakly signiﬁcant (i.e., only at the 10% level). Exporting ﬁrms appear to have superior performance when compared to nonexporting ﬁrms for 1994, though not for the 1990–1994 period as a whole. For 1994, the dummy variable suggests that exporters had returns 814 percentage points above nonexporters. The export sales ratio suggests that returns were an increasing function of the proportion of exports, such that each percentage point of exports increased returns by 21–22 basis points. MNEs appear to have superior performance when compared to domestic enterprises. The results for 1994 are stronger, both in magnitude and statistical signiﬁcance, but the results for 1990–1994 still afﬁrm the ﬁnding. For 1994, the dummy variable suggests that multinationals had returns nearly 6 percentage points higher than domestics, and the comparable ﬁgure for 1990–1994 is 112 percentage points. Although weaker than the results for the dummy variable, the coefﬁcients on the multinational sales ratio suggest that returns were an increasing function of multinationality, such that each percentage point of sales from foreign operations increased returns by about 10 basis points in 1994 or 4 basis points during 1990–1994. Among all the variables capturing international involvement, entropy might be the most important. The coefﬁcient on entropy is statistically

120

REID W. CLICK

signiﬁcantly positive at the 5% level in both tables, suggesting that multinational diversiﬁcation increases the total shareholder return. The average ﬁrm with an entropy value of 0.567 had a return 7.1 percentage points higher in 1994, and an average return 1.6 percentage points higher over 1990–1994, than an otherwise similar domestic ﬁrm (i.e., a ﬁrm in the same industry, with the same beta, of the same size, with the same book- to market-equity ratio, and the same leverage). This is strong evidence that the value of international involvement is tied to the degree of diversiﬁcation in the ﬁrm, such that a MNE as a whole is worth more than the sum of its subsidiary parts. Further discussion of this is provided in the conclusion.

5. SUMMARY International strategic management is the process through which value is created by an individual or group of individuals operating across a national border. This deﬁnition thus distinguishes international strategic management from general management with the phrase ‘‘operating across a national border,’’ and distinguishes international management from international business by being focused on the ‘‘process through which value is created’’ internationally. This paper answers three questions from the premise that skills creating the most value in cross-border management are the most important elements in the domain of international strategic management. First, how important is international management? The relative importance of international management is determined by the value created internationally compared to that created domestically, and Section 2 presents several statistics that reveal the importance of international activities. Factor income from abroad, which is predominantly proﬁts on direct foreign investment, represents 4.6% of total corporate proﬁts in the U.S. Among publicly traded nonﬁnancial ﬁrms, sales and proﬁts from foreign operations amount to 34% and 24% of corporate sales and proﬁts (respectively) for MNEs, and 20% and 15% of aggregate sales and proﬁts of all U.S. ﬁrms. In addition, ﬁnancial ﬁrms create value from their international ﬁnancial activities. Second, what is the domain of international management? Since the domain is potentially unbounded under the deﬁnition provided here, this paper considers a strategy approach in which international management is the aggregation of value created through international production, marketing, and ﬁnancial activities. Section 3 develops this framework and uses it to

Value Creation Perspective of International Strategic Management

121

deﬁne the core domain of international management. The conclusion is that the domain of international management is quite broad even under the criterion of value creation. Indeed, the diversity of international management is one appealing aspect that draws many managers into the ﬁeld. Third, does international management make the whole multinational corporation worth more than the sum of its parts? Effective management of production, marketing, and ﬁnancial activities is capable of creating more value in an MNE than in an otherwise similar purely domestic ﬁrm. Section 4 presents some preliminary empirical evidence that this is indeed the case for U.S. ﬁrms by showing that shareholder returns are positively related to measures of international involvement (and, by inference, the amount of international management). In particular, shareholder returns are positively related to the ﬁrm’s geographic diversiﬁcation; the average ﬁrm with an entropy value of 0.567 had a return 7.1 percentage points higher in 1994 (and an average return 1.6 percentage points higher over 1990–1994) than an otherwise similar domestic ﬁrm.

REFERENCES Adler, M., & Dumas, B. (1984). Exposure to currency risk: Deﬁnition and measurement. Financial Management, Summer, 41–50. Adler, N. (1996). International dimensions of organizational behavior (3rd ed.). Cincinnati, OH: South-Western College Publishers. Aliber, R. Z. (1989). The debt denomination decision. In: R. Z. Aliber (Ed.), The handbook of international financial management (pp. 435–451). Burr Ridge, IL: Dow Jones-Irwin. Amran, M., & Kulatilaka, N. (1999). Real options: Managing strategic investment in an uncertain world. Boston, MA: Harvard Business School Press. Berger, P., & Ofek, E. (1995). Diversiﬁcation’s effect on ﬁrm value. Journal of Financial Economics, 37, 39–65. Bernard, A., & Jensen, J. (1999). Exceptional exporter performance: Cause, effect, or both? Journal of International Economics, 47, 1–25. Besanko, D., Dranove, D., Shanley, M., & Schafer, S. (2003). Primer: Economic concepts for strategy. The Economics of Strategy, 1–37. Boddewyn, J. J. (1999). The domain of international management. Journal of International Management, 5, 3–14. Brewer, H. (1981). Investor beneﬁts from corporate international diversiﬁcation. Journal of Financial and Quantitative Analysis, 16, 113–126. Comment, R., & Jarrell, G. (1995). Corporate focus and stock returns. Journal of Financial Economics, 37, 67–87. de Meza, D., & van der Ploeg, F. (1987). Production ﬂexibility as a motive for multinationality. Journal of Industrial Economics, 35, 343–351. Douglas, S. P., & Craig, C. S. (1989). Evolution of global marketing strategy: Scale, scope, and synergy. Columbia Journal of World Business, Fall, 47–59.

122

REID W. CLICK

Dufey, G., & Srinivasulu, S. L. (1983). The case for corporate management of foreign exchange risk. Financial Management, Winter, 54–62. Dunning, J. H. (1988). Explaining international production. Boston, MA: Unwin Hyman. Fama, E. F., & French, K. R. (1992). The cross-section of expected stock returns. Journal of Finance, 47, 427–465. Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33, 3–56. Fama, E. F., & French, K. R. (1995). Size and book-to-market factors in earnings and returns. Journal of Finance, 50, 131–155. Fatemi, A. (1984). Shareholder beneﬁts from corporate international diversiﬁcation. Journal of Finance, 34, 1325–1344. Ferdows, K. (1997). Making the most of foreign factories. Harvard Business Review, March– April, 73–88. Flaherty, T. (1986). Coordinating international manufacturing and technology. In: M. Porter (Ed.), Competition in global industries (pp. 83–93). Harward Business School Press. Flaherty, T. (1996). Global operations management. New York: McGraw-Hill. Ghoshal, S. (1987). Global strategy: An organizing framework. Strategic Management Journal, 8, 425–440. Ghoshal, S., & Bartlett, C. A. (1990). The multinational corporation as an interorganizational network. Academy of Management Review, 15, 603–625. Hamel, G., & Prahalad, C. K. (1985). Do you really have a global strategy? Harvard Business Review, July–August, 139–148. Harrigan, K. R. (1992). A world-class company is one whose customers cannot be won away by competitors: Internationalizing strategic management. In: A. M. Rugman & W. T. Stanbury (Eds), Global perspective: Internationalizing management education (pp. 251–263). British Columbia, Canada: Centre for International Business Studies, University of British Columbia. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversiﬁcation: Effects on innovation and ﬁrm performance in product-diversiﬁed ﬁrms. Academy of Management Journal, 40, 767–798. Hofstede, G. (1984). Culture’s consequences: International differences in work-related values. Beverly Hills, CA: Sage. Hout, T., Porter, M. E., & Rudden, E. (1982). How global companies win out. Harvard Business Review, September–October, 98–108. Karrenbrock, J. (1990). The internationalization of the beer brewing industry. Federal Reserve Bank of St. Louis Review, November–December, 3–19. Kashani, K., & Quelch, J. A. (1990). Can sales promotion go global? Business Horizons, May– June, 37–43. Knetter, M. M. (1994). Exchange rates and corporate pricing strategies. In: Y. Amihud & R. M. Levich (Eds), Exchange rates and corporate performance (pp. 181–219). Burr Ridge, IL: Irwin Professional. Kogut, B. (1985). Designing global strategies: Proﬁting from operational ﬂexibility. Sloan Management Review, Fall, 27–38. Kogut, B., & Kulatilaka, N. (1994). Operating ﬂexibility, global manufacturing, and the option value of a multinational network. Management Science, 40, 123–139. Lessard, D. R., & Lightstone, J. B. (1986). Volatile exchange rates can put operations at risk. Harvard Business Review, July–August, 107–114.

Value Creation Perspective of International Strategic Management

123

Levitt, T. (1983). The globalization of markets. Harvard Business Review, May–June, 92–102. Mikhail, A., & Shawkey, H. (1979). Investment performance of U.S.-based multinational corporations. Journal of International Business Studies, 10, 53–66. Morris, W. (Ed.) (1975). The American heritage dictionary of the English language. Boston: Houghton Mifﬂin. Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press. Porter, M. E. (1986a). Changing patterns of international competition. California Management Review, 28, 9–40. Porter, M. E. (1986b). Competition in global industries. Boston, MA: Harvard Business School Press. Porter, M. E. (1990). The competitive advantage of nations. New York: Free Press. Prahalad, C. K., & Doz, Y. L. (1987). The multinational mission: Balancing local demands and global vision. New York: Free Press. Quelch, J. A., & Bloom, H. (1999). Ten steps to a global human resources strategy. Strategy & Business, First Quarter, 18–29. Quelch, J. A., & Hoff, E. J. (1986). Customizing global marketing. Harvard Business Review, May–June, 59–68. Quian, G., & Li, J. (1998). Multinationality, global market diversiﬁcation, and risk performance for the largest U.S. ﬁrms. Journal of International Management, 4, 149–170. Ricks, D. A. (1993). Blunders in international business. Cambridge, MA: Blackwell. Shapiro, A. C. (1983). International capital budgeting. Midland Corporate Finance Journal, Spring, 26–45. Standard and Poor’s Corporation. (1997). Compustat PC-Plus. Stobaugh, R. B., Jr. (1969a). Where in the world should we put that plant? Harvard Business Review, January–February, 129–136. Stobaugh, R. B., Jr. (1969b). How to analyze foreign investment climates. Harvard Business Review, September–October, 100–108. Theil, H. (1967). Economics and information theory. Chicago: Rand McNally. U.S. Department of Commerce. (2005). National income and product accounts of the United States. Survey of Current Business, August, pp. 36–172.

This page intentionally left blank

OWNERSHIP STRUCTURE, DIVERSIFICATION STRATEGY, AND PERFORMANCE: IMPLICATIONS FOR ASIAN EMERGING MARKET MULTINATIONAL ENTERPRISES Juichuan Chang ABSTRACT The purpose of this study is to provide an integrated framework that conceptualizes multifaceted antecedents pertaining to international expansion of emerging market businesses in relation to firm performance. This paper develops multiple-item measures of multiple dimensions to clarify ownership structure and three diversification strategy relationships to performance. We test how ownership structure and diversification strategy affect emerging market multinational enterprises’ financial performance. The result shows that the relationship between ownership structure and firm performance is a nonlinear relationship (S shape). We also found that excessive international diversification, product diversification, and geographic scope of the expansion process negatively moderate the impact of Asia Pacific multinational enterprises’ performance. Value Creation in Multinational Enterprise International Finance Review, Volume 7, 125–148 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07006-3

125

126

JUICHUAN CHANG

1. INTRODUCTION International expansion is accompanied by both opportunities and threats. However, empirical research and theory development have traditionally focused on US and European multinational company data, which may reﬂect features of the corporate behavior of multinational enterprises (MNEs) in developed countries. However, the rise of emerging market multinational ﬁrms is a more recent phenomenon and has attracted limited empirical research attention (Pananond & Zeithaml, 1998). It seems that existing corporate internationalization literature lacks an integrated framework that conceptualizes multifaceted antecedents pertaining to international expansion of emerging market businesses in relation to ﬁrm performance. Therefore, the primary purpose of this study is to ﬁll this gap in the context of multinational ﬁrms. Asian emerging market MNEs are expected to internationalize and globalize at a faster pace in the future. They enhance their competitive cost advantages and tap foreign markets. Some of Asia MNEs are moving outside their Asian bases to North America and Europe to strategically position themselves for new markets and technologies in the 21st century. Some studies suggest that there are good reasons to believe that ﬁrms from lessdeveloped countries differ in their internationalization from ﬁrms in developed countries. Therefore, we build a theoretical argument and examine the effects of ownership structure, three diversiﬁcation strategies – international diversiﬁcation strategy, geographic scope, and production diversiﬁcation strategy – of MNEs’ international expansion process in a sample of Asian emerging market MNEs over the period 1998–2002. In particular, our hypotheses were tested on panel data on 115 Asia Paciﬁc MNEs (Taiwan, Singapore, Hong Kong, and South Korea) that expanded abroad. The concept model we test is illustrated in Fig. 1.

2. THEORETICAL BACKGROUND AND HYPOTHESES DEVELOPMENT 2.1. Ownership Structure and Performance Numerous empirical studies have tried to highlight evidence of the relationship between ownership structure and corporate performance. According to agency theory, several categories of shareholders can have an inﬂuence on the managers’ efﬁciency: the managerial shareholders, the ﬁnancial shareholders, and the institutional shareholders. In this paper, we only focus on

Ownership Structure, Diversification Strategy, and Performance

127

Outside Block Ownership Ownership Structure Managerial Ownership

Diversification Strategy

H1

International Diversification

H2

Product Diversification

H3

Interaction of International and Product Diversification

Performance

H7

H4

Geographic Diversification

H5 H6

Intra-market Extra-market

Fig. 1. Conceptual Model.

managerial ownership. McConnell and Servaes (1990) deﬁned management ownership as equity owned by corporate ofﬁcers and members of the board of directors. Now three main hypotheses exist to explain the relationship between ownership structure and performance: convergence in interest hypothesis, neutrality hypothesis, and entrenchment hypothesis. The ﬁrst is Jensen and Meckling’s (1976) ‘‘convergence in interest hypothesis.’’ They show that managerial ownership increases a ﬁrm’s value by reducing agency costs. When managers own a large proportion of the ﬁrm’s shares, they beneﬁt to a larger extent of the beneﬁts of their effort. The second is Demsetz’s (1983) hypothesis, which shows that corporate performance depends on environmental constraints. This hypothesis is known as the ‘‘neutrality hypothesis.’’ The third is Shleifer and Vishny’s (1989) hypothesis, which suggests that the greater the percentage of shares held by the manager, the less the other shareholders can compel him to manage the ﬁrm in their interests. This hypothesis is known as the ‘‘entrenchment hypothesis.’’

128

JUICHUAN CHANG

The nonlinear relation between a ﬁrm’s ownership and performance is conﬁrmed by different empirical studies. For example, Morck et al. (1988) indicate that corporate value rises ﬁrst with increases of internal ownership below 5 percent, decreases between 5 and 25 percent, and ﬁnally increases slightly when internal ownership exceeds 25 percent. McConnell and Servaes (1990) show a negative effect of internal ownership between 5 and 25 percent and a nonsigniﬁcant one for ownership values exceeding 25 percent. Shleifer and Vishny (1989) argue that the positive relationship between Tobin’s Q and managerial ownership is sustained at higher levels of ownership for small ﬁrms than it is for large ﬁrms. Empirical research and theory development have traditionally focused on US and European multinational company data; there may also be good reasons to believe that the relationship between ownership structure and performance is different for ﬁrms in emerging market MNEs. Therefore, we hypothesize the relationship between the ﬁnancial performance of Asian emerging market MNEs and their ownership structure as follows: Hypothesis 1. There is a nonlinear relationship (S shaped) between managerial ownership and performance in Asian emerging market MNEs. In this paper, we focus on three diversiﬁcation strategies: international diversiﬁcation, geographic diversiﬁcation, and product diversiﬁcation. 2.2. The Effect of International Diversification Strategy on Performance There may also be good reasons to believe that the internationalization– performance relationship is different for ﬁrms in emerging market MNEs. First, ﬁrms in developing countries are generally less advanced in managerial and technological knowledge in comparison with ﬁrms in developed countries (Tolentino, 1993). Second, they may face different macroeconomic conditions and institutional environments compared with ﬁrms in developed countries (Khanna & Rivkin, 2001). Third, developed countries and developing countries differ in international competitiveness by sector (Nachum, 2004). Some studies have found a positive relationship between internationalization and ﬁrm performance (Geringer et al., 1989; Rugman et al., 2000). Some studies have found a negative relationship (Kumar, 1984; Michel & Shaked, 1986). However, recent studies suggest that the form of the relationship is an inverted U-relationship (Geringer et al., 1989; Hitt, Hoskisson, & Kim, 1997). More recently, Ruigrok and Wagner (2003) have found a standard U-shape relationship between internationalization and ﬁrm performance.

Ownership Structure, Diversification Strategy, and Performance

129

In stage 1, the relationship is a negative slope as costs and barriers to initial international expansion outweigh the beneﬁts. Initially, the emerging market MNEs typically have large learning costs because of unfamiliarity with foreign markets, cultures, and environments. Second, the emerging market early internationalizers often suffer from less efﬁcient production and less internationally competitive knowledge-base industries; they also often suffer from weak or non-existing institutions (Khanna & Rivkin, 2001) as well as underdeveloped capital markets (Singh, 1995). Apart from economic and legal barriers, barriers to emerging market MNEs internationalization include cultural distance (Johanson & Vahlne, 1977) and establishing the ﬁrm’s legitimacy abroad (Zaheer & Mosakowski, 1997). For early emerging market internationalizers, the small scale of global operations is insufﬁcient to recoup costs of creating an internationalization (Gongming, 1998; Hitt et al., 1997). In stage 2, the relationship becomes positive as beneﬁts of emerging market MNEs’ international expansion are now realized. Further geographical expansion makes possible efﬁciencies that improve resource utilization (Kim, Hwang, & Burgers, 1989; Porter, 1985). While market-seeking ﬁrms are better able to scan for market opportunities, the incremental beneﬁts of further international expansion are now greater than the incremental costs of a further stage 2 expansion. Other beneﬁts of stage 2 international expansion are the ability of some companies to exercise global market power (Grant, 1987) and to extend the product cycle (Vernon, 1966). Vernon argues that competitive advantages of developing country MNEs do not lie in frontier technology. Rather, they can take the form of adaptation of imported technology, development of products suitable for developing countries, or innovations of small-scale production techniques. For instance, while the manufacturing and commodity sectors have seen a decline in many industrialized countries, the manufacturing and commodity sectors have been growing in many developing countries. Young, Wansley, and Lane (1999) indicate that some unique features of Asia Paciﬁc emerging market MNEs’ internationalization strategies can be highlighted in stage 2: (1) sectoral specialization through vertical integration, (2) diversiﬁcation into unrelated businesses, and (3) family ownership and management. Therefore, we hypothesize the relationship between performance of a multinational ﬁrm and its internationalization as follows: Hypothesis 2. Moderate levels of internationalization are positively related with ﬁrm performance, but a high-level internationalization should be negatively related with the ﬁrm’s performance.

130

JUICHUAN CHANG

2.3. Product Diversification Strategy and Performance The relationship between product diversiﬁcation and ﬁrm performance has long been studied by management scholars although the results are puzzling, given that most are theoretical arguments. Salter and Weinhold (1981) indicate that a diversiﬁed ﬁrm can apply particularly skills and knowledge acquired in one business to solve problems and exploit opportunities in other business. Wrigley (1974) distinguished two modes of product diversiﬁcation: related and unrelated diversiﬁcation. Then, Hitt and Ireland (1986) found that related diversiﬁcation facilitates the sharing of activities and the transfer of skills across businesses to increase ﬁrm value; therefore, related diversiﬁcation has potential sources of value increases at both the corporate and business levels. Conversely, unrelated diversiﬁcation provides few operational synergies and, therefore, must rely on ﬁnancial synergies for increasing value. Wrigley (1974) suggested that related diversiﬁed ﬁrms emphasize strategic control to achieve superior performance and that unrelated diversiﬁed ﬁrms emphasize ﬁnancial controls to achieve superior performance. However, at some point these efforts are also associated with major costs. For example, Grant, Jammine, and Thomas (1988) highlight the growing strain on top management as it tries to manage an increasingly disparate and less familiar portfolio of businesses. Markides (1992) identiﬁes other costs, such as control and effort losses (due to increased shirking), coordination costs, and other diseconomies related to organization, inefﬁciencies from conﬂicting ‘‘dominant logics’’ between businesses, and internal capital market inefﬁciencies. However, some researchers argue for a curvilinear relationship between product diversiﬁcation and performance with net beneﬁts increasing to a maximum and decreasing as costs increase. Therefore, we hypothesize that the emerging market MNEs’ moderate levels of product diversiﬁcation have a beneﬁcial inﬂuence on their performance when they did not engage heavily in product diversiﬁcation, but when emerging market MNEs go far beyond their original industries, then their performance turned out to be negative with further high levels of product diversiﬁcation. The effect of diversiﬁcation on business group performance is as follows: Hypothesis 3. Moderate levels of product diversiﬁcation are positively related with emerging market MNEs’ performance, but a higher product diversiﬁcation is negatively related with the emerging market MNEs’ performance.

Ownership Structure, Diversification Strategy, and Performance

131

2.4. Geographic Diversification of Emerging Market MNEs’ Operation In this study, we examine geographic scope effects on the relationship between internationalization and Asian emerging market MNEs’ performance. There are good reasons to believe that the internationalization– performance shape of the curve may vary with the country of origin of the ﬁrm and the markets it enters. The Uppsala model of internationalization (Johanson & Vahlne, 1977) proposed that ﬁrms internationalize incrementally from ‘‘psychically close’’ countries to ‘‘psychically distant’’ countries. This would predict a pattern of internationalization in which one would ﬁnd internationalization in familiar countries in the ﬁrst stage and internationalization in less familiar countries in the latter stages. By deﬁnition, intraregional markets are located at a closer distance. When emerging market businesses ﬁrst internationalize, they usually enter countries whose culture and business customs most resemble their home (OECD, 1997). If they are in the same time zone, it is also easier to coordinate activities and reduce transportation costs, the costs of coordination, as compared to extraregionally dispersed operations. By concentrating on markets in the same region, ﬁrms may also avoid diseconomies of time compression (Vermeulen & Barkema, 2002). Finally, markets in the same region are also likely to be part of the same trade blocs and so beneﬁt from reduced market entry barriers. Therefore, a regional strategy may allow emerging market MNEs to achieve a better balance between the two demands, gaining some of the beneﬁts of globalization while remaining responsive to local market needs (Ghemawat, 2003; Ghoshal, 1987). However, Ghoshal (1987) suggests that as ﬁrms concentrate on extra-regional international markets, more exposure to foreign exchange ﬂuctuations systematically increases the foreign exchange risk, and then the agency problem, communication, coordination, and motivation problems stemming from cultural diversity in the MNEs. Therefore, the relationship between internationalization and ﬁrm’s performance is a crucial question for researchers. We assume that these effects seem to be conﬁrmed by the preponderance of regionalization strategies among Asia emerging market MNEs. The logic described above suggests the following hypotheses: Hypothesis 4. A higher geographic diversiﬁcation negatively moderates the relationship between international diversiﬁcation and the ﬁrm’s performance. Hypothesis 5. The internationalization of intra-regional operations will be positively related to the ﬁrm’s performance.

132

JUICHUAN CHANG

Hypothesis 6. The internationalization of extra-regional operations will be negatively related to the ﬁrm’s performance. 2.5. The Interaction Effects of International Diversification and Product Diversification on a Firm’s Performance It is now widely recognized that internationalization and diversiﬁcation effects are related. A small number of existing studies have examined the combined effects of product diversiﬁcation and internationalization. Geringer et al. (1989) found that the effect of the interaction of product diversiﬁcation and internationalization on a ﬁrm’s performance has no signiﬁcant effects. Franko (1989) indicates that a high level of product diversiﬁcation in geographically diverse internationalization leads to lower performance. Kim et al. (1989) suggested no effect of global diversiﬁcation on related diversiﬁed ﬁrm performance. Hitt et al. (1997) found that high levels of product diversiﬁcation will raise the cost of governing ﬁrms; therefore, excessively high degrees of product and international diversiﬁcation together should depress performance. The interactive and direct inﬂuence of product diversiﬁcation on international operations has also been elaborated in many previous studies (Geringer et al., 1989; Grant et al., 1988; Hitt et al., 1997; Kim et al., 1989). Overall, these studies indicate that signiﬁcant interactive effects exist between international diversiﬁcation and product diversiﬁcation. We argue that moderate levels of product diversiﬁcation and internationalization are positively related with ﬁrm performance, but a high level of diversiﬁcation and internationalization should be negatively related with a ﬁrm’s performance. Therefore, we hypothesize the interaction effects of internationalization and diversiﬁcation on business groups as follows: Hypothesis 7. Moderate interaction of international diversiﬁcation and product diversiﬁcation are positively related with emerging market MNEs’ performance, but a higher interaction of the expansion process is negatively related with the emerging market MNEs’ performance.

3. METHODOLOGY 3.1. Sample and Data collection First, we identiﬁed the Asian emerging market multinational companies from The Forbes 2000, a comprehensive ranking of the world’s top 2000

Ownership Structure, Diversification Strategy, and Performance

133

largest companies over the period 1998–2002, measured by a composite of sales, proﬁts, assets, and market value. Second, data on ownership structure, the international operations (foreign sales/total sales and foreign asset/total asset), and return on sales. Returns on asset were obtained from stock exchanges of Taiwan, Hong Kong, South Korea, and Singapore, company annual reports, and supplementary materials comprising group founders’ biographies, governmental statistics, corporation disclosure statements, and business newspaper and journal reports. Firms used in this study had to have a least 10 percent of their sales originating overseas, an implicit criterion used in many of the previous studies (see, e.g., Daniels & Bracker, 1989; Geringer, Tallman, & Olsen, 2000). Companies that did not distinguish between export and foreign subsidiary sales and those that did not provide the ﬁgures necessary for calculating the ratio were removed from the sample. Table 1 shows the distribution by four countries and across seven industries. The ﬁnal sample amounted to 115 multinational ﬁrms in Asia Paciﬁc MNEs (Hong Kong, South Korea, Taiwan, Singapore) over the period 1998–2002. The breakdown by country and industry sector is shown in Tables 1a and 1b, respectively. Table 1a.

Sample Distribution of Asia Paciﬁc MNEs across Countries.

Asia Paciﬁc MNEs

Total Samples

Hong Kong Taiwan South Korea Singapore

43 29 33 10

Total

Table 1b.

115

Sample Distribution of Asia Paciﬁc MNEs across Industries.

No. 1 2 3 4 5 6 7

Industry

Asia Paciﬁc MNEs

Conglomerates Consumer goods Food Primary manufacturing Secondary manufacturing Service Utility

7 18 10 15 35 26 4

Total

115

134

JUICHUAN CHANG

3.2. Variables and Measures 3.2.1. Dependent Variable Performance: We measured ﬁrm performance by two methods: (1) Return on assets (ROA): income before extraordinary item, divided by total assets. (2) Return on sales (ROS): income before extraordinary item, divided by total sales revenue 3.2.2. Independent Variables Managerial ownership: We deﬁne managerial ownership as the percentage of share held by the ﬁrm’s board of directors. Degree of internationalization (DOI) was measured using the sum of foreign sales as a percentage of total sales, FSTS (e.g., Grant, 1987) and foreign assets as a percentage of total assets, FATA (Gomes & Ramaswamy, 1999). DOI (degree of internationalization) ¼ FSTS+FATA Intra-regional: Asia Paciﬁc region Extra-regional: outside Asia Paciﬁc region Global region: intra-region+extra-region Product diversification index: We use a Herﬁndahl-type quantitative index as a measure of ﬁrm diversiﬁcation (Grant et al., 1988). The index satisﬁes the following important requirements: it varies directly with the number of different products produced; it varies inversely with the increasingly unequal distribution of products across product lines; and it is bounded between zero and unity. The product diversiﬁcation index takes into account the number of segments in which a ﬁrm operates and the relative importance of each segment in sales. X Product Diversification Index ¼ 1 ðsiÞ2 where si is the proportion of a ﬁrm’s sales reported in product group i. Geographic diversification: the number of countries in which the ﬁrm operates. 3.2.3. Control Variables Size: This was measured by the natural logarithm of number of employees (Gomes & Ramaswamy, 1999) and was used to control for the potential

Ownership Structure, Diversification Strategy, and Performance

135

effect of scale economy differences. Logarithmic transformation makes the results easier to interpret, because the changes in the logarithm domain represents the relative (percentage) changes in the original metric and also makes the distribution of data closer to normality. Debt: To control for changes in the ﬁrms’ capital structure, we include a leverage variable, measured by total debt over total assets. Table 2 indicates the operationalization of variables, the sources of information, and its previous usage. 3.3. The Empirical Model We use a cross-sectional heteroskedastic, timewise autoregressive model, which makes it possible to reveal individual exploratory variables as well as the dynamics over time. The multiple regression analysis was carried out on ﬁve-year (1998–2002) averaged data for internationalization, performance, and other control variables. To test Hypothesis 1, we used a cubic regression model with ﬁrst-, second, and third-order terms for degree of ownership as follows: Performance ¼ a þ b1 ðmanagerial ownershipit Þ þ b2 ðmanagerial ownershipit Þ2 þ b3 ðmanagerial ownershipit Þ3 þ S b3þm I m þ b3þmaxðmÞþ1 C 1 þ b4 ðoutside block ownershipÞ þ b5 ðSizeÞ þ b6 ðDebt ratioÞ þ it MARS model. In the second stage, we used the Multivariate Adaptive Regression Spline (MARS) methodology developed by Friedman (1991) to determine the turning points in the ownership–performance relationship. The piecewise linear regression model allowing for two changes in the slope coefﬁcient of the degree of internationalization is as follows: Degree of managerial ownership: DMO Performance ¼ a þ b1 ðDMO1i Þ þ b2 ðDMO2i Þ þ b3 ðDMO3i Þ þ b4 ðSizeÞ þ i

136

Table 2. Deﬁnition of Variables and Data Sources. Variable Dependent variables Performance

Managerial ownership Outside block Ownership Im m I t

Deﬁnition

Adapted From

(1) Return on assets (ROA): income before extraordinary item, divided by total assets (2) Return on sales (ROS): income before extraordinary item, divided by total sales revenue The percentage of shares held by the ﬁrm’s board of directors The percentage of shares held by institutions and other external ownership interests Set of dummy variables to control sub-sector effects Knowledge and capital intensive service sector Represents the companies in the study (cross-sectional component) Corresponds to the different years (timeseries component)

Dunning (1995), Grant (1987), Daniels and Bracker (1989)

McConnell and Servaes (1990) McConnell and Servaes (1990) Contractor et al. (2002) Contractor et al. (2002) Contractor et al. (2002) Contractor et al. (2002)

JUICHUAN CHANG

Independent variables International diversiﬁcation ¼ FSTS+FATA ¼ DOI (degree of internationalization), Sullivan (1994) FSTS Foreign sales as a percentage of total sales Rugman (1986), Haar (1990) FATA Foreign assets as a percentage of total assets Daniels and Bracker (1989), Sambharya (1996) (International diversiﬁcation)2 Quadratic of international diversiﬁcation Contractor et al. (2002) P 2 Product diversiﬁcation Product diversiﬁcation index ¼ 1 (si) si is the proportion of a ﬁrm’s sales in Grant et al. (1988) product group i

Size Debt ratio

The number of countries in which the ﬁrm operates Natural log of number of employees Total debt over total assets

Grazialetto-Gilies (1998)

Industries dummies I1 I2 I3 I4 I5 I6

Using the Utility sector as the baseline variable, to reach lowest Cronbach’s alphas 1 ¼ Conglomerates 1 ¼ Consumer goods 1 ¼ Food 1 ¼ Primary manufacturing 1 ¼ Secondary manufacturing 1 ¼ Service

Haveman (1993) Hitt et al. (1997)

Ownership Structure, Diversification Strategy, and Performance

Geographic diversiﬁcation

137

138

If If If If If If If

JUICHUAN CHANG

DMOiolevel one then DMO1i ¼ DMOi DMOi>level one then DMO1i ¼ level one DMOiolevel one then DMO2i ¼ 0 level onerDMOiolevel two then DMO2i ¼ DMOilevel one DMOiZlevel two then DMO2i ¼ level twolevel one DMOiolevel two then DMO3i ¼ 0 DMOiZlevel two then DMO3i ¼ DMOlevel two

4. RESULTS 4.1. The Relationship between Ownership Structure and Firm Performance: A Nonlinear Relationship (S Shape) Table 3 and Fig. 2 conﬁrm Hypothesis 1. The results show that the relationship between a ﬁrm’s performance and managerial ownership is a nonlinear relationship (S shape). The results of Models 1 (see Table 3) are strongly signiﬁcant, with high F-values. The results (Fig. 2 and Model 1) indicate that corporate performance (ROA) rises ﬁrst with increases of managerial ownership below 8 percent (beta ¼ 0.163, po0.01), decreases between 8 and 23 percent (beta ¼ 1.018, po0.05), and ﬁnally increases when managerial ownership exceeds 23 percent (beta ¼ 2.213, po0.001). Hence, the results are in accordance with those obtained by Morck et al. (1988). In the second step, we use a cross-sectional heteroskedastic, timewise autoregressive model and the MARS methodology to determine the turning points in the ownership–performance relationship (Fig. 2). Initially, as the managerial ownership increase, the corporate performance improves. But we also ﬁnd that the square of the level of managerial ownership is negatively related to a ﬁrm’s performance (8 percentomanagerial ownershipo23 percent). As managerial ownership exceeds 23 percent, the managerial ownership is likely to beneﬁt corporate performance because mangers who own enough stock to dominate the board of directors could get entrenched. Therefore, managerial ownership increases a ﬁrm’s value by reducing agency costs. As the greater the percentage of shares held by managers, the lesser the other shareholders can compel him to manage the ﬁrm in their interest. The results of Table 3 also show that a ﬁrm’s performance is positively related to outside block ownership (beta ¼ 0.142, po0.05), indicating that outsider owners are active in monitoring management. The results also show that a ﬁrm’s performance is positively related to the ﬁrm’s size and negatively related to debt ratio.

Ownership Structure, Diversification Strategy, and Performance

Table 3.

139

Regression of Asia Paciﬁc MNEs’ Performance (Samples N ¼ 115) (Cubic Regression Model).

ROA (Dependent variable) (Test Hypothesis 1) Intercept Managerial ownership (Managerial ownership)2 (Managerial ownership)3 Outside block ownership I1 I2 I3 I4 I5 I6 Firm size Debt ratio R2 F

Model 1 1.891 0.163 1.018 2.213 0.142 0.028 0.414 0.136 0.125 0.053 0.079 0.542 0.173 5.394 7.027

Notes: I1: sector effect (dummy, 1 ¼ Conglomerates; 0 ¼ otherwise). I2: sector effect (dummy, 1 ¼ Consumer goods; 0 ¼ otherwise). I3: sector effect (dummy, 1 ¼ Food; 0 ¼ otherwise). I4: sector effect (dummy, 1 ¼ Primary manufacturing; 0 ¼ otherwise). I5: sector effect (dummy, 1 ¼ Secondary manufacturing; 0 ¼ otherwise). I6: sector effect (dummy, 1 ¼ Service; 0 ¼ otherwise). Using the Utility sector as the baseline variable to reach lowest Cronbach’s alphas. po0.05; po0.01; po0.001.

4.2. The Interaction Effects of International Diversification and Product Diversification on a Firm’s Performance In Table 4, Model 3 revealed a clear quadratic relationship between international diversiﬁcation and performance. The results of Model 3 are strongly signiﬁcant, with high F-values. International diversiﬁcation and performance were signiﬁcantly positively related up to a point in the 0–64 percent range (beta ¼ 1.367, po0.001), after which an increase in international diversiﬁcation was associated with declining performance (beta ¼ 1.068, po0.01). Therefore, the results support Hypothesis 2. In the case of product diversiﬁcation, the quadratic term was also similar.

140

JUICHUAN CHANG Performance (ROA)

% 8% 23% Managerial Ownership (The percentage of shares held by the firm’s Board of Directors)

Fig. 2.

Ownership Structure and Performance (MARS Regression Model).

Table 4.

Regression of Asia Paciﬁc MNEs’ Performance (Samples N ¼ 115).

ROS (Dependent variable) (Test Hypotheses Model 2 (Linear Form) 2, 3, 4, and 7) Intercept Internationalization diversiﬁcation (Internationalization diversiﬁcation)2 Product diversiﬁcation (Product diversiﬁcation)2 Geographic diversiﬁcation (Internationalization diversiﬁcation) (Product diversiﬁcation) (Internationalization diversiﬁcation)2 (Product diversiﬁcation)2 Firm size Debt ratio R2 F

1.147 2.692 1.385 2.361

Model 3 (Quadratic Form) 2.636 1.367 1.068 2.183 2.273 2.879 1.363 3.485

0.681 0.219 0.553 6.231

0.724 0.195 0.683 7.248

po0.05; po0.01; po0.001.

The performance is positively related to product diversiﬁcation in the 0–48 percent range (beta ¼ 2.183, po0.01) and negatively related up to 48 percent (beta ¼ 2.273, po0.05). Therefore, the results support Hypothesis 3. This suggests that the scope of foreign operations may exert great inﬂuences on a ﬁrm’s performance.

Ownership Structure, Diversification Strategy, and Performance

141

In Table 4, Model 3 also shows that the interactive effect of international diversiﬁcation and product diversiﬁcation (Internationalization diversiﬁcation Product diversiﬁcation) was positively related to performance (beta ¼ 1.363, po0.01), when MNEs did not engage heavily in foreign activities. MNEs’ performance turned out to be negative (beta ¼ 3.485, po0.001) with further high levels of international diversiﬁcation and product diversiﬁcation: (Internationalization diversification)2 (Product diversification)2. This suggest that MNEs’ performance will be positive at lower and moderate levels of interaction of international diversiﬁcation and product diversiﬁcation, but may be negatively affected with further increase in the degree of international diversiﬁcation and product diversiﬁcation. This result supports Hypothesis 7. Model 3 also shows that the geographic diversiﬁcation is negatively related to a ﬁrm’s performance (beta ¼ 2.879, po0.01). This result supports Hypothesis 4. 4.3. Geographic Diversification of Emerging Market MNEs’ Operation In Table 5, we examine geographic scope effects on the relationship between internationalization and Asian emerging market MNEs’ performance, and Regression Models 4 and 5 were used to test Hypotheses 5 and 6. There are good reasons to believe that the internationalization–performance shape of the curve may vary with the country of origin of the ﬁrm andP the markets it enters. Model 4 (Performanceintra ¼ a+b1 (DOIit)intra+ b1+m Im+b2 (Firm size)+b3 (Debt ratio)+eit) tests the intra-region international expansion path of Asian emerging P market MNEs and Model 5 (Performanceextra ¼ a+b1 (DOIit)extra+ b1+m Im+b2 (Firm size)+b3 (Debt ratio)+eit) tests extra-region (outside Asia paciﬁc region). Model 6 (PerP formanceglobal ¼ a+b1 (DOIit)intra l+b2 (DOIit)extra+ (b1+m Im+ b2+m Im)+b2 (Firm size)+b3 (Debt ratio)+eit) tests the global international expansion path. The results of Table 5 are strongly signiﬁcant, with high F-values. The results of Model 4 strongly support Hypothesis 5. The internationalization of intra-regional operations is positively related to a ﬁrm’s performance (beta ¼ 0.583, po0.001). On the other hand, the results of Model 5 indicate that the internationalization of extra-regional operations is negatively related to its performance (beta ¼ 0.114, po0.05), the results supporting Hypothesis 6. Why should this be the case? We propose that intraregional international expansion processes reap the positive beneﬁt of internationalization easier, gain the beneﬁts of the multinational network more effectively, easily coordinate activities, beneﬁt from diseconomies of time

142

JUICHUAN CHANG

Table 5. Regression of Performance (Samples: 115 Asia Paciﬁc MNEs). ROS (Dependent Variable) (Test Hypotheses 5 and 6) Intercept DOI (intra-regional markets) DOI (extra-regional markets) I1 I2 I3 I4 I5 Firm size Debt ratio R2 F

Model 4, Intraregion 0.615 0.583 0.317 0.145 0.059 0.113 0.192 0.075 0.248 0.413 9.726

Model 5, Extraregion

Model 6, Global Region

1.891

2.162 0.416 0.106 0.377 0.151 0.001 0.173 0.159 0.182 0.204 0.386 8.326

0.114 0.414 0.136 0.125 0.053 0.179 0.063 0.183 0.372 7.219

Notes: DOI (extra-regional markets): degree of internationalization within intra-region. DOI (extra-regional markets): degree of internationalization within extra-region. DOI (global regional markets): degree of internationalization within global region (global region ¼ intra-region+extra-region). I1: sector effect (dummy, 1 ¼ Electronic/electrical equipment; 0 ¼ otherwise). I2: sector effect (dummy, 1 ¼ Food; 0 ¼ otherwise). I3: sector effect (dummy, 1 ¼ Chemical; 0 ¼ otherwise). I4: sector effect (dummy, 1 ¼ Industrial equipment; 0 ¼ otherwise). I5: sector effect (dummy, 1 ¼ Software and services; 0 ¼ otherwise). Using the transportation equipment as the baseline variable to reach lowest Cronbach’s alphas. po0.05; po0.01; po0.001.

compression, reduce transportation costs and market entry barriers, and allow ﬁrms to optimize on the cost and changes in prices of goods, interest rates, labor, and raw materials by shifting operations across nearer nations (Pantzalis, 2001). But extra-regional international expansion processes require a larger global scale before the net beneﬁts of expansion are realized.

5. DISCUSSION This paper contributes to open debate about the link between ownership structure and ﬁrm performance. Previous studies considering a nonlinear relation between managerial ownership and a ﬁrm’s value report different

Ownership Structure, Diversification Strategy, and Performance

143

breaking points probably due to differences in the samples used (Kole, 1995). The results of Table 3 show a statistically signiﬁcant S shape relationship between managerial ownership and ﬁrm performance for the sample of Asian emerging market MNEs. This result suggests that initially, as mangers’ equity increases, their interests coincide more with those of outside shareholders. Initially, as the managerial ownership increases, the corporate performance improves. But we also ﬁnd that the square of the level of managerial ownership is negatively related to a ﬁrm’s performance (8 percentomanagerial ownershipo23 percent) in our samples, the reason is that corporate performance depends on environmental constraints (Demsetz, 1983). This hypothesis is known as the ‘‘neutrality hypothesis’’. As managerial ownership exceeds 23 percent, the managerial ownership is likely to beneﬁt corporate performance because mangers who own enough stock to dominate the board of directors could get entrenched. Therefore, managerial ownership increases a ﬁrm’s value by reducing agency costs. The greater the percentage of shares held by managers, the lesser the other shareholders can compel him to manage the ﬁrm in their interests. The results are in accordance with those obtained by Morck et al. (1988). Finally, we ﬁnd that a ﬁrm’s performance is positively related to outside block ownership, indicating that outsider owners are active in monitoring management. All of our empirical results of Table 4 conﬁrm Hypotheses 2, 3, 4, and 7. Multiple regression analysis was performed to examine their effects on performance for a sample of Asian emerging MNEs. The results show that international diversiﬁcation and product diversiﬁcation will exert both individual and interactive effects on performance. We observed that international diversiﬁcation had also a curvilinear relationship with performance. This result is fully consistent with Hypothesis 2. Initially (Stage 1), Asian emerging market MNEs typically have large learning costs because of unfamiliarity with foreign markets, cultures, and environments. They are also generally less advanced in managerial and technological knowledge in comparison with ﬁrms in developed countries and face different macroeconomic conditions and institutional environments when compared with ﬁrms in developed countries (Khanna & Rivkin, 2001). For early emerging market internationalizes, the small scale of global operations is insufﬁcient to recoup costs of creating an internationalization (Hitt et al., 1997). In stage 2, the relationship becomes positive as beneﬁts of emerging market MNEs’ international expansion are now realized; further geographical expansion makes possible efﬁciencies that improve resource utilization (Kim et al., 1989; Porter, 1985). Asian emerging market MNEs’ internationalization strategies can be highlighted in the second

144

JUICHUAN CHANG

stage: (1) sector specialization through vertical integration, (2) diversiﬁcation into unrelated businesses, and (3) family ownership and management. Moderate product diversiﬁcation increases the ﬁrm’s capacity to exploit different market opportunities when they engage in foreign activity. In contrast, ﬁrm performance will turn out to be negative when MNEs heavily expand their product offerings and geographic market. This is due mainly to the increase in both environmental and technological complexities associated with increase of geographic scope of foreign operations. Movement into more international markets signiﬁcantly increases managerial transaction costs and information-processing demands. As ﬁrms encompassed increasingly broader geographic markets, the costs associated with geographic dispersion began escalating sometimes quite rapidly, thus eroding proﬁt margins (Geringer et al., 1989). This suggests that foreign operations eventually become highly complex and difﬁcult to manage (Hitt et al., 1997). All of our empirical results of Table 5 conﬁrm Hypotheses 5 and 6. The internationalization of intra-regional operations by Asian emerging market MNEs is positively related to its performance. On the other hand, the internationalization of extra-regional operations is negatively related to its performance. The results show that Asian emerging market MNEs’ international expansion paths have signiﬁcant differences between intra-regional markets and extra-region markets. Why should this be the case? We found several reasons: First, when Asian emerging market MNEs operate in intraregional markets, it could be the case that MNEs may gain a higher return or margins than when they operate in extra-regional markets, because of the greater risk involved in extra-region operations. This would predict a pattern of internationalization in which one would ﬁnd MNEs’ internationalization in familiar countries in the ﬁrst stage and internationalization in less familiar countries in the second stage. The Uppsala model of internationalization (Johansson & Vahlne, 1977) proposed that ﬁrms internationalize incrementally from ‘‘psychically close’’ countries to ‘‘psychically distant’’ countries. Second, extra-regional international expansion processes require a larger global scale before the net beneﬁts of expansion are realized. Extra-regional subsidiaries in different circumstances ask for different organizational systems and processes (Gupta & Govindarajan, 1991) and companies active in a variety of cultures need to adapt their structures as well (Ghoshal, 1987). Building these systems and structures takes considerable time and attention. Third, as Asian emerging market MNEs typically have more ﬂexibility in intra-regional markets, they are able to react quickly and efﬁciently to both market and technological changes. Fourth, because extra-regional expansion is accompanied by an increase in organizational and environmental

Ownership Structure, Diversification Strategy, and Performance

145

complexity, it is therefore subject to time compression diseconomies. That is, hasty international expansion will over exhaust a ﬁrm’s absorptive capacity within a given time frame, thereby causing inadequate structural adaptation that, in turn, triggers negative performance effects. Therefore, Asian emerging market MNEs would be able to choose intra-region markets where the competitive conditions are more suitable for their resource proﬁle compared to being forced to operate in extra-region markets where they might not possess the greatest advantage. Rugman (2000) has shown that most multinational companies tend to serve primarily markets in the intraregion rather than in extra-regions, presumably because of competitive advantages in the home region.

6. CONCLUSIONS This paper develops multiple-item measures of multiple dimensions to clarify ownership structure and three diversiﬁcation strategy relationships to performance. All of our empirical results conﬁrm our hypotheses. The results of this study suggests that Asian emerging market MNEs’ ownership structure, institutional heritage, industry characteristics, and its own internationalization strategy are likely to create speciﬁc preference for operating in differing geographic regions. Although developed countries’ MNEs initially dominated the international competitive landscape, Asian emerging market MNEs have also entered the race owing to the narrowed gap in competitive advantage in international markets (Oviatt & McDougall, 1994). However, Asian emerging market MNEs have certain advantages over developed countries’ MNEs, including greater ﬂexibility, efﬁciency, quality and advantage-seeking behavior, which allow Asian MNEs to develop capabilities to succeed in the international markets. Finally, because external globalization drivers, internal organization resources, global strategy, and global performance are all multidimensional, we suggest that Asian emerging MNEs must set clear performance goals before formulating their global strategy. It is also important that businesses constantly monitor the external environmental changes and adjust their global strategy accordingly to ensure that the performance goals are being achieved. Hopefully, this study may prove useful to Asian emerging market MNEs facing expansion decisions. However, these ﬁndings still require further elaboration. Because of some inherent limitations of this study, several major areas remain unexplored or need further study to verify the conclusions and explore related areas.

146

JUICHUAN CHANG

Moreover, further enlarging the sample may offer a way to secure additional insight that would permit broader, more valid generalization and provide a better base from which to make inferences and predictions.

REFERENCES Contractor, F. J., Kundu, S. K., & Hsu, C. C. (2002). A three-stage theory of expansion of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34, 5–18. Daniels, J. D., & Bracker, J. (1989). Proﬁt performance: Do foreign operations make a difference? Management International Review, 29, 46–56. Demsetz, H. (1983). The structure of corporate ownership: Causes and consequences. Journal of Political Economy, 93, 1155–1177. Dunning, J. H. (1995). Reappraising the eclectic paradigm in an age of alliance capitalism. Journal of International Business Studies, 26, 461–492. Franko, L. G. (1989). Unrelated product diversiﬁcation and global corporate performance. International Strategic Management, 1, 221–241. Friedman, J. H. (1991). Multivariate adaptive regression splines. Annals of Statistics, 19, 1–50. Geringer, J. M., Beamish, P. W., & Costa, R. C. (1989). Diversiﬁcation strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10, 109–119. Geringer, M. J., Tallman, S., & Olsen, D. M. (2000). Product and international diversiﬁcation among Japanese multinational ﬁrms. Strategic Management Journal, 21, 51–80. Ghemawat, P. (2003). Semi-globalization and international business strategy. Journal of International Business Studies, 34, 138–152. Ghoshal, S. (1987). Global strategy: An organizing framework. Strategic Management Journal, 8, 425–440. Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between internationalization and performance. Journal of International Business Studies, 30, 173–188. Gongming, Q. (1998). Determinants of proﬁt performance for the largest US ﬁrms 1981-92. Multinational Business Review, 6, 44–51. Grant, R. M. (1987). Internationalization and performance among British manufacturing companies. Journal of International Business Studies, 18, 79–89. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 18, 79–89. Grant, R. M., Jammine, A., & Thomas, H. (1988). Diversity, diversiﬁcation, and proﬁtability among British manufacturing companies, 1972–84. Academy of Management Journal, 31, 71–80. Grazialetto, A., & Gilies, R. M. (1998). International diversification: Effects on innovation and firm performance in product diversified firms. New York: Palgrave. Gupta, A. K., & Govindarajan, V. (1991). Knowledge ﬂows and the structure of control within multinational corporations. Academy of Management Journal, 34, 138–152. Haar Jerry., Dandapani, K., & Stanley, P. (1990). The American Depository Receipt (ADR): A creative ﬁnancial tool for multinational companies. Global Finance Journal, 1, 163–171. Haveman, J. D. (1993). New dimensions in regional integration. Cambridge University Press.

Ownership Structure, Diversification Strategy, and Performance

147

Hitt, M. A., Hoskisson, R. E., & Ireland, R. D. (1986). A mid-range theory of the interactive effects of international and product diversiﬁcation on innovation and performance. Journal of Management, 20, 297–326. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversiﬁcation: Effects on innovation and ﬁrm performance in product diversiﬁed ﬁrms. Academy of Management Journal, 40, 767–798. Jensen, M. C., & Meckling, W. H. (1976). Theory of the ﬁrm; managerial behavior, agency cost and ownership structure. Journal of Financial Economics, 3, 305–360. Johanson, J., & Vahlne, J. E. (1977). The internationalization process of the ﬁrm: A model of market knowledge and increasing foreign market commitments. Journal of International Business Studies, 8, 23–32. Khanna, T., & Rivkin, J. W. (2001). Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22, 45–74. Kim, W. C., Hwang, P., & Burgers, W. P. (1989). Global diversiﬁcation strategy and corporate proﬁt performance. Strategic Management Journal, 10, 45–57. Kole, S. R. (1995). Measuring managerial equity ownership: A comparison of sources of ownership data. Journal of Corporate Financial, 1, 413–435. Kumar, M. S. (1984). Growth acquisition and investment: An analysis of the growth of industrial firms and their overseas activities. Cambridge, UK: Cambridge University Press. McConnell, J. J., & Servaes, H. (1990). Additional evidence on equity ownership and corporate value. Journal of Financial Economics, 27, 315–595. Michel, A., & Shaked, I. (1986). Domestic corporations: Financial performance and characteristics. Journal of International Business Studies, 18, 89–100. Morck, R., Andrei, S., & Robert, W. V. (1988). Management ownership and market valuation: An empirical analysis. Journal of Financial Economics, 20, 293–351. Nachum, L. (2004). Geographic and industrial diversiﬁcation of developing country ﬁrms. Journal of Management Studies, 41, 274–294. OECD (1997). Towards a new global age: Challenges and opportunities, Policy Report. OECD, Paris. Oviatt, B. M., & McDougall, P. P. (1994). Toward a theory of international new ventures. International Business Study, 25, 45–64. Pananond, P., & Zeithaml, C. P. (1998). International expansion process of MNEs from developing countries: A case study of Thailand’s CP group. Asia Pacific Journal of Management, 15, 163–184. Pantzalis, C. (2001). Does location matter? An empirical analysis of geographic scope and market valuation. Journal of International Business Studies, 32, 133–155. Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press. Rugman, A. M. (1986). Strategies for national competitiveness. Long Range Planning, 20, 92–97. Rugman, A. (2000). The end of globalization: Why global strategy is a myth & how to profit from the realities of regional markets. New York: Amacom. Rugman, A. M., & Verbeke, A. (2000). Six cases of corporate strategic responses to environmental regulation. European Management Journal, 18, 377–385. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43, 61–84. Salter, M. S., & Weinhold, W. A. (1981). Choosing compatible acquisitions. Harvard Business Review, 59, 117–127.

148

JUICHUAN CHANG

Sambhraya, R. B. (1996). The combined effect of international and product diversiﬁcation strategies on the performance of U.S.-based multinational corporations. Management International Review, 35, 197–218. Shleifer, A., & Vishny, R. W. (1989). Large shareholder and corporate control. Journal of Political Economy, 94, 461–488. Singh, A. (1995). Corporate financing patterns in industrializing economies. International Finance Corporation, Technical Paper, Vol. 2, Washington, DC. Sullivan, D. (1994). Measuring the Degree of Internationalization of a Firm. Journal of International Business Studies, 25, 325–342. Tolentino, P. E. E. (1993). Technological innovation and third world multinationals. New York: Routledge. Vermeulen, F., & Barkema, H. (2002). Pace, rhythm, and scope: Process dependence in building a proﬁtable corporation. Strategic Management Journal, 23, 637–653. Vernon, R. (1966). International investment and international trade in the product life cycle. Quarterly Journal of Economics, 80, 191–207. Wrigley, L. (1974). Divisional autonomy and diversification. Ph.D. thesis, Harvard Business School, Boston. Young, H., Wansley, J., & Lane, W. (1999). Stock market recognition of internationalization and international events. Journal of Business Finance and Accounting, 12, 263–274. Zaheer, S., & Mosakowski, E. (1997). The dynamics of the liability of foreignness: A global study of survival in ﬁnancial services. Strategic Management Journal, 18, 439–464.

SUCCESSFUL ADAPTATION STRATEGIES OF MULTINATIONAL ENTERPRISES IN CENTRAL AND EASTERN EUROPE Roxana Wright ABSTRACT The paper explores strategies of adaptation to the environment as employed by multinational corporations in Central and Eastern Europe. Organizations are treated as adaptive systems that have to match the complexity of their environments. The justification of the research lies in the complex nature of the market institutions emerging from transition that emphasizes the need for new managerial frameworks. Adaptive approaches such as vertical integration and/or value-chain development, leveraging autonomy and integration, local knowledge acquisition, and embedding in the social and political environment are explored in their relationship to success in the region.

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 149–167 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07007-5

149

150

ROXANA WRIGHT

INTRODUCTION The present study analyzes the strategic responses of multinational corporations (MNCs) to the relatively idiosyncratic environment that still persists in countries of Central and Eastern Europe (CEE). The environment in the region has gone through periods of rapid shifts through the transition from centrally planned to a market economy. The paper asserts that the MNCs that strive in the CEE economies are the ones that are actively maintaining and renewing their competitive advantages. Successful ﬁrms may ﬁnd ways to limit the impact of institutions that are set or slow changing, take opportunities to turn deﬁciencies into advantages, and actively acquire and incorporate local knowledge into their operations. This paper analyzes the common patterns of strategic adaptation to the economic environments in CEE. The study takes a dynamic approach with wide applicability to other economies in transition and beyond. The paper starts with a presentation of the theoretical framework that aids in modeling companies’ responses to the complex environment in the region. The patterns of organizational adaptation are classiﬁed into strategies of complexity reduction or complexity absorption. The formal hypotheses are derived from the theoretical framework and assert these approaches impact on company performance, such as the potential positive effect of value-chain integration, medium degree of autonomy, active knowledge acquisition, and local embedness. The selection criteria are substantiated in the subsequent section and argue for the inclusion of subsidiaries in sectors with signiﬁcant foreign investment, growth, and impact from reforms. The performance outcomes measures are presented separately as a combined quantitative and qualitative assessment of company success. The following data and methodology section describes and justiﬁes the use of logistic regression in evaluating the impact of adaptive approaches, measured as quantitative and qualitative variables, on the probability of success. The ensuing results are argued and discussed. It is mainly found that the level of integration may not affect the probability of success due to low internalization potential, a medium level of autonomy is essential for a successful local operation and endogenous learning, that subsidiaries beneﬁt from the general ability of the multinational to process knowledge, and that a strategy of local embedding is not necessarily increasing the probability of success. The closing section reﬂects on the conﬁrmed effectiveness of complexity reduction strategies as means to anticipate changes toward fully functioning market structures and institutions in CEE. The paper concludes that the patterns of successful adaptation rely on

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

151

responsiveness as well as strong capabilities and learning routines at corporate level.

THEORETICAL FRAMEWORK AND RESEARCH HYPOTHESES The analysis treats organizations as adaptive systems that have to match the complexities of their environments. The nature of transition may mean that multinational ﬁrms need new tools to function in dynamic and complex environments. The challenge for the managers is how to model complexity and analyze the implications. The current paper assists academics and managers in meeting this challenge. To handle complexity, it is asserted that two major strategic responses are available to ﬁrms operating in the environments of CEE. The ﬁrst is reduction, translated into strategies that attempt to reduce complexity by bringing it under apparent control such as vertical integration and/or value-chain development, combining existing business models with localized strategies, medium degree of autonomy and integration, and complementary specialization. The second alternative is complexity absorption, deﬁned here as active knowledge acquisition, processing and incorporation, gaining from imperfect market structures/ beneﬁting from and enhancing imperfect competition, embedding in the social and political environment, and matching local deﬁciencies with core businesses. Four of these aspects are treated in separate formal hypotheses, as presented below. The rest have been added to the empirical model for completeness and are discussed brieﬂy in the ﬁndings. The paper examines the empirical evidence of the impact of such strategies on the ﬁrm performance. Each of the hypotheses considers that the complexity reduction and absorption strategies lead to increased performance. The alternative hypotheses are also discussed, suggesting that these strategies could increase transaction costs; engage the ﬁrm in a greater level of variance that it is familiar with, which in turn may limit its ability to relate policies and practices in CEE countries to its worldwide system. The research thus explores the possibility that the adaptive approaches considered may have a negative impact on performance outcomes. The hypotheses are complemented by the investigation of potential explanations for either alternative. Hypothesis 1. Primary value-chain organization: Vertical integration has a positive impact on the success of CEE operations.

152

ROXANA WRIGHT

It is asserted that successful MNCs may eliminate inefﬁciencies in the value chain through vertical integration either by ownership or by building a network of alliances. Eliminating inefﬁciencies may also mean that MNCs need to be actively developing the capabilities of their local network of suppliers and distributors. We expect that the successful MNC adopts the role of a strategic bridge among organizations that are having difﬁculties cooperating, thus eliminating inefﬁcient practices in the value chain. The literature1 has shown that the dominant mode of growth in CEE countries has been based on networking and foreign acquisitions. It could thus be hypothesized that companies integrated vertically perform well. There are however challenges in actively building values chains and this strategy may have not been in fact successful. Companies seeking to create or expand a distribution network would encounter difﬁculties in locating and evaluating distributors – it is thus possible that some MNCs may make poor or expensive decisions. It is also possible that the distribution systems in transition economies in CEE have not achieved their potential costefﬁciency. The literature (e.g., Dyker, 2001; Pavlinek, 2004) has noted that foreign investors in some of the CEE countries rarely use local ﬁrms as ﬁrsttier suppliers in their supply networks and develop no or only few linkages with local ﬁrms because they ﬁnd it difﬁcult to acquire the supplies at the required level of sophistication or quality. Research (Radosevic, 2004) has interpreted the predominance of vertical alliances as being driven by unexploited market opportunities and cost differential rather than displacement of competition. We would then expect that as competition intensiﬁes, horizontal expansion and alliances will become more prevalent. Vertical integration may not necessarily have a positive impact on performance. Hypothesis 2. Degree of autonomy and corporate integration: Medium degree of autonomy and loose integration with the corporation lead to successful operation in CEE. It is hypothesized that successful corporations operating in CEE allow for a medium degree of autonomy of their local ventures and a loose integration with the rest of the operations. A medium level of autonomy is essential for a successful local operation: allowing subsidiaries to test strategy/new products and engage in endogenous learning. MNCs may gain from providing subsidiaries with autonomy and refraining from imposing worldwide practices and short-term efﬁciency targets. A more loosely integrated afﬁliate has access to the resources of the MNC and local managerial freedom to

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

153

generate diversity and enhance the corporation’s global capabilities. However, diversity and high degrees of adaptation generate high coordination costs and may result in a continual search within unrelated capabilities. The successful MNC needs to balance the trade-off. A time line may also be associated with the relation between performance and level of autonomy (high autonomy in the beginning creates only medium-term success). Retaining and developing context sensitivity and diversity have been shown to be valuable for the MNC (Lieb-Doczy & Meyer, 2000). From an evolutionary perspective, permitting a variety of managerial practices and organizational structures gives the company a long-term beneﬁt of more options and a greater pool of capabilities. In the idiosyncratic transition environment, loose integration with the headquarters may permit for twoway learning and thus enhance the MNC’s global capabilities. Allowing a local operation to develop capabilities and solutions in the transition context would require not only a reasonable degree of autonomy but also access to resources for developing these capabilities. Literature has suggested (e.g., Meyer, 2000) that foreign acquirers risk losing valuable local capabilities if they concentrate on the transfer of their established best practice and neglect development of variety by fostering indigenous capabilities. Rigid integration of internal routines may increase allocative efﬁciency, but MNCs may yield more long-term beneﬁts if they allow for some degree of autonomy in local managerial practice, organizational arrangements, and technology. Many local practices may initially be perceived to be inferior, yet they may be better adapted to the environment. Experimentation may be needed to develop new managerial practices that correspond to the local cultural values and resources. New practices developed locally may outperform the established ones. As a caveat, loose integration with the rest of the corporation could sabotage consistency in the MNC’s overall strategy and ultimately have a negative effect on performance. Hypothesis 3. Local capability development: Successful MNC subsidiaries pursue active knowledge acquisition and incorporation in CEE. Successful MNCs may pursue active knowledge acquisition, as well as knowledge processing and incorporation into their operations through speciﬁc learning outcomes. MNCs that followed a strategy of using their ﬁrm-level capabilities and strong market positions were successful at the beginning of transition. Active generation and development of competitive advantages may be essential in current and future successful operations in

154

ROXANA WRIGHT

CEE. Success may be related to putting together capabilities that build on activation of distinctive local knowledge or technology. Research (e.g., Radosevic, 2004) has found that companies that have acquired essential knowledge about the local environment and clients are in a better position in bargaining with foreign partners. It would thus appear that knowledge of the local market can be effectively traded for technology access. Alternatively, given the fact that local managerial know-how and marketing capabilities are somewhat limited in CEE, MNCs may beneﬁt largely by applying their ﬁrm-level capabilities. Successful operation in the transition environments of CEE may imply the use of existing business models and familiar strategy when the MNCs pursue wealthy markets and localized strategies for poor markets. In pursuing top-of-the-pyramid markets in emerging economies, MNCs may rely on proven capabilities to incrementally adapt existing business models and a familiar subsidiary strategy based on controlling resources, extracting knowledge, and leveraging economies of scale and scope. When operating in CEE, it might be more appropriate to develop separate strategies for the wealthy rising class and the poor customers. From a marketing strategy perspective, for example, the emphasis on global brands and products with little adaptation to local demand, distribution structures, and cultures may actually be detrimental in the long term. Although the introduction of global brands may initially take advantage of pent up demand, the mediumand long-run beneﬁts of such strategy are uncertain. Hypothesis 4. Embedding into the local environment: Successful MNCs follow a strategy of embedding their organizations into the local environment. Creating close relations with banks and local governments, lobbying, and pursuing reputation-building tactics (playing an active role as a good citizen in the community) are part of an embedding strategy that many MNCs are following in the CEE region. However, the local embedding may only be a support strategy of diminishing limitations while taking full advantage of opportunities and may only have an insigniﬁcant impact on performance.

CHOICE OF FIRMS OPERATING IN CEE For the purpose of the investigation a few selection criteria identify a relevant sample. The focus is on subsidiaries in technology and information

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

155

technology, consumer goods, transport, metallurgy, and infrastructure companies. These areas are among the sectors with most foreign investment in the CEE; they are rapidly growing sectors in the region and are largely affected by reforms. The selection allows for comparisons across industries and has the beneﬁt of contrasting evidence across successful and unsuccessful companies. These areas encompass not only some of the most successful companies but also among the largest loss-making ﬁrms in their respective economies (Political Risk Services, Investment Climate, 2004; Romanian Business Digest, 2004). Only investment projects with foreign equity of over $1 million are selected. UNCTAD (2005) observes a shift of FDI toward services, which can be noticed in CEE countries as well. Most countries have liberalized their services FDI regimes that have made larger inﬂows possible, especially in industries previously closed to foreign entry, such as utilities. In general, the countries of CEE outside the CIS are characterized by substantial FDI penetration in infrastructure services (e.g., banking, telecommunications, water, electricity). This trend justiﬁes the inclusion of services and IT as areas of investigation in this research. Literature suggests (Radosevic & Rozeik, 2005) that the value creation potential of CEE as a global automotive location has not yet been fully exploited. According to this recent research, ﬁnding the patterns of successful strategies of MNC subsidiaries in this sector could bring a signiﬁcant contribution. The automotive industry in the region is concentrated in Central Europe (Czech Republic, Slovakia, Hungary, Slovenia, and Poland) with vast beneﬁts as regards clustering of supplier network. Potential discovery of clustering in other countries of the CEE and cross-country patterns are possible and could further complement the discussion on the link between business environment and company strategy. Therefore, MNCs’ subsidiaries in the automotive industry are included in the present analysis.

PERFORMANCE OUTCOMES MEASURES The study uses two measures of company performance: numerical measures of profitability (proﬁts and proﬁts as percentage of sales and productivity as revenues per employee are used also in the discussion) that are easy to use in investigating possible correlations between variables, and a qualitative evaluation of performance that classiﬁes companies as successful or unsuccessful. The present research deﬁnes a successful company as an organization that

156

ROXANA WRIGHT

fulﬁlls the following conditions: It generates positive profits: As a cautionary note, although we would expect proﬁtability to be a sign of success, there are other factors that need to be taken into consideration – in some cases, losses may be temporal and due to exogenous factors, productive enterprises may show losses and transfer their proﬁts elsewhere, etc. A company may also generate positive proﬁts occasionally, while being a loss maker on average. Calculating an average proﬁt across time may have its limitations due to lack of reliable deﬂators. It has an output that is not excessively volatile (although possibly ﬂuctuating in response to market conditions, demand, etc.): We would expect that a successful ﬁrm shows stable performance not only in terms of proﬁts but also in terms of output. A highly volatile nominal output may be a sign of instability. It invests into fixed assets and has a long-term strategy at the foreign location that reflects a sustainable position in the market: A qualitative interpretation is also necessary for some companies – a ﬁrm may not invest because it has spare capacity, it accumulates money for a large project, purchases other ﬁrms’ equity, etc.

DATA AND METHODOLOGY The paper treats subsidiaries of MNCs in selected CEE countries as the unit of analysis. The research identiﬁes MNCs’ subsidiaries established in CEE between 1990 and 2004. For statistical investigation, a number of variables have been employed as quantitative and qualitative measures of the aspects considered in the hypotheses. Due to the limited nature of data availability on the variables speciﬁed in the model, a combination of a variety of sources has been used. The sources for data used in statistical analysis include the following: Thomson Gale Business and Company Resource Center,2 Mergent Online3 International Company Data, and Kompass,4 USA. Additional information was obtained from The American Chamber of Commerce in CEE – through Mr. Elias Wexler, representative in New York, USA; Arisinvest – Romanian institution promoting foreign direct investment – through Mr. Alexandru Mitroi, representative; AVAS – Romanian Institution for the Valoriﬁcation of State-owned Assets – through Mr. Viorel Dinu, representative; and the Romanian Chamber of Commerce and Industry – through Mr. Adrian Grecescu.

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

157

A logistic regression model is used for statistical generalization. The sample of subsidiaries respects the criteria described previously. Data for a total of 570 companies, subsidiaries of MNCs, have been compiled on the variables presented in the methodology description. From these companies, 542 had complete information – this sample has been used for further investigation and in comparative models. The number of companies in various countries across the region generally reﬂects the amount of FDI existent in each of these countries, although there were some limitations due to availability of data. The representation is reasonably close to expected proportions among manufacturing and services industries in the region. Companies’ representation is illustrated in Table 1. A logistic regression has been used for statistical generalization. The model used for statistical analysis is formulated as follows: pðxÞ ¼ EðyÞ ¼

eu 1 þ eu

where y ¼ success ¼ S u ¼ b0 þ b1 VI þ b2 AD þ b3 AI þ b4 SM þ b5 LC þ b6 MS þ b7 EM þ b8 LD þ b9 RE þ b10 SZ þ b11 IN Probability of success is calculated as pðSÞ ¼

eb0 þb1 VIþb2 ADþb3 AI þb4 SMþb5 LCþb6 MSþb7 EMþb8 LDþb9 REþb10 SZþb11 IN 1 þ eb0 þb1 VIþb2 ADþb3 AI þb4 SMþb5 LCþb6 MSþb7 EMþb8 LDþb9 REþb10 SZþb11 IN

where S – measure of success: 1 – yes, 0 – no (at least one of the conditions for success is not met). A discussion of the conditions considered for successful operation is presented in detail in the methodology description section. Income after tax in equivalent US dollars has been used as proﬁtability measure. The success measure also includes the volatility of output and investment in ﬁxed assets. VI – number of stages in the value chain in which the company is involved. AD – level of adaptation (1 – low to 5 – high). SM – strategic motivation: 1 – market seeking, 2 – efﬁciency seeking, 3 – knowledge seeking, 4 – hybrid motivation. AI – level of autonomy and integration (1 – low, 5 – high). LC – local capability development (1 – low, 5 – high). MS – market structure: number of local companies at the three-digit level of NAISC industry code.

158

ROXANA WRIGHT

Table 1. Companies Representation in the Sample by Industry and Country. Companies Representation by Industry

Companies Representation by Country

Industry

% of MNC Subsidiaries

Country

% of MNC Subsidiaries

29

Albania

1

19

Belarus

1

Manufacturing, high technology Manufacturing low technology Construction Retail, food, beverages, tobacco Retail, IT Retail, pharmaceuticals, and other Wholesale, food, beverages, tobacco Wholesale, IT, equipment, cars Wholesale pharmaceuticals and other Banking, insurance, accounting, administrative management Utilities generation and distribution Telecom Hotels, transportation, other Total

Source: Author’s own.

3 2

Bosnia and Herzegovina Bulgaria

3 2

Croatia Czech Rep.

2

Hungary

5

6

Estonia

5

4

Latvia

3

23

Lithuania

2

2

Macedonia

Less than 1%

2 3

Moldova Poland

Less than 1% 25

Romania Russia Serbia and Montenegro Slovakia Slovenia Ukraine

8 5 Less than 1%

100

Less than 1% 6 3 22

8 2 2

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

159

EM – local embedness (0, 1). LD – matching local deﬁciencies (1–5). Control Variables: RE – regional effects: 1 – EU, 2 – US, 3 – Other. SZ – size: number of employees. IN – industry: 1 – manufacturing, high tech; 2 – manufacturing, low tech; 3 – construction; 4 – retail food, beverages, tobacco; 5 – retail IT, equipment; 6 – retail pharmaceutical and other; 7 – wholesale food, beverages, tobacco; 8 wholesale IT, equipment; 9 – wholesale pharmaceutical and other; 10 – banking, insurance, accounting, administrative management; 11 – utilities generation and distribution; 12 – telecom; 13 – hotels, transportation, farming. The MNCs’ subsidiaries included in the analysis represent successful as well as unsuccessful companies in approximately equal numbers. The average revenues across the sample are of USD 2,160,000, with an average number of 1,700 employees. The majority of the ﬁrms come from the European Union. The parametric descriptive statistics show there is signiﬁcant positive correlation (signiﬁcance below 10% in a two-tails test) between the success measure and the level of local embedding, and between the success measure and the industry. Nonparametric correlation measures (Kendall’s Tau and Spearman’s Rho) are also signiﬁcant and positive between the success variable and the local embedding variable and also between the success variable and the industry. Nonparametric correlations are signiﬁcant and negative between success and size variables. Nevertheless, the correlations between variables do not grant any conclusion of causal relationships or impact on probability of success. The ﬁndings of the logistic regression presented below allow for such generalizations.

RESULTS AND INTERPRETATION The statistical results for the complete model (570 observations, with 29 missing cases representing 5.1% of the data) are not signiﬁcantly different from the model that included only the sample for which all information was available (542 companies, with no missing cases). For this model, results show no inordinately large parameter estimates or standard errors, which means that there is no reason to suspect a problem with outcome groups perfectly predicted by any variable. There is also no indication of violation of the linearity in the logit for the model proposed. The results show no

160

ROXANA WRIGHT

problem with convergence, nor are the standard errors for parameters exceedingly large. No multicollinearity is evident, although there is some correlation between the level of adaptation variable and local capability development variable. No outliers were found. A test of the full model with all predictors against a constant-only model was statistically reliable, with a chi-square of 31.129 with low signiﬁcance (0.01), showing statistical signiﬁcance at 1% level. This indicates that the predictors, as a set, reliably distinguish between successful and unsuccessful ﬁrms. The variance in level of success accounted for is small, however, with Nagelkerke R square and Cox and Snell R square values low. Prediction of success is reasonable, with an overall success rate of 61%. More speciﬁcally, the value of chi-square as a measure of improvement due to the introduction of the independent variables (likelihood ratio as a measure of goodness of ﬁt) and its low p-value show that the outcome might be predicted from the set of variables considered. From the two measures of strength of association (low values for Nagelkerke R square and Cox and Snell R square), we cannot conclude on a strong association between the outcome and the variables. The log likelihood high value may also be an indication of imperfect ﬁt. However, Tabachnick and Fidell (2001) suggest that for large samples differences between models might have no practical importance. For large samples, classiﬁcation might be good (as is the case here) even though the model deviates from a perfect model. The authors conclude that a statistical difference between a ﬁtted model and the observed frequencies may not indicate a poor model with large samples. The analyst should thus keep in mind both the effects of sample size and the way the test works while interpreting the results. The classiﬁcation has a cut-off probability criterion of 0.5, and the 61% is reasonable, as a method of assessing the success of the model. The classiﬁcation evaluates the ability to predict correctly the outcome category of cases for which the outcome is known. The only way to improve the overall accuracy of classiﬁcation is to ﬁnd a better set of predictors. The classiﬁcation is reported for various models, however, and no signiﬁcant contribution to the correct percentage seemed to be added by other models. Selected results for the logistic regression run with no missing cases are included in Table 2. Based on the original model, the results show (Wald test) that three variables have an impact on the probability of success: level of adaptation (AD), market structure (MS), and industry (IN) at the 20% level of signiﬁcance. Only the industry variable is signiﬁcant at less than 1% signiﬁcance level. These ﬁndings suggest that higher levels of local adaptation as

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

Table 2.

161

Summary of Results for Logistic Model.

Summary of Tests/Results Omnibus tests of model coefﬁcients Measures of strength of association Classiﬁcation

Variables in the Equation VI AD SM AI LC MS EM LD RE SZ IN

w2 ¼ 31.126

Signiﬁcance ¼ 0.001

Cox and Snell R2 ¼ 0.056

Nagelkerke R2 ¼ 0.075

Overall percentage ¼ 60.900

Cut value ¼ 0.500

Coefﬁcient

Wald

Signiﬁcance

0.052 0.179 0.044 0.098 0.078 0.012 0.129 0.018 0.009 0.000 0.114

0.173 1.929 0.384 1.329 0.298 1.692 0.349 0.018 0.003 0.018 22.475

0.677 0.165 0.535 0.249 0.585 0.193 0.555 0.893 0.956 0.895 0.000

Source: Author’s own.

reﬂected in use of local brands, supply system, local management personnel, business models, etc. increase the probability of success for a multinational subsidiary in the CEE region. More competition in the market also increases the probability of success. The ﬁndings regarding industry show that companies operating services have a higher probability of success. The high signiﬁcance level for the adaptation and market structure variables should be noted. Although these variables could be interpreted to have some impact on success, the set probability of making a Type I error is high (20%). The part of the distribution that remains under the curve for the known population but is beyond critical value in the region of rejections is large. Only at this set probability are the two variables signiﬁcant and the interpretation above is valid. Such high potential for a Type I error cautions against the limitations of the statistical ﬁndings regarding the adaptation and market structure impact on success of CEE subsidiaries. Because some level of correlation between AD (level of adaptation) and LC (local capability development) was found, a model from which the AD

162

ROXANA WRIGHT

variable was removed was run. The results are similar to the original model, as regards goodness of ﬁt and classiﬁcation. The MS and IN variables were again the only signiﬁcant ones. The inclusion of interaction terms and the application of stepwise procedure provided no additional information. The model run with the entire sample, including the missing cases, shows slightly better goodness of ﬁt and classiﬁcation, but the variables found signiﬁcant in their impact on the probability of success are the same as in the results for the model run excluding missing cases. Reduced models (various combinations of explanatory variables) show low practicality. Results based on statistical inference show evidence supporting the alternative of Hypothesis 1 – vertical integration does not necessarily lead to successful operation in CEE. Although it was suggested that successful MNCs operating in CEE may use vertical integration to eliminate inefﬁciencies in the value chain, bridging across activities may not necessarily increase the probability of success. Successfully supporting suppliers to reach the level of quality required by their operations, as well as other ways of creating and improving distribution networks have proven to increase the abilities of foreign subsidiaries and the opportunities for proﬁt at the beginning of transition. We must consider, however, that the capabilities of local suppliers, distributors, and other members of the value chain have evolved signiﬁcantly in recent years. The ﬁnding that a company’s level of integration may not affect the probability of success could be related to the fact that foreign ﬁrms do not have to ‘‘internalize’’ different levels of operations, as local alternatives have similar potential. On the other hand, in countries and industries where supply and distribution systems are not developed satisfactorily, challenges in actively building an integrated value chain have not been successful, due to high costs, administrative blockages and a lack of experience/ability in creating a local or regional value chain. The statistical ﬁndings support Hypothesis 2 – medium degree of autonomy and loose integration with the corporation lead to successful operation in CEE. This result suggests that an intermediate level of autonomy is essential for a successful local operation, in that it lets subsidiaries test strategy and try different local tactics, allowing for endogenous learning processes to take place. A more loosely integrated afﬁliate has access to the resources of the MNC and local managerial freedom to generate diversity and enhance the corporation’s global capabilities. As discussed previously in hypothesis deﬁnition, retaining and developing context sensitivity and diversity can be valuable for the MNC. From an evolutionary perspective, permitting a variety of managerial practices and organizational structures gives the company a long-term beneﬁt of more options and a greater pool of capabilities.

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

163

In the idiosyncratic transition environment, loose integration with the headquarters may permit two-way learning and thus enhance the MNC’s global capabilities. Allowing a local operation to develop capabilities and solutions in the transition context requires not only a reasonable degree of autonomy but also access to resources for developing these capabilities. Experimentation may be needed to develop new managerial practices that correspond to the local values and resources that outperform the established ones. The results of the model presented here with regard to Hypothesis 3 advocate that MNCs do not necessarily increase their probability of successful operation in CEE countries by developing subsidiary-level capabilities. Although it is conceivable that companies that have acquired essential knowledge about the local environment and clients are in a better position, the ability to acquire the local knowledge may come from the general ability of the multinational to process knowledge. Although networking, establishing connections with national and local governments, and other behaviors of embedding in the local environment are considered by previous literature as essential for success in CEE, the ﬁndings here suggest that establishing close relationships with governments and communities is not related to the success of CEE subsidiaries (ﬁndings relating to Hypothesis 4). Creating close relations with banks and local governments, lobbying and pursuing reputation building tactics (playing an active role as a good citizen in the community) are part of an embedding strategy that many MNCs have followed in the CEE region. However, the local embedding may only be a support strategy of diminishing limitations while taking full advantage of opportunities, and may only have an insigniﬁcant impact on performance on its own. Finally, the level of adaptation is found to have a signiﬁcant impact on the probability of success. High adaptation as regards local management, local distribution systems, local products, and brands appears to be the key in operating winner subsidiaries. Finally, one of the control variables (industry) is noteworthy in its impact on success. Conversations with industry regulators, managers, and representatives of foreign chambers of commerce have consented on the high proﬁtability in the services sectors. In fact, the representative of the American Chamber of Commerce for CEE noted that most US multinationals are services companies, as the success expectation in the manufacturing sector is very low. The results presented support this expectation. The ﬁndings show no evidence to suggest that success may be related to strategic motivation of MNCs operating in the CEE region or to the capability of the ﬁrm to match local deﬁciencies.

164

ROXANA WRIGHT

In conclusion, it can be inferred that the strategies of complexity reduction are most effective in increasing the probability of success of a foreign subsidiary in CEE. The ﬁndings suggest low impact of complexity absorption strategies that may reﬂect the fact that thriving operations in countries in this region do not attempt to ‘‘absorb’’ the environment but rather, they anticipate change and convergence toward features of fully functioning market structures and institutions, adapt to the local customer, and reduce the perceived complexity of the setting. Although they may use localized strategies, they draw from strong multinational and headquarters capabilities. They also ﬁnd a trade-off between high levels of local autonomy that encourage experimentation (with possible loss of consistency, and continuous testing of unsuccessful strategic avenues) and low levels of autonomy hindering learning and potential synergy feedback. The services sector is more conducive to success. Competitive markets are usually the most dynamic and provide more incentives for strategies that generate sustainable advantages. The results of this analysis show the evolution of winning strategies in CEE in recent years. As the markets progress, competition increases and consumers become more sophisticated, companies need the backing of strong capabilities, high responsiveness to the market, and an adequate level of autonomy that allows for ﬂexibility but does not promote wasteful iterations on losing strategies.

CONCLUSIONS The present study investigates MNCs’ strategies to cope with resilient components of the idiosyncratic transition environments in CEE (such as embedding and drawing advantages from regional deﬁciencies) as well as strategies to react to dynamic components of the environment (such as knowledge acquisition and building or improving value chains). The research argues the appropriateness of a business model that allows for local responsiveness and relates successful implementation of strategies to local autonomy and combination of resources across company network. The dynamic environment in the region is viewed as an incubator for new and innovative strategies. The assumption is that of ‘‘accumulated wisdom’’ at the location, however, the adaptive strategies are viewed to be either positive (successful adaptation) or negative as a result of complexity and imperfect foresight. The descriptive and inferential statistics using a logistic model show that vertical integration does not necessarily lead to successful operation in the

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

165

CEE region, whereas a medium degree of autonomy does. One of the main statistical ﬁndings is that the MNCs do not necessarily increase their probability of success by developing subsidiary-level capabilities but rather can rely effectively on proven capabilities and a familiar strategy. Establishing close relations with governments is not related to success. It would appear that where companies have fewer options to rely on market mechanisms, they attempt to absorb the local environment. Strategies of complexity reduction are most effective in increasing the probability of success of a foreign subsidiary in CEE. Thriving operations in the region do not attempt to ‘‘absorb’’ the environment but anticipate change and convergence toward features of fully functioning market structures and institutions, adapt to the local customer, and reduce the perceived complexity of the setting. Although they may use localized strategies, they draw from strong multinational and headquarters capabilities. They also ﬁnd a trade-off between high levels of local autonomy that encourage experimentation (with possible loss of consistency and continuous testing of unsuccessful strategic avenues) and low levels of autonomy hindering learning and potential synergy feedback. The main contribution of the research presented here is that of a clear demarcation of what constitutes successful adaptation to the environment. The strategic approaches analyzed represent patterns of solutions. As a general and ﬁnal conclusion it is noted that the best performing strategies are those that are based on decisive action ensuring high quality learning and responsiveness. Leveraging strong global and regional capabilities and the knowledge acquisition experience of the ﬁrm are prerequisites for success. The limitations of the research presented here are related to the limited access to information and the accuracy of shared information from primary sources. The statistical methodology is based on qualitative data that are drawn from existing evidence. The quality of the data and the results depends on the quality and sometimes the interpretation of this evidence. The statistical results have limitations due to their general nature. The companies included represented as many heterogeneous agents across which the generalization occurred. The future research agenda could be expanded to other transition economies; also as an opportunity to test the validity of the framework in a different region with a similar environment. A comparable model may be applicable to MNCs’ subsidiaries operating in dynamic environments in general. The discussion of data included in the paper includes references to distinctive patterns of adaptation for companies from various home countries. The model can be expanded to include explanations related to cultural and psychic distance. Future research can focus exclusively on

166

ROXANA WRIGHT

multinationals from a particular country, with potential ﬁndings’ justiﬁcations drawing upon speciﬁc relationships between home and host country. The present research reaches across industries and generalizes on overall strategies. Industry-speciﬁc subsets of strategies can improve the accuracy of the ﬁndings and add essential details. As multinational subsidiaries become established in the region, future investigation can emphasize on the requirements to build sustainable advantage. The ﬁndings suggest that some companies with a ‘‘diversiﬁed’’ portfolio of knowledge and strategic options might be in a position to catch up from an inferior position. It would be interesting to see whether such cases exist or can be predicted in the near future. Transition toward market structures in the region may have also meant a transition in companies’ strategies – from using their ﬁrm-level capabilities to developing ﬁrm-level practices of incorporating new capabilities and local innovations.

NOTES 1. A case study of a French MNC (Yoruk, 2001) has shown that the company has been successfully supporting its suppliers to reach the level of quality required by its operations. In a study of marketing obstacles in a number of CEE countries, Batra (1996) has found that MNCs have been at the vanguard of creating and improving distribution networks at the beginning of transition. 2. Business and Company Resource Center is an integrated resource with company proﬁles, brand information, rankings, investment reports, company histories, chronologies, and periodicals. The information on companies (MNCs’ subsidiaries) located in CEE countries provided by this source includes addresses and contact information, a complete classiﬁcation by NAISC codes, annual sales and proﬁts, number of employees, productivity, and company management and proﬁle. 3. Mergent Online provides Internet-based access to a global company database, including business description, history, property, subsidiaries, ofﬁcers, and directors, as well as ﬁnancial statements of subsidiaries in CEE countries. The International Annual Reports module is integrated into the International Company Data module, Mergent’s global company database. The coverage includes subsidiary contact information, company proﬁle, primary NAISC classiﬁcation, business summary, and ﬁnancial highlights. 4. The main source of data for this research is Kompass Database. Kompass provides business information mainly for the international commercial and industrial community. Information collected using this database includes company information, proﬁle, and ﬁnancial data. This database is particularly useful as it reveals CEE subsidiaries’ associations with local organizations, a list of the MNC’s subsidiaries and their activities, a comprehensive history of the subsidiary, a description of operations owned in the primary value chain as well as a complete ﬁnancial proﬁle and a discussion of the local activities/operations.

Successful Adaptation Strategies of MNEs in Central and Eastern Europe

167

ACKNOWLEDGMENTS The author would like to thank Dr. Massood Samii, Chair of International Business, Southern New Hampshire University, Dr. Phillip Vos Fellman, Professor of International Business, Southern New Hampshire University, Dr. Yusaf Akbar, Professor of International Business, Southern New Hampshire University, and Dr. Bulent Aybar, Professor of International Business, Southern New Hampshire University for their support and guidance on this paper.

REFERENCES Batra, R. (1996). Marketing issues and challenges in transitional economies (Working Paper No. 12). Ann Arbor, MI: William Davidson Institute. Dyker, D. (2001). The dynamic impact on the Central-Eastern European economies of accession to the European Union: Social capability and technology absorption. Europe-Asia Studies, 53(7), 1001–1021. Lieb-Doczy, E., & Meyer, K. (2000). Context specificity of post-acquisition restructuring: An evolutionary perspective (Discussion Paper No. 6). London: London Business School, Center for New and Emerging Markets. Meyer, K. (2000). International business research on transition economies (Working Paper No. 32). Copenhagen, Denmark: Copenhagen Business School, Center for East European Studies. Pavlinek, P. (2004). Regional development implications of foreign direct investment in Central Europe. European Urban and Regional Studies, 11(1), 47–70. Political Risk Services. (2004). March 1 release: Investment climate – Romania country conditions. From The PRS Group International Country Risk Guide online http:// www.prsgroup.com/icrg/icrg.html Radosevic, S. (2004). Growth of enterprises thru alliances in Central Europe (Working Paper No. 36). London: University College London, Centre for Study of Social and Economic Change in Europe. Radosevic, S., & Rozeik, A. (2005). Foreign direct investment and restructuring in the automotive industry in Central and East Europe (Working Paper No. 53). London: University College London, Center for the Study of Economic and Social Change in Europe. Romanian Business Digest. (2004). Larive Romania: Direct investments in Romania. From http://rbd.doingbusiness.ro/2004_02/larive2004.pdf Tabachnick, B., & Fidell, L. (2001). Using multivariate statistics (4th ed.). Needham Heights, MA: Allyn & Bacon. UNCTAD. (2005). April 4 press release: World FDI ﬂows grew at an estimated 6% in 2004. From http://www.unctad.org/Templates/webﬂyer.asp?docid=5700. Yoruk, D. E. (2001). Industrial integration and growth of firm in transition economies: The case of a French multinational company (Working Paper No. 20). London: University College London, Centre for the Study of Economic and Social Change in Europe.

This page intentionally left blank

THE RELATIONSHIP BETWEEN ORGANIZATIONAL STRUCTURES AND PERFORMANCE: THE CASE OF THE FORTUNE 500 Nicole Avdelidou-Fischer ABSTRACT This paper investigates the relationship between organizational structures and the performance of FORTUNE 500 companies, which have always been among the most profitable and admired in the world. After a discussion of whether companies should organize regionally, nationally, or globally, the important assumption is made that each structural type utilizes resources differently in generating profit. Performance is conceptualized as Return on Capital Employed (RoCE) and Return per Employee (RpE). A sample of 50 companies was randomly selected. Testing revealed that structural types are positively related to financial performance, calculated as RoCE, with Multidivisional-structured companies outperforming Functional-structured ones; structural types are not related to human resource performance, calculated as RpE.

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 169–206 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07008-7

169

170

NICOLE AVDELIDOU-FISCHER

INTRODUCTION Problem Statement The National Bureau of Economic Research declares that the United States entered recession in March 2001 (BBC-News, 2001b). A slew of companies from various sectors issue proﬁt warnings, following the terrorist attacks on New York and the Pentagon (BBC-News, 2001a). In October 2002, BBC broadcasts that the US economy is still struggling to banish the specter of recession (BBC-News, 2002). For Ramesh (2002) the cause is the overinvestment and over-borrowing spree of the 1990s and the slowdown is a welcome corrective to these excesses. Miller (cited in Harrington, 2003) makes CFOs responsible for using the old accounting rules to pump up revenues, so as to have a higher stock price. A number of scapegoats could be blamed for this depressing state of affairs – terrorism bringing a sluggish economy, 2002’s sagging stock market, and a seemingly unending string of scandals come to mind. Still, given all the stumbles and falls in corporate America over the past years, most of the FORTUNE 500 companies look like Olympic sprinters (Boyle, 2002), considering that Wal-Mart, the top company on the 2002 list made over $6,670 million and even the bottom end, The New York Times, made over $444 million proﬁts – and that in bad times. No wonder that FORTUNE 500 companies have always been among the biggest, most proﬁtable, and most admired companies in the world (Kahn, 1998). Creating a multinational organization like GM or Coca-Cola that can respond quickly and creatively when crises strike or opportunity knocks, requires more than luck. So why do some companies perform better than others?

Possible Explanations Early economic theory depicted the organization as synonymous with the larger-than-life entrepreneur, the ‘‘peak coordinator’’ (Mintzberg, 1983; Papandreou, 1952). But experts agree that old style CEOs who were driving organizations to make a number rather than do the right thing, are out of fashion, or awaiting prison time; a different kind of leadership is called for (Gertner, 2002; Sellers, Mulcahy, & Notebaert, 2002; Stein, 2000). This confession would deﬁnitely be a relief for Mintzberg (1983, p. 115), who predicted that converting external inﬂuence into internal action effectively,

The Relationship between Organizational Structures and Performance

171

is beyond the capacity of one person. The CEO must coordinate people and align tasks through good organizational design; it is the ‘‘whole’’ that matters. Bower (2003) agrees that a person’s effectiveness, job satisfaction, and zest for work are all vitally affected by the structure of the organization in which he or she works. Back in 1998, Kahn highlighted that global strategy with a vision was what this year’s winners had in common. They were in the forefront of creating the truly multinational corporation, demonstrating abilities to seize opportunity from chaos and commit to long-term plans (Kahn, 1998). But even if the thought global strategy is as smart as P&G’s or Gillette’s, it will not get the ﬁrm far if it cannot be implemented. ‘‘You cannot be adaptive if there are 12 levels of decision-making for any given change’’ says Pfau, vice president of the Hay Group, and argues that an effective, ﬂat organizational structure is necessary (Kahn, 1998, p. 208). Whittington (1993, p. 112) concurs that management spends more time and energy on choosing strategies rather than implementing them, since even strategies that are well chosen fail because of poor implementation. Getting the organizational structures right for a particular strategy is thus clearly critical to practical performance. What these possible explanations have in common is that they all come down to one term: organizational structures. As suggested by OECD (1987, pp. 7–9), the way in which multinational enterprises structure and organize their activities is of great managerial importance, raising a general and wellknown issue; governments and labor are notably interested to the extent that these will inﬂuence country beneﬁts from direct investment, such as the level and type of investment, employment, or R&D. Equally, employee representatives have a legitimate interest to be informed about the decisionmaking structure within the enterprise. Robbins and Coulter (2002) stress the importance of measuring organizational performance and emphasize how vital an organizational structure is, since it is the vehicle through which managers can coordinate the activities of the various functions or divisions to exploit fully their expertise and capabilities. Business failure frequency is increasing in the high-competitive world industries (Eidleman, 1995). Several studies have concluded that soft keys, such as management and employees’ alignment are key determinants in producing value (CGEY, 2000; Ernst & Young, 1997; KPMG, 1999). And also for Porter (1980, pp. 50–51), organizational structure is one component of competitor analysis that requires comprehensive diagnosis. Although literature underlines the importance organizational structures play in ﬁnancial performance, and although earlier studies have chosen the FORTUNE 500 list as population of interest (Bethel & Liebeskind, 1993;

172

NICOLE AVDELIDOU-FISCHER

Capon, Farley, Lehmann, & Hulbert, 1992; Cummings, 2004; Horgan & Garnick, 2000; Moore & Reichert, 1983; Walker & Bain, 1989) there are, to my knowledge, no prior studies that have examined the direct relationship between FORTUNE 500 company performance and structural types. The purpose of this paper is to answer the following question: Is there a relationship between organizational structures and the performance of FORTUNE 500 companies?

REVIEW OF THE LITERATURE Organizational Structures The organization ﬁrst comes into being when an initial group of inﬂuencers join together to pursue a common mission (Mintzberg, 1983, p. 22). The way individuals relate to work towards this goal is determined by the structure of the organization (Duncan, 1981). Chandler deﬁnes structure as the design of organization through which the enterprise is administered: This design, whether formally or informally deﬁned, has two aspects. It includes, ﬁrst, the lines of authority and communication between the different administrative ofﬁces and ofﬁcers and, second, the information and data that ﬂow through these lines of communication and authority. Such lines and such data are essential to assure the effective coordination, appraisal, and planning so necessary in carrying out the basic goals and policies and in knitting together the total resources of the enterprise. (Chandler, 1966, p. 14)

Simply, organizational structure is the process by which organizations formally divide, group, and co-ordinate job tasks. The value creation activities of organizational members are meaningless unless some type of structure is used to assign people to tasks and connect activities (Galbraith, 1973). Coordination of effort within a function and between functions within an organization is by no means an easy task (McKenna, 2000, p. 423). Each function needs a structure designed to allow it to develop its skills and become more specialized and productive. Left to themselves, the functions may have little to say to one another, and value creation opportunities will be lost (Hill & Jones, 2001, p. 384). Accordingly, scholars suggest that organizational design is related to the revenue side of the equation (Hill & Jones, 2001, p. 386; Yunker, 1983, pp. 51–64). New information technologies, especially groupware and client-server, fast changing customer needs, competitor offerings, and more complicated products, require better integration of manufacturing, design, and marketing

The Relationship between Organizational Structures and Performance

173

functions. The workforce is becoming more heterogeneous sexually, racially, culturally, individually, and so on due to changing demographics and globalization of the labor market. This diversity is a source of conﬂict/ communication problems because of different styles of interaction, presentation, physical appearance, and dress (Borgatti, 2001). Companies cannot afford to ignore the fact that they are now operating in extremely complex environments where survival depends on the ability to understand and respond to multiple demands and opportunities (Ghoshal & Bartlett, 1999, p. 121). As Welch puts it, ‘‘Anytime there is change, there is opportunity. So it is paramount that an organization gets energized rather than paralyzed’’ (Kahn, 1998, p. 207). Organizational structures provide the framework for a social-operational-control system, inﬂuencing greatly individual and group behavior, helping organizations actively adapt to their environments. The basic building blocks of organizational structure are differentiation, which refers to the way in which a company divides itself into parts (divisions and functions), and integration, which refers to the way in which the parts are then combined. The two processes determine together how an organizational structure will operate (Hill & Jones, 2001, p. 385). Horizontal differentiation focuses on the division and grouping of tasks to meet the objectives of the business. Perhaps the ﬁrst person to address this issue formally was Harvard Business historian Chandler (cited in Hill & Jones, 2001, p. 394). Chandler’s famous hypothesis is that structure follows strategy. Companies are quickly recognizing that if they are to retain their ﬂexibility, they need to develop a range of internal perspectives (in order to understand the environment) and diversity of resources and capabilities (in order to respond to it). Restating the general thesis, Chandler perceives strategic growth as resulting from an awareness of the opportunities and needs to employ existing or expanding resources more proﬁtably (Chandler, 1966, p. 15). ‘‘In such a world,’’ advise Ghoshal and Bartlett (1999, p. 121), ‘‘a company’s organizational structure must reﬂect, not deny, the complexity of its external environment.’’ That is, unless new structures are developed to meet new administrative needs, which result from opportunities and strengths, economies of growth and size cannot be realized. For example, when General Motors (GM) faced bankruptcy during the 1920 crisis, Durant’s retirement made it possible for Sloan and du Pont to create a decentralized management structure to cope effectively with the company’s widely varying products and markets, an approach adopted at DuPont just a year later. Sloan ﬁrmly believed that divisional independence encouraged initiative and innovation. By the time du Pont retired from the

174

NICOLE AVDELIDOU-FISCHER

Board of Directors in 1928, GM became the largest corporation in the world (Chandler, 1966; DuPont Heritage, 2002). So is structure a reﬂection of strategy? The question ‘‘what is strategy?’’ leads to the work of Whittington (1993). He devised a historical typology of strategy theory, outlining the nature and the assumptions of four ‘‘conceptions’’ within which most of the literature could be placed. His view is that a dominant ‘‘concept’’ is identiﬁable in each decade: the Classic in the 1960s, the Processual in the 1970s, the Evolutionary in the 1980s, and the Systemic in the 1990s (Whittington, 1993, p. 40). According to this view, Porter is positioned in the Classic approach. He is conﬁdent that ‘‘every ﬁrm competing in an industry has a competitive strategy. This may have been developed explicitly through a planning process, or implicitly through the activities of the various functional departmentsy the emphasis being placed on strategic planning today in ﬁrms in the United States and abroad reﬂects the proposition that there are signiﬁcant beneﬁts to gain through an explicit process of formulating strategy, to insure that at least the policies (if not the actions) of functional departments are coordinated and directed at some common set of goals’’ (Porter, 1980, p. xiii). This raises two questions. First, could the best strategy be developed explicitly through a planning process? ‘‘No’’ according to Mintzberg (1994), who concludes that strategy cannot be planned because planning is about analysis and strategy is about synthesis. That is why the process has failed so often and dramatically. Second (also in agreement with Whittington, 1993, p. 111), is strategic planning what unites all individuals who make up an organization around the effective pursuit of a coherent goal? Mintzberg is skeptical: ‘‘There can be little doubt that planning can serve as important mechanism to knit disparate activities together. But to consider this imperative is another matter. Coordination can be effected in other ways too y And even when plans do serve in a coordinating role, it cannot be assumed that planning has created those plans’’ (Mintzberg, 1994, p. 17). The message of the Classical and Evolutionary Schools is obvious: change structure to match changes in strategies. Donaldson’s (1987) statistical analysis of the correlation between performance and appropriate structure among large American corporations certainly conﬁrmed the ﬁnancial logic of this argument: diversiﬁed ﬁrms with Multidivisional structures and nondiversiﬁed ﬁrms with Functional structures performed better in terms of proﬁt margin and return on investment growth than their mismatched peers. One could say, that according to these results and at least in the United States, ﬁtting strategy to structure pays. Porter’s Competitive Strategy has met the needs of both academics and managers who were looking for a

The Relationship between Organizational Structures and Performance

175

strategy theory and models that could be empirically tested, e.g.,the ﬁveforces analysis. Harﬁeld (1998) ﬁnds Porter’s work attractive because it provides the illusion of security, control, and legitimacy in the face of uncertainty, but also problematic because if every business adopted the strategies advocated, none would be able to secure a competitive advantage. The need to match structure with strategy is well accepted but as Whittington, Pettigrew, and Ruigrok argue, the old notions need to be re-examined (FT, 1999). For the Processualists, the direction of causality is reversed. They not only point to how organizational structures in practice fail to ﬁt strategies, but also how they can actually shape strategies (Whittington, 1993, p. 111). The Processual view is that organizations are partnerships of individuals, each of whom brings their own personal objectives and cognitive biases to the organization. For scholars like Cyert, March, or Simon, people are only ‘‘restrictedly rational’’ (Harﬁeld, 1998). Because of this limitation, strategy is a gradual process of negotiation and adjustment of routines to environmental messages, which eventually force themselves on the manager’s attention. Thus, strategies are often ‘‘emergent,’’ their coherence accruing through action and perceived in retrospect, while successive, small steps eventually merge into a pattern (Harﬁeld, 1998). So, is strategy a reﬂection of structure? In practice, the strategies of most organizations are probably a combination of the intended and the emergent (Hill & Jones, 2001, p. 22). Mintzberg (1994, p. 25) maintains that few, if any, strategies can be purely intended, and few can be purely emergent. One suggests no learning, the other no control. All real-world strategies need to mix these in some way – to attempt to control without stopping the learning process. An organization’s capability to produce them is fostered by the organization’s structure. Still, research evidence seems to indicate that formal planning systems do help companies make better strategic decisions; e.g., the study from Miller and Cardinal in 1994 (cited in Hill & Jones, 2001, p. 23) analyzed in detail the results of 26 previously published studies on the relationship between strategic planning and company performance, concluding that, on average, strategic planning does indeed have a positive impact on company performance and hence is a valuable activity. But should companies organize regionally, nationally, or globally; centralize or decentralize control; build around functions, markets, or products? Taking a look at 21st century organizations, a variety of ways to structure appear: The Simple Structure. This is normally used by the small entrepreneurial company producing a single product or a few related ones for a speciﬁc market segment; no formal arrangements regarding organization exist. If

176

NICOLE AVDELIDOU-FISCHER

the company wants to grow and perform all tasks required by a rapidly expanding market, it needs a more complex form of horizontal differentiation (Hill & Jones, 2001, p. 394). It should be noted that none of the FORTUNE 500 companies is structured this way (Fortune, 2003). The Functional Structure. This is based on the primary activities that have to be carried out, such as production, ﬁnance, and accounting, marketing and personnel. When a company enters and competes in an international market, it just adds a subsidiary in the foreign country to handle sales and distribution. The Functional structure is typically found in smaller companies, or those with narrow, rather than diverse, product ranges. It allows greater operational control at a senior level – the CEO is in touch with all areas, and there is a clear deﬁnition of roles and tasks. It is a very popular structure; however, there are disadvantages, particularly as organizations become larger or more diverse. In such circumstances, senior managers might be burdened with everyday operational issues, neglecting the overall strategy; they rely on their specialist skills rather than taking a strategic perspective on problems (Johnson & Scholes, 1999, p. 403). In comparison to other forms, such as the Matrix, the Functional structure is inﬂexible and, thus, less suited for ‘‘ad hoc’’ projects. Perhaps the biggest weakness is the conﬂict that tends to arise between departments. It is not uncommon for ‘‘demarcation lines’’ to be generated with managers from each department unwilling to work with others. There is a tendency to ignore the ‘‘big picture’’ (RBR, 2001). The Geographic Structure or Territorial Grouping. When an organization is geographically dispersed, groupings are created by region within a country or, where appropriate, by world regions. Frequently, sales or marketing groups are divided by geographic region (McKenna, 2000, p. 433). Signiﬁcant beneﬁts accruing to this type of structure are that the organization can cater for the speciﬁc needs of a given location, achieving better market coverage, and that it can provide training to familiarize managers with operations in the ﬁeld (Hodgetts, 1991). A study by Berenbeim (1983) exposed a relationship between Geographic structure characteristics and an advantageous approach towards delegation and control. In comparison to other structures, the Geographic is characterized by a greater degree of local autonomy, which fosters entrepreneurship and motivation (Leavitt & Lipman-Blumen, 1995, p. 113; Nayak & Ketteringham, 1997, p. 369; Timmons, 1999, p. 224). Drawbacks could include an increase in administrative expenses, advertising, promotion and sales outlet costs, and the isolation the work groups in

The Relationship between Organizational Structures and Performance

177

the regions feel when distanced from the nerve center of the organization. This could cultivate loyalty to the regional work group that is of greater magnitude than commitment to the larger organization. A difﬁculty for top management is the exercise of overall control in the face of signiﬁcant devolution of authority to Territorial Groupings (McKenna, 2000, p. 433). The Product-Team Structure. This has been a major innovation in recent years. Its advantages are similar to those of the Matrix structure, but it is much easier and far less costly to operate because of the way people are organized into permanent cross-functional teams, instead of being assigned only temporarily to different projects. Tasks are divided along product or projects lines to reduce bureaucratic costs and to increase management’s ability to monitor and control the manufacturing process. This results in much lower costs associated with coordinating activities than in a Matrix structure, in which tasks and reporting relationships change rapidly (Hill & Jones, 2001, pp. 404–405). When a company embarks on a global strategy, it locates manufacturing and other value creation activities in the lowest-cost global location to increase efﬁciency, quality, and innovation. A product-group headquarters (similar to SBU headquarters) is created to coordinate the activities of the domestic and foreign divisions within the product group, providing, at the same time, a centralized control. This way, managers can decide which value creation activities should be performed in which country to increase efﬁciency (Hill & Jones, 2001, pp. 464–465). The Multidivisional Structure. Many commentators compare the current period of organizational change to the birth of the divisional organization 80 years ago in the great American corporations like DuPont and GM. Consistent with Chandler’s premise, these companies having changed their strategies from focus to diversiﬁcation, needed the new decentralized structure to match (FT, 1999). This structure is subdivided into units (divisions) on the basis of products, services, geographical areas, or the processes of the enterprise. This way, each division is able to concentrate on the problems and opportunities of its particular business environment. The products and markets in which the company operates may be so diverse that it would be impractical to bring the tasks together in a single body. Thus, a division can be created, which relates closely to an SBU, allowing a tailoring of the product/market strategy to the requirements of that SBU and improving the ownership of the strategy by divisional staff (Johnson & Scholes, 1999, p. 404). When a company enters an international market, it adds an International Division to coordinate the ﬂow of different products across different countries (Hill & Jones, 2001, p. 464).

178

NICOLE AVDELIDOU-FISCHER

The research that Armour and Teece performed in 1978 (Hill & Jones, 2001, pp. 398–399) suggests that large companies that adopt the Multidivisional structure outperform those that retain the Functional; perhaps, this is the reason why this structure is said to be adopted by more than 90% of all large US corporations. However, as with all of the structural types, there are disadvantages to the Multidivisional structure. Some, mentioned by Hill and Jones (2001, pp. 399–402), are the following: establishing the divisional-corporate authority relationship, distortion of information, competition for resources, transfer pricing, short-term Research and Development focus, and, ﬁnally, high bureaucratic costs. Nearly a century after the innovations of DuPont and GM, it has been discovered that the old divisional form needs extending (FT, 1999). Most multinationals are now in the third phase of international development (Kahn, 1998). Firms ﬁrst regarded international operations as an ugly stepchild – simply, ‘‘one old executive with an airline ticket.’’ Later, they realized the value of internationalization but tended to clone their operations, creating a loose federation of businesses, each of which duplicated the hierarchy of the home ofﬁce. Now, power has shifted to units responsible for performing a given function globally, and the emphasis is on optimizing processes worldwide. This third phase is the ‘‘Matrix.’’ The Matrix Structure. This was developed by NASA in the United States and based on two forms of horizontal differentiation (RBR, 2001). It could also be said that it is a combination of structures, operating in tandem. Activities on the vertical axis are grouped by function or geographical divisions, while the horizontal pattern is based on differentiation by project or product group. The result is a complex network of reporting relationships among projects and functions. Employees inside the Matrix have two bosses, namely, a functional boss, who is the head of a function, and a project boss, who is responsible for managing the individual projects (Hill & Jones, 2001, p. 402). And this is perhaps its most obvious disadvantage; the classical school as represented by Taylor, Fayol, and Urwick (Armstrong, 2001, p. 183) has been stressing, from early times, the importance of unit of command. That is, each individual employee should have only one supervisor. The division of priorities can lead to conﬂict and lack of clarity (RBR, 2001). This structure has many strengths, which help in overcoming many problems within the Matrix (Bolton, 2002; Decker, 1996; Hill & Jones, 2001):

The Relationship between Organizational Structures and Performance

179

Team members control their own behavior. This freedom provides the autonomy to motivate employees. Flexibility. Decisions on major projects are made quickly, customer response time increases. Participation in project teams allows team members to monitor other members and learn from each other. As the project goes through different stages, different specialists from different functions are required. Therefore, the Matrix can make maximum use of employees’ skills as existing projects are completed and new ones come into existence. A Matrix structure is ﬂat and requires a minimum of direct hierarchical control by supervisors, leaving top management free to concentrate on strategic issues. This operating reality allows companies to break away from the functionally dominated model institutionalized by the old organizational chart. For example, Intel’s Matrix relationships link products and functional units, shared relationships becoming the norm (Ghoshal & Bartlett, 1999, p. 122). In their study, Backstro¨m and Ladan (2002) introduce a Matrix structure to dysfunctional organizations, also with the goal of enabling the principles of a learning organization. Positive effects like holistic awareness, alignment, ﬂexibility, learning, motivation, cooperation, and shorter time to decision are reported. It is remarkable how each structural type shapes the six elements that divide and group labor (Candea, 1976; Sablynski, 2002): work specialization, departmentalization, chain of command, span of control, centralization vs. decentralization, and formalization. Several examples illustrate how thought-through organizational structures rejuvenated PepsiCo’s Frito-Lay division and raised product quality (Sellers & Barlyn, 1996), supported Kmart’s plan to exit bankruptcy (Howell, 2002, p. 1), and pulled Shell out of its midlife crisis, moving it into proﬁtable new businesses (Guyon, 1997). Managing the strategy-structure relationship can be really difﬁcult; when the number of hierarchical levels in a Functional structure becomes too great, the structure also becomes too expensive; motivational problems and information distortion appear. Depending on a company’s situation, problems can be solved and bureaucratic costs reduced by a decentralized Product-Team Structure. Globalization and the search for unsaturated markets bring growth through Geographical dispersion to administer local ﬁeld units. As company size increases, however, decentralized structures may

180

NICOLE AVDELIDOU-FISCHER

cause coordination problems. Careful organizational design can help a ﬁrm grow and diversify and increase people performance while economizing on costs without becoming too tall or too decentralized. Does this mean that the type of structure can make a difference? Is there really a relationship between organizational structures and performance?

SELECTION OF PERFORMANCE MEASURES AND FORMULATION OF HYPOTHESES Performance In the previous pages, the imprecise term ‘‘performance’’ was used. The diversity of its possible measures stresses the necessity of a comprehensive framework that relates it to structure and the importance of selecting the appropriate dimension. Agreeing upon how to measure performance is not that simple since there are so many aspects of it. Each academic ﬁeld offers a different point of view. They range from general ﬁnancial outcomes to speciﬁc subjective ones (Bello & Gilliland, 1997; Campbell, 1990; Chaudron, 2001; Levinson, 1992; Robbins & Coulter, 2002). Nevertheless, accepting the notion that the ultimate objective of a ‘‘for proﬁt organization’’ is the quest for proﬁts and their maximization (Mintzberg, 1983, p. 9) and also accepting that monetary performance is the critical, ‘‘bottom line’’ kind of result that companies must deliver to survive (Chaudron, 2001), then the core of the concept consists of a plethora of outcome-based ﬁnancial indicators. Absolute performance measures will be excluded because they face a comparability problem and limit the focus of the study. They may be vulnerable to the size of the company, market segment, industry, and current market share, to mention only a few factors. They are also hard to compare in cross-country studies (Styles, 1998, pp. 18–19). This is why accounting numbers will be translated into relative values; the technique to be used for evaluating companies is ratio analysis. A combination of measures will be considered in order to capture the rich complexity of performance (Bello & Gilliland, 1997). Given that there are dozens and dozens of acknowledged ﬁnancial ratios, one needs to be selective in their use (Kerr, 2001). Speciﬁcally, the selection of the measures that will be used to gather the data should be driven by the research problem and the focus of the study (Agathangelou, 2001). In topics like organizational structure and performance, the efﬁciency with which

The Relationship between Organizational Structures and Performance

181

resources are utilized needs to be carefully assessed, e.g., ofﬁces, buildings, and costs from operations made for generating proﬁt. For example, when an organization is Geographically structured, it clones its operations. If this structure is not aligned to the strategy, dividing sales or marketing groups by geographic region and duplicating the hierarchy of the home ofﬁce will needlessly increase administrative expenses, advertising, promotion, and sales outlet costs. Financial performance might be negatively affected. As noted by Armstrong (2001, pp. 143–153), the role/job situation individuals are in is one of the inﬂuencers of human resource performance. A job consists of a group of ﬁnite tasks to be ‘‘performed’’ (Armstrong, 2001, p. 350). Since each structural type divides, groups, and coordinates job tasks in a different way, it can be expected that each structural type would have a different impact on human resource performance. For example, the Matrix structure is a complex network of reporting relationships among projects and functions. Its most obvious disadvantage is that employees inside it have two bosses. Without a unit of command, the division of priorities can lead to conﬂict and lack of clarity. Human resource performance might be negatively affected. As exempliﬁed in the two previous paragraphs, several structural characteristics have power over company performance in terms of ﬁnancial and human resources. Interpreting these two dimensions with the aim of understanding the performance of a ﬁrm in the recent past requires identifying factors believed to be relevant to these changes. The ﬁrst factor is the efﬁciency with which the long-term capital (property, ofﬁces, plants, machinery, and intangible assets) has been employed in a for-proﬁt organization. This will be measured with the Return on Capital Employed ratio (RoCE ¼ Proﬁt/Capital Employed 100), which is considered crucial for testing proﬁtability and ﬁnancial performance (Kerr, 2001). It is intended to focus attention on the efﬁciency with which all of the resources available to the company have been utilized (Pendlebury & Groves, 2001, p. 112). The second factor is the effectiveness of the utilization of labor. A quick glance at a company’s Proﬁt and Loss account shows that there are several levels of proﬁt. Which one of those would make sense to use for the human resource performance test? As mentioned before, the type of organizational structure can be held responsible for the level of distribution costs, cost of sales, and administrative expenses of the company. But it cannot be held responsible for the level of interest payments, as the markets set interest rates, and it cannot be held responsible for the amount of taxation the company pays, as the government sets taxation rates (Kerr, 2001). For this

182

NICOLE AVDELIDOU-FISCHER

reason, it would deﬁnitely make sense to use Operating Proﬁt as the proﬁt measure. The absolute numbers will be translated into relative values to facilitate company comparisons, through the Return per Employee ratio (RpE ¼ Operating Proﬁt/Number of Employees), which gives the proﬁtability of the company per employee (Pendlebury & Groves, 2001, p. 276).

Hypotheses Organizations are now operating in extremely complex environments where survival depends on the ability to understand and respond to a wide variety of forces, some internal to the enterprise, others external. As these forces change over time, and many of them have changed quite dramatically in the last decades, it can be expected that these will often be associated with changes in what is felt to be the most suitable organizational structure (OECD, 1987, p. 42). Unless new structures are developed to meet new functional and administrative needs, company performance cannot be realized. The central question: ‘‘Is there a relationship between organizational structures and the performance of FORTUNE 500 companies?’’ is now developed into testable hypotheses cast in a ‘‘hypothetico-deductive’’ form.

Structures and Return on Capital Employed The Functional Structure is based on the primary activities that have to be carried out, and, as organizations become larger or more diverse, the number of levels increases, together with bureaucratic costs; its local character lifts distribution costs. The duplication of specialist services in the Multidivisional structure is expensive to run and manage. When an organization is Geographically structured, it clones its operations. Dividing sales or marketing groups by geographic region and duplicating the hierarchy of the home ofﬁce, increases administrative expenses, advertising, promotion and sales outlet costs. All these factors can affect a company’s ﬁnancial performance negatively. In the Product-Team structure, functional specialists are placed in permanent cross-functional teams and tasks are divided along product or projects lines to reduce staff costs. Employees inside the Matrix structure have two bosses, which raises bureaucratic costs. But the structure’s ﬂexibility makes maximum use of employee’s skills; decisions on major projects are made faster, and customer response time increases. These factors can affect a company’s ﬁnancial performance positively.

The Relationship between Organizational Structures and Performance

183

Although the degree to which each structure affects a company’s ﬁnancial performance cannot be deﬁned in advance, it is accepted that some attributes make companies perform better than others. We can also distinguish how each of these attributes ‘‘groups’’ companies under a structural name. Consequently, it is suggested that a signiﬁcant RoCE difference will be observed at least between two of the structural groups. Hence, HA. IF companies are categorized by structural type THEN a signiﬁcant RoCE difference will be observed between two or more of the groups. FORTUNE 500 companies are among the largest in America. A centralized organizational structure is counterproductive in this case, because it slows response time and lowers the incentives of those at the local level (Porter, 1980, p. 211). From a practical point of view, Multidivisional structures are ‘‘better suited’’ to be run in a decentralized manner, while Functional structures suggest that a rather centralized management process be enforced (Candea, 1976, p. 4). Accordingly, Armour and Teece’s research from 1978 (cited in Hill & Jones, 2001, p. 399) suggests that large companies that adopt the Multidivisional structure outperform those that retain the Functional. These arguments suggest HB. IF companies have a Multidivisional structure THEN the RoCE will be higher than from companies with a Functional structure. A major innovation in recent years has been the Product-team structure. It enjoys the advantages of the Matrix structure, but it is much easier and far less costly to operate. Tasks are divided along product or projects lines to reduce costs associated with coordinating activities and to increase management’s ability to monitor and control processes. The following hypothesis is, therefore, suggested: HC. IF companies have a Product-Team structure THEN the RoCE will be higher than from companies with a Matrix structure. Structures and Return per Employee In a Functional structure, senior managers are burdened with everyday operational issues, neglecting the overall strategy. Functional structure is inﬂexible, and there is conﬂict between departments. The Matrix structure is a complex network of reporting relationships among projects and functions. Its most obvious disadvantage is that employees inside it have two bosses. Without a unit of command the division of priorities can lead to conﬂict and

184

NICOLE AVDELIDOU-FISCHER

lack of clarity. All these factors can affect a company’s human resource performance negatively. A signiﬁcant beneﬁt of the Geographically structured organization is that it can offer Cross-Cultural Marketing and Team Building Trainings, to familiarize managers with operations in the ﬁeld and better cater for the speciﬁc needs of a given location. Work groups cultivate regional loyalty. The Multidivisional structure enables divisional staff to concentrate on SBU problems and opportunities and tailor the product/market strategy to the requirements of that SBU. In the Product-Team structure, permanent crossfunctional teams result in better coordinating activities. Dividing tasks along product or projects lines increases management’s ability to monitor and control the processes. All these factors can affect a company’s human resource performance positively. Again the degree to which each structure affects a company’s human resource performance cannot be deﬁned in advance, but it is accepted that some companies perform better than others. Consequently, it can be suggested that a signiﬁcant RpE difference will be observed at least between two of the structural groups. Hence, HD. IF companies are categorized by structural type THEN a signiﬁcant RpE difference will be observed between two or more of the groups. The study by Berenbeim (1983), which examined various decision areas according to the features of the company, found a number of relationships between such company characteristics and the general approach towards delegation and control. In comparison with other structures, companies organized along geographic lines may see themselves more as a federation of local companies, associated, therefore, with a greater degree of local autonomy. Work group loyalty and motivation is a signiﬁcant advantage of the Geographic structure. The following hypothesis is therefore suggested: HE. IF companies have a Geographic structure THEN they will have a higher RpE than other companies. A relationship between organizational structure and company performance is generally hypothesized. An important assumption that is made, supported by the literature review, is that each structural type utilizes resources differently in generating proﬁt. The hypotheses will be empirically examined and discussed in the next parts of the paper.

The Relationship between Organizational Structures and Performance

185

METHODOLOGY Data The previous sections provided a list of constructs (variables). There are ﬁve basic types of organizational structures: Functional, Geographic, ProductTeam, Multidivisional, and Matrix. Although the variable is qualitative, it can be operationalized in a way that allows reducing it to some quantiﬁable index by assigning numbers to these types using a nominal scale. This is a pure labeling activity; there is no logical/natural order. That is, they are simply listed in categories alphabetically. These categories are mutually exclusive and exhaustive. RoCE is equal to Proﬁt divided by Capital Employed ( 100) and RpE is equal to Operating Proﬁt divided by the Number of Employees. Values are snapshot ﬁgures for ﬁscal 2002; the terms involved are deﬁned as follows: Proﬁt is Revenue minus cost. Capital Employed is taken to be the total assets of the company: property, ofﬁces, and plants; machinery, stock, and intangible assets (Dyson, 2001, p. 44). Operating Proﬁt/Operating Income is the revenue from a ﬁrm’s regular activities after cost of sales, distribution costs, and administrative expenses, but before the deduction of interest and tax (Harvey, 2003). Number of Employees is the number of permanent or full-time equivalent employees, as reported by the company in its 10-K report, who work in return for ﬁnancial/other compensation. All terms are numeric and hence, quantiﬁable. The data were collected from audited Annual Reports and 10-K Forms, prepared under the regulations of the US Securities and Exchange Commission and recognized as reliable sources of ﬁnancial and other signiﬁcant information.

Population and Sampling For almost 50 years, Fortune Magazine has been ranking the top 500 USbased corporations with the largest revenue in the past year – calculated using publicly available data. The result is their annual FORTUNE 500 list. This is the population of interest. The basic unit to be examined is each company. The necessary sample size was found to be n ¼ 46:25; and was rounded up to 50, which is 10% of the

186

NICOLE AVDELIDOU-FISCHER

population. The complete 2002 FORTUNE 500 frame was downloaded and saved in the Statistical Package for Social Sciences (SPSS), keeping the original rank-order by revenues (descending), where each company was numbered from 1 to 500. The random sampling facility generated 50 ﬁgures. The corresponding companies were selected and results were stored in one major ﬁle. Data Collection The sample was contacted in two phases. In the ﬁrst phase, copies of the audited Annual Reports and 10-K Forms were ordered electronically. All companies offered in their Web sites several contact possibilities; some provided e-mail addresses to directly contact an expert. This e-mail also functioned as a notiﬁcation of the research, encouraging the receiver to mail back the contact details of a person – preferably an executive – who is most directly involved in HR or corporate strategy matters. Eleven companies (22%) offered the link and the option to download the entire reports or speciﬁc sections; most of them had chosen not to distribute hard copies of the 2002 reports in 2003, and the rest were not yet ready with printing. The requested ﬁnancial reports were received, on average, 22.31 days after they were ordered (Table 1). It was necessary to conduct several follow-up e-mailings, spaced about two weeks apart. Two companies did not answer; their reports were downloaded. The minimum interval was 0 days (one company answered the same day), and the maximum was 48 days. The second phase was concerned with identifying the structural types that companies in the sample had during ﬁscal 2002. The instrument to collect the desired primary data was a PowerPoint presentation, which described and showed the ﬁve basic structural types that the literature review provided. The pilot was administered to ﬁve of the punctual respondents who had participated in phase 1, was corrected according to comments, and was then sent to the rest of the 45 companies. During this phase, the response duration fell remarkably: 23 (46%) of the companies answered within the ﬁrst week (Table 2). One of them refused to provide the requested information due to its strategic and competitive nature. Nine companies (18%) answered during the second week; ﬁve companies, unable to settle on a type, requested a telephone interview to answer speciﬁc questions for which they would not feel comfortable mailing/writing answers due to their strategic nature. Seven companies did not answer at all. Although the ﬁrst letter encouraged receivers to preferably identify executives who are most directly involved in HR or Corporate strategy

The Relationship between Organizational Structures and Performance

Data Collection: Phase I.

Table 1.

Annual reports 48 2

Duration in days Mean Median

22.31 19

Mode

19.00

Std. deviation

12.61

Minimum

0

Maximum

48

10 9 8 7 6 5 4 3 2 1 0

N=48

0 5 10 15 20 25 30 35 40 45 50 Duration in days, shown in midpoints

7

2

20

3

3

Customer service/ affairs

2

1

3

Human resources

2

3

1

6

E-Center/ web site

2

1

2

5

4

1

1

6

21

16

6

43

Media/Public relations Department

Corporate comm./ offices Total

Frequencies

Female Male No Info Total 11

Investor relations

Direct e-mail

N:12

Mean

19.67

Minimum

7

Maximum

36

Mail to IR dept.

N:21

Mean

26.71

Minimum

9

Maximum

48

Web contact form

N:15

Mean

18.27

Minimum

0

Maximum

45

Data Collection: Phase II.

Table 2.

Respondent demographics

Duration by contact style

Duration statistics

Valid Missing

Frequencies

N

187

25

N=43

20 15 10 5

0 1st 2nd 3rd 4th Response duration shown in weeks

matters, only 13 of the respondents were from management. Accordingly, the qualitative data are collected from employees embodying a diversity of departments and job-titles, ranging from the simple Representative of the Investor Relations Department, to the Vice President of Corporate Communications. The majority (46.5%) of the respondents were from the Investor Relations Department. People on various positions may hold different views regarding patterns of interaction and organizational structures. For example, managers derive a sense of identity from the afﬁliation with a company or their connection to social groups (SBU, function) within the company; in large and heterogeneous organizations, they tend to identify more strongly with their immediate work groups than with the organization as a whole (Houston, Walker, Hutt, & Reingen, 2001, pp. 19–35). As the aim is not simply to save results

188

NICOLE AVDELIDOU-FISCHER

of which structure the respondents ‘‘believe’’ their organization has, all answers were controlled for reliability. Data were collected independently, and a transcript with results was prepared during the time it took the companies to respond. These were compared with the respondents’ results; disagreements in seven cases were captured but were resolved by discussion. An independent judge was consulted for the cases where no answer was received.

ANALYSIS OF FINDINGS Descriptive Statistics All records were ﬁrst collected in an Excel ﬁle in order to compute the ratios. On average, ﬁrms in the sample made $969.96 millions proﬁt, with 3.87% RoCE and about $251,850 thousands RpE (Table 3). Summarizing the categorical variable ‘‘organizational structure’’ in a frequency table displays that the highest percentage of cases falls under the Multidivisional category. This result was expected to be higher, considering the statement from Hill and Jones (2001) that Multidivisional structure has now been adopted by more than 90% of all large US corporations. Since many statistical tests assume data are normally distributed, the ﬁrst step was to check the data distribution. Inspecting the skewness, kurtosis, and their standard errors (Table 3), it was concluded that none of the distributions is a normal one. Exploring the data (see the appendix) revealed two most extreme scores. After eliminating them, RoCE does not differ Table 3.

Descriptive Statistics.

Case summaries

N

Minim

Profit

50

-1,303.00

7,829.00

969.96

1,732.20

2.09

.34

4.86

.66

Total assets

50

1,095.70

887,515.00

61,589.77

163,348.46

3.88

.34

15.54

.66

Return on capital empl Operating profit

50

-23.19

17.11

3.87

7.19

-.97

.34

3.60

.66

50

-894.90

46,335.00

2,866.37

7,093.32

5.11

.34

29.80

.66

Number of employees

50

2,003

355,421

51,128.56

67,776.50

3.04

.34

10.36

.66

Return per employee

50 -73,678.58 9,653,125.00 251,849.14

1,359,453.95

7.03

.34

49.57

.66

Structures Frequency Valid percent

Functional

Maxim

Mean

Geographic

Matrix

Std. deviation

Skewness Std. E Kurtosis Std. E

Multidivisional

Product-Team

Total

12

8

7

14

9

50

24%

16%

14%

28%

18%

100%

Note: All values in $millions, except RoCE (in %), employees and RpE (in thousands).

The Relationship between Organizational Structures and Performance

189

signiﬁcantly from a normal distribution. Therefore, RoCE hypotheses are tested with parametric tests and RpE hypotheses with non-parametric ones. The hypotheses are empirically examined and discussed in the same sequence they appeared at the end of the literature review section. Structures and Return on Capital Employed HA. One-way ANOVA tests for differences between group means. The dependent variable is quantitative; the factor variable is categorical. The following null and alternative hypotheses are implied: The null hypothesis states that there will be no observed differences between group means; The alternative hypothesis states that there is a signiﬁcant difference, at least between two of the means. A report table lists the selected statistics, describing the distribution of the dependent variable for each group. For the Geographic and Product-team groups, the mean is approximately equal to the median, suggesting a symmetrical distribution of RoCE. Plotting the means for the groups illustrates how they differ, but if one conducts a study with two or more groups, the resulting group means will almost never be identical. The task is to decide if the independent variable has inﬂuenced the group means; this decision will be based on the outcome of a signiﬁcance test. The ANOVA table compares the means for the different groups, partitioning the total variation into two components: Between Groups represents variation of the group means around the overall mean, and Within Groups represents variation of the individual scores around their group means, sometimes referred to as error variation. The F test is the ratio between these two measures of variation. In Table 4, the F value is large and the signiﬁcance level is less than 0.05, indicating that at least one of the groups differs from the others. The null hypothesis can be rejected. The small signiﬁcance value of Linearity and the high signiﬁcance value of Deviation from Linearity would indicate that a linear relationship exists between the variables, if the categories of the independent variable were ordered. When the null hypothesis is false, the difference between means is due to error variance plus the treatment effect. How much this effect depends on the independent variable (especially when this is nominal) is tested with the Crosstabs procedure (the word ‘‘depends’’ has no implications for a causal relationship). The continuous numeric RoCE data are converted to a discrete number of categories, each one containing approximately the same

190

NICOLE AVDELIDOU-FISCHER

Table 4. Mean Difference between Multiple Groups.

0.520

4.943

-2.179

4.937 8

8 5.165

5.010

1.998

-0.531

0.237 6

Matrix

7 7.278

5.910

5.203

0.249

-1.884

Multidivisional 14 4.188

1.815

6.597

0.879

-0.235

Product-team

8.450

5.142

0.496

-1.138

9 8.706 ANOVA table

2 0 -2

F

Sig.

Combined

570.300

4

142.575

5.176

0.002

Linearity

484.484

1

484.484

17.588

0.000

Deviation from Linearity

85.816

3

28.605

1.038

0.385

Within groups

1212.007

44

27.546

Total

1782.307

48

Between groups

Sum of squares

4

PRODUCT-TEAM

11 -1.161

Geographic

MULTIDIVISIONAL

Functional

Mean line-chart

MATRIX

Skewness Kurtosis 10

GEOGRAPHIC

N Mean Median Std. dev

FUNCTIONAL

Structures

df

Mean square

number of cases. The value of 1(low) was assigned to cases below the 25th percentile, 2(medium) to cases between the 25th and 50th percentile, 3(high) to cases between the 50th and 75th percentile, and 4(very high) to cases above the 75th percentile. Selecting Structures as the row variable, and levels of RoCE as the column variable, the crosstabulation shows the frequency of each RoCE level at each structure. If each structure inﬂuences ﬁnancial performance similarly, the pattern of RoCE should be similar across structures. The following null and alternative hypotheses are generated: The null hypothesis states that the two variables are independent; the alternative hypothesis states that the two variables are not independent (they are related). From the clustered bar chart, it is evident that each pattern of bars is not constant across structures; this indicates that the two variables are not independent. The Functional group has RoCE incidents only in the ‘‘low’’ and ‘‘medium’’ categories. The Multidivisional is the only group with a complete set of colored bars and, hence, frequencies in every category. An interesting observation is that the Matrix and Multidivisional appear to have the same frequencies for the ‘‘high’’ and ‘‘very high’’ levels (Table 5). From the crosstabulation and the bar chart, it is impossible to tell whether these differences are real or due to chance variation. For that reason, the chi-square statistic is used to measure the discrepancy between the observed cell counts and what would be expected if the rows and columns were unrelated. Both signiﬁcance values are below 0.05, indicating that there is some relationship between the two variables. The Likelihood ratio results

The Relationship between Organizational Structures and Performance

Table 5.

Structures Count Observed

Functional Expected % within Observed

Geographic Expected % within Observed

Matrix

Expected % within Observed

MultiExpected divisional Productteam

Clustered bar chart: Count

7

% within Observed Expected % within

low

191

Crosstabulation and Relation.

Asymp.Sig Low Medium High V. High Total Value df Chi-square test (2-sided) 7 4 0 0 11 34.074* 12 0.001 2.7 2.7 2.9 2.7 11.0 Pearson chi-sqre 40.291 12 0.000 63.6% 36.4% 0.0% 0.0% 100.0% Likelihood ratio 49 0 1 6 1 8 N of valid cases 2.0 2.0 2.1 2.0 8.0 *20 cells (100.0%) expected count less than 5 0.0% 12.5% 75.0% 12.5% 100.0% Directional measures Value Appr. sig. 0 2 2 3 7 Symmetric 0.296 0.013 Lambda 1.7 1.7 1.9 1.7 7.0 RoCE dependent 0.361 0.007 0.0% 28.6% 28.6% 42.9% 100.0% Uncertainty Symmetric 0.277 0.000 5 4 2 3 14 coefficient RoCE dependent 0.297 0.000 3.4 3.4 3.7 3.4 14.0 Value Appr. sig. 35.7% 28.6% 14.3% 21.4% 100.0% Symmetric measures 0.834 0.001 0 1 3 5 9 Phi 0.481 0.001 2.2 2.2 2.4 2.2 9.0 Cramer's V 0.640 0.001 0.0% 11.1% 33.3% 55.6% 100.0% Contingency coefficient

medium

high

very high

6 5 4 3 2 1 0

Functional

Geographic

Matrix

Multidivisional

Product-team

(a goodness-of-ﬁt statistic similar to Pearson’s chi-square) are more reliable because 100% of the cells have expected values less than 5. The null hypothesis can be rejected. The direction of the relationship is studied with Lambda and Uncertainty coefﬁcient measures of association. The low signiﬁcance values point to a relationship between the two variables; but the average values for both test statistics indicate that the relationship is an intermediate one. The Uncertainty coefﬁcient result moreover indicates that knowledge of structural type reduces error in predicting values of RoCE by almost 30%. The values of the three Symmetric measures are medium to high, and all signiﬁcance values are 0.001. It is safe to say that the relationship is not due to chance and that it is a medium to strong one. Testing conﬁrmed the expectations deﬁned in HA. The ﬁndings obtained from the chi-square statistics, are cross-validated with the ones from the powerful ANOVA. But ANOVA alone would not have offered enough information on each group distribution and insight to draw some more

192

NICOLE AVDELIDOU-FISCHER

conclusions regarding the very ﬁrst syllogism that led to testing for difference – how some structural attributes can affect a company’s ﬁnancial performance negatively while others can affect it positively. The Product-Team and Matrix ﬂatter arrangement, which Lawler (cited in Wheeler, 2000) called a ‘‘high-involvement structure,’’ allows for quicker reactions to the business environment, involving individuals throughout the organization in the information ﬂow and decision-making capacities. This approach is consistent with democratic principles and, thus, may be especially appropriate in this increasingly democratic world (Wheeler, 2000). The ﬂexibility that characterizes Product-Team and Matrix structures supports better coordination and makes maximum use of assets, reducing costs. These factors were expected to affect a company’s ﬁnancial performance positively, and apparently they do, comparing the means in Table 4 and also the RoCE patterns in the clustered bar chart – none of the two types has RoCE incidents in the ‘‘low’’ categories. HB. For the independent-samples t-test the Functional and the Multidivisional group were deﬁned by specifying two values; cases with any other values were excluded from the analysis. The following null and alternative hypotheses are generated: The null hypothesis states that the two means are equal; The alternative hypothesis states that the two means are not equal. The Standard Error of Mean (in the Group Statistics table in Table 6), demonstrates how much the value of the mean may vary from sample to sample taken from the same distribution. Since the independent-samples t-test compares the two group means, it is useful to know what the mean values are. In this case, the mean value of the Multidivisional group appears Table 6.

Structures N Multidivisional 14 Functional 11

Equal variances Assumed Not assumed

Mean Difference between Two Groups.

Group statistics Mean Std. deviation Std. error mean 4.188 6.597 1.763 1.490 -1.161 4.943

Fisher and Yates t table 0.050 0.020 Levels for two-tailed Levels for one-tailed 0.025 0.010 df 23 2.069 2.500

Independent samples test 95% Confidence interval of Levene's test for t-Test for equality of means equality of variances the difference Sig. Mean Std. error F Sig t df (2-tailed) difference difference Lower Upper 1.779 0.195 2.237 23 0.035 5.349 2.391 0.402 10.295 2.317 22.970 0.030 5.349 2.309 0.573 10.125

The Relationship between Organizational Structures and Performance

193

to be higher than the mean value of the Functional group. But is this difference signiﬁcant? In the Levene test for homogeneity of variances the signiﬁcance value is high (more than 0.05), so results that assume equal variances for both groups will be used. The t value is signiﬁcant –equals the value between those shown in the Fisher and Yates’ table. The low signiﬁcance value for the t-test (0.035) and the fact that the conﬁdence interval for the mean difference does not contain zero indicate that there is a signiﬁcant difference between the two group means. The null hypothesis can be rejected. The effect of Multidivisional structure on ﬁnancial performance was positive and signiﬁcant as predicted. This ﬁnding complies with Armour and Teece’s research from 1978 (cited in Hill & Jones, 2001, p. 399), suggesting support for HB. HC. The independent samples t-test is again used to compare the means for the Product-Team and the Matrix groups. The following null and alternative hypotheses are implied: The null hypothesis states that the two means are equal; The alternative hypothesis states that the two means are not equal. After running the test, the mean value of the Product-team group appears higher than the mean value of the Matrix group (Table 7). Is this difference signiﬁcant? The Levene’s test for equality shows homogeneous variances for both groups, as the F value is small and the signiﬁcance level is higher than 0.05. To be signiﬁcant, the t obtained from the data, should have been equal to or larger than the value shown in the Fisher and Yates’ table – here the t is much lower. The signiﬁcance value for the t-test is high, and the conﬁdence interval for the mean difference contains zero, hence, it cannot be Table 7.

Structures Product-team Matrix

Equal variances Assumed Not assumed

N 9 7

Mean Difference between Two Groups.

Group statistics Mean Std. deviation Std. error mean 8.706 5.142 1.714 1.967 7.279 5.203

Fisher and Yates t Table 0.100 0.050 Levels for two-tailed Levels for one-tailed 0.050 0.025 df 14 1.761 2.145

Independent samples test 95% Confidence interval of Levene's test for t-Test for equality of means equality of variances the difference Sig. Mean Std. error F Sig t df (2-tailed) difference difference Lower Upper 0.039 0.847 0.548 14 0.592 1.427 2.605 -4.159 7.013 0.547 12.966 0.594 1.427 2.609 -4.210 7.064

194

NICOLE AVDELIDOU-FISCHER

concluded that there is a signiﬁcant difference between the two group means. The null hypothesis cannot be rejected. Consistent with HA and HB, it was expected that structural type would continue to inﬂuence ﬁnancial performance. But remember that ProductTeam and Matrix structures both enjoy the advantages of ﬂexibility and ﬂatness, while operating in teams – the ﬁrst one in permanent crossfunctional teams, the second one in temporary project teams. Apparently, this small variation is not imperative for shaping ﬁnancial performance. This suggests lack of support for HC. Structures and Return per Employee HD. The Median test is used to assess the RpE difference between the groups, if any. The following null and alternative hypotheses are generated: The null hypothesis states that there will be no observed differences between group medians; The alternative hypothesis states that there is a signiﬁcant difference, at least between two of the medians. The Percentiles Statistic Table (Table 8) gives a numerical representation of the shape of the distribution. Across all 49 subjects, the RpE median is a score below 28. The null hypothesis for the median test is that this particular value is a good approximation of centre for each of the groups. To test this hypothesis, each group is divided into two subgroups: those whose scores fall at or below the median, and those whose scores are above it. The result is a two-way Frequencies Table with two rows and 5 columns (the grouping variable). If the groups do not differ, for each group, half of the scores should be above and half of the scores below this overall median. For the Geographic, Matrix, and Multidivisional groups, the median does what the null hypothesis says it should do: it divides them into two equal subgroups. But for the rest, the null hypothesis does not provide a good approximation of centre. The Median test now determines whether this pattern represents group differences that are signiﬁcant. From the two-way Frequencies Table, a chisquare statistic is calculated to test row and column independence. In fact, the median test is a chi-square test of independence between group membership and the proportion of cases above and below the median. For each cell, the distance between the observed and expected counts is squared, then divided by the expected value. Finally, these quantities are summed across

The Relationship between Organizational Structures and Performance

Median Difference between Multiple Groups.

Table 8.

Structures Functional

N Mean Median 11 50,555 19,504

Geographic Matrix Multidivisional

Percentiles statistic table

Std. dev Skewn. Kurtosis 112,482

1.692

2.122

N=49

25th 11,105

27,734

78,026

2

3

4

8 68,983

27,181

123,675

2.698

7.439

RpE Structures

7 47,764

44,289

43,971

1.267

1.297

14 65,694

20,551

80,672

1.685

3.210

9 64,142

37,770

67,835

2.468

6.564

Product-team

195

Frequencies table

50th(Median)

Median test table N Median

Structures Functional Geographic Matrix Multidivisonal Product-Team

75th

Chi-square

> Median

2

4

4

7

7

df

<= Median

9

4

3

7

2

Asymp. sig.

49 27,734 7.358 4 .118

Note: Grouping Variable: Structural Categories 6 cells (0.0%) have expected frequencies less than 5. The minimum expected cell frequency is 3.4.

all cells. In the Median test table, the chi-square value is small and the p-value is larger than 0.05. None of the groups differs signiﬁcantly from the others. The null hypothesis cannot be rejected. According to the views of Chandler (1966), Mintzberg (1983), and Hill and Jones (2001), organizational structures provide the framework for a social-operational-control system, inﬂuencing greatly individual and group behavior for pursuing a common outcome. Each structural type formally divides, groups, and coordinates job tasks in a different way. Thinking of how Brumbach (1988, pp. 387–402) embraces both behavior and outcome in a deﬁnition of human resource performance, one could expect that each structural type would have a different impact on it. Testing has revealed that the effect of structural types on human resource performance could not be proven signiﬁcant; this suggests lack of support for HD. Although the data do not appear to conform to the requirement (signiﬁcantly different) for testing the next hypothesis, an analysis will be conducted – simply for directly addressing the particular suggestion. It should also be noted that when no statistically signiﬁcant differences have been found, it does not mean that no differences have been found. But in terms of statistical testing for signiﬁcance, a difference that fails to meet the criteria of ‘‘signiﬁcance,’’ is deemed to be ‘‘no difference’’ (Balestrini & Agathangelou, 2001). HE. The point-biserial analysis is used when the one variable is continuous and the other genuinely dichotomous, in this case a company can have a Geographic structure or not. For the Dummy coding each company with a Geographic structure is assigned a 1, and all the rest a 0. The

196

NICOLE AVDELIDOU-FISCHER

Table 9.

Point-Biserial Correction.

Y1 Y0 Sy n1 n0

= mean of the continuous variable for the Geographic type = mean of the continuous variable for the non-Geographic = standard deviation of all the continuous scores = number of companies with a Geographic structure = number of companies with a non-Geographic structure n = total number of companies

68,983.554 58,230.818 87,784.616 8 41 49

0.122 rpb= 0.046 0.139

Fisher and Yates r table

rpb = [Y1−Y0 /S y] n1n 0 /n(n −1)

Levels for two-tailed df 45

0.100

0.050 Two-tailed 0.100

0.2428 0.2875

df 50

0.050

0.2306 0.2732

following null and alternative hypotheses are implied: The null hypothesis states that the two variables are not positively correlated; The alternative hypothesis states that the two variables are positively correlated. A positive value for ‘r’ suggests that a higher score on one variable is associated with a higher score on the other. There is a 0.046 correlation between structure and RpE (Table 9). The point-biserial correlation coefﬁcient is tested for statistical signiﬁcance using the same table of critical values as the Pearson r. The degrees of freedom are n2 (Grimm, 1993, p. 468), in this case 47. Testing the obtained correlation using a nondirectional test with an alpha level of 0.05, it is evident that the absolute value of the rpb is much lower than the possible critical values. The null hypothesis cannot be rejected. Results seem to conﬂict with Berenbeim’s (1983) study that exposed a relationship between Geographic structure characteristics and an advantageous approach toward delegation, which fosters motivation. No positive correlation between Geographic structure and human resource performance could be found, suggesting lack of support for HE. Even though the effect of structural type on the distribution of RoCE is obvious, and according to the results the two variables are not independent, the same cannot be presumed for the effect of structural type on RpE. It is believed that this two-edged result exactly reﬂects reality. There is a relation between organizational structures and the performance of FORTUNE 500 companies, but structural type is not the sole or universal performance determinant.

The Relationship between Organizational Structures and Performance

197

CONCLUSIONS Discussion Organizations vary in performance, just like they vary considerably in structural attributes. Prior studies have identiﬁed some correlation between performance and appropriate structure (Bowman, Singh, Useem, & Bhadury, 1999; Donaldson, 1987); this research reveals a somewhat divergent result of the two tested performance dimensions. Although results prove that structural types are positively related to ﬁnancial performance, calculated as RoCE, no relation between structural types and employee performance, calculated as RpE, could be proved. Regarding the second proposition, it appears that other inﬂuencers have a stronger impact on FORTUNE 500 human resource performance. Rephrasing this, consider Armstrong’s theory (2001, pp. 143–153) that performance is not just a product of the role/situation individuals are in (the organizational context and direction or inﬂuence exercised from above or elsewhere in the organization), but also of their own skills, competences, individual differences, orientation, attitudes, and personality. Borgatti (2001) proclaims that this is attributable to the fact that, when human beings interact with each other over a long period of time, they develop a social structure that is only partly based on the formal organizational structure. Formal organizations are not as formal as they may seem, even if they are bureaucracies. This agrees with the Processual view that organizations are partnerships of individuals, each of whom brings their own personal objectives and cognitive biases to the organization. In an MIT study, Cross, Nohria, and Parker (2002) support the view that the ‘‘boundaryless’’ organization is increasingly common, but the general health and ‘‘connectivity’’ of work groups can have a signiﬁcant impact on strategy execution and organizational effectiveness. Working with a group of FORTUNE 500 companies and government agencies, the authors assessed more than 40 networks in 23 organizations. They discovered in all cases that informal networks provide strategic and operational beneﬁts by enabling members to collaborate effectively. However, managers must overcome harmful myths about how networks operate to better support them. An interesting criticism was made by Ghoshal and Bartlett (1999), who believe that the more managers understood about these new demands, the more they realized that their historic focus on adjusting the strategystructure linkage was only part of a much more profound metaphoric change their companies had to make. Indeed, as companies adopted new

198

NICOLE AVDELIDOU-FISCHER

structures and sophisticated processes were described, most were stumbling on a similar constraint: their managers were simply unable to adapt to the demands being placed on them by the new organization. One manager captured the situation accurately when he suggested that his company ‘‘was trying to implement third-generation strategies through second-generation organizations run by ﬁrst-generation managers.’’ Evolutionists are perfectly prepared to accept the difﬁculties of organizational change, but their conclusion is more brutal. If managers will not change, change the managers; if the organization does not adapt, sell it and buy another one (Whittington, 1993, p. 132). This brings the discussion back to the objective of understanding why organizations have the structure that they do. Chandler’s dictum ‘‘structure follows strategy’’ dominates as an answer. Findings reinforce the literature that describes the positive effect of Alfred Sloan’s innovative Multidivisional structure on the ﬁnancial performance of large American corporations. But one should not assume, like Taylor, Fayol, Weber and other theorists did, that there is a single best way to structure an organization. And the advice that structure must follow strategy may be basically sound, but it is also simplistic. Sometimes, it will be ineffective or even dangerous to impose Multidivisional structures upon diversiﬁed businesses. Systemic theorists stress that the link between strategy and structure may not follow the precise form of the textbooks (Whittington, 1993, p. 121). Also, Borgatti (2001) argues that ‘‘institutionalization’’ should be treated with caution. Under conditions of uncertainty, organizations imitate others that appear to be successful; one reason why this happens is the fear of litigation or, simply, blame. This can cause whole industries to adopt similar structural features. Thus, although results associate Product-team and Matrix structures with high ﬁnancial performance, one ought to pledge that there is no one best way to organize. What is important is that there be a ﬁt between the organization’s key processes, objectives and strategies, employees and other assets, social system, its technology, and the requirements of its environment (Kotter, 1995). Furthermore, in view of the fact that results do associate Product-team and Matrix structures with high ﬁnancial performance, it could be inferred that these companies achieved a higher ﬁt than the others. Different designs are better for different situations. This raises the issue of the importance of the subject of reorganization. External events having an impact on life-cycle stages (Jawahar & McLaughlin, 2001), changes in management trends or personnel, declining economic performance (Candea, 1976), or growth without structural adjustment can only lead to economic inefﬁciency

The Relationship between Organizational Structures and Performance

199

(Chandler, 1966). Twenty-ﬁrst century organizations will require the capacity to keep up with an intense pace of change as well as the capacity to reshape themselves continually. No longer can organizational houses be built on the obsolete assumption that they will last for 100 years (Leavitt & Lipman-Blumen, 1995, p. 116). To sum up, the accumulated examples simply do not provide enough evidence that organizational structures have an inﬂuence on the human resource performance of FORTUNE 500 companies, calculated as RpE, but prove that there is an inﬂuence on the ﬁnancial performance of FORTUNE 500 companies, calculated as RoCE. The inﬂuence of organizational structures on the performance of FORTUNE 500 companies – taken together – may not be imperative, but is deﬁnitely noteworthy.

Implications for Managers First of all, the results have implications with respect to how managers view organizational design. Asked to describe their structure, managers usually respond by drawing an organization chart, seeing lines of command but failing to see lines of communication. This is better expressed in the example from Ghoshal and Bartlett (1999), where the ‘‘ﬁrst-generation managers were trying to implement third-generation strategies through secondgeneration organizations.’’ The fundamental necessity for a company’s strategic managers is to understand the twofold role of organizational structures; ﬁrst, to coordinate activities effectively for implementing a strategy that increases competitive advantage and, second, to motivate employees and provide them with the incentives to achieve superior efﬁciency, quality, innovation, or customer responsiveness (Hill & Jones, 2001, p. 383). Moreover, managers must accept change as constant and organizational restructuring as resulting from it. They should be able to recognize when there is a gap between desired and actual company performance and see this as the ﬁrst step of the change process. Because threats and opportunities vary with life-cycle stages, organizations are likely to have different needs at each stage, also in terms of human resources. Speciﬁc stakeholders are likely to become more or less important as an organization evolves from one stage to the next (Jawahar & McLaughlin, 2001, pp. 397–415). Managers should understand the impact of organizational restructuring on different stakeholder groups and the reasons for resistance to change (Armstrong, 2001, pp. 265–274), and acknowledge that strategic goals are more likely to be attained with the help and guidance in analysis and diagnosis of a specialist

200

NICOLE AVDELIDOU-FISCHER

HR function. In this research, HR respondents were the most proﬁcient on the subject. The informal social structure has as much to do with the way the organization runs as does the formal structure (Borgatti, 2001). Managers should bear in mind that organizational structures are just one of the seven cultural web elements, and their informality and ﬂuidity do not mean casualness or sloppiness when it comes to goals, standards, or clarity of direction and purpose. Rather they translate into responsiveness and readiness to absorb and assimilate rapid changes while maintaining ﬁnancial and operational cohesion (Johnston, 2000, pp. 4–22; Timmons, 1999, p. 245). The most important managerial tool available for assessing patterns of relationships in informal networks is called social network analysis. It is in management’s interest to both make critical patterns of interaction visible and understand the social structure, so that they can predict how different stakeholders will react to things and, if needed, to inﬂuence them so as to achieve strategic alignment. Managers should incorporate strategies into each staff-member’s job, building alignment around the organization’s structure. As concluded in the above discussion, strategic ﬁt turns organizational visions into reality (Bradford, 2002, p. 14). Because of the size of most multinational organizations (the minimum number of employees in the sample was 2,000), one ﬁnal suggestion for managers is to expand their universe with respect to diversity of culture, identity, and thought and focus more on the mission of their work and less on institution building (Wheeler, 2000).

Limitations of the Research To put ﬁndings in a proper perspective, some limitations of the research must be considered. First of all, the results are based on a modest sample size of 50 ﬁrms. Although the sample size was calculated from a reliable formula, and the two general rules on sampling have been applied (samples smaller than 30 should be treated cautiously; subsamples smaller than 5 create problems when using tests like chi-Square), the size is based on a 90% level of conﬁdence. No evidence was found that the period of six months over which the data were collected might have affected the outcomes, but it should be recognized that they may be inﬂuenced by the economic climate in which the ﬁnancial reports were generated. Qualitative data were collected from employees embodying a diversity of departments and job-titles, ranging from the

The Relationship between Organizational Structures and Performance

201

simple Representative of the Investor Relations Department, to the Vice President of Corporate Communications. Although respondents were, on average, familiar with structures in their organizations, many could not guarantee specialized knowledge.

Recommendations for Future Research In the case of the sample, scholars could use a 99% level of conﬁdence in future studies; the size would be 99.75, rounded up to 100, which would make conclusions more powerful. Some tests assume that data are from normally distributed populations. While this assumption is not too important with large samples, it is important with small subsamples, especially with unequal sizes (Motulsky & Searle, 1998, p. 47). Stratiﬁed random sampling would ensure that every structural group gets a better representation and gives higher precision with the same number of cases. The drawback would be that a complete frame of the 500 structures would be needed. The people who responded to the survey were, on average, knowledgeable about structures in their organizations. However, for a future replication of this research, HR specialists would be recommended, as they proved to be the most reliable contacts. Firm size was included, because it is a standard variable in strategy research and it captures, in a crude way, the level of ﬁrm resources (Rust, Moorman, & Dickson, 2002). Drucker stated that the FORTUNE 500 are downsizing (Harris, 1993, p. 116). Several examples show how many companies eradicate or add value to the organization through workforce rationalization. This is an avenue for future research that can provide interesting insights into the dynamics of downsizing on organizational structures and performance. According to Styles (1998, p. 13), business performance is an important and often used dependent variable in the marketing literature, as well as in the strategic management literature. Walker and Ruekert (1987, pp. 15–33) suggest that the relevance and importance of performance dimensions vary across stakeholder groups (investors, employees, or customers), and each one could be examined with a different performance dimension. Two dimensions were used that reﬂect on the efﬁciency with which resources are utilized in generating proﬁt; together, they reveal a more complete portrait of effects. Further research might examine the way in which structural types inﬂuence different performance indicators. Addressing the bureaucratic costs issue, a future study could test the theory of Hill and Jones (2001, p. 462) about complex structures having

202

NICOLE AVDELIDOU-FISCHER

higher bureaucratic costs. Consistent with this theory, structural types will be ordered from simple to complex and bureaucratic costs from low to high, facilitating analyses for linear relationships.

ACKNOWLEDGMENTS I would like to thank Dr. Andrew Adcroft from the School of Management, University of Surrey, for his very useful comments, and also the FORTUNE 500 companies that participated in the study.

REFERENCES Agathangelou, T. (2001). Electronic lecture 1: Business decisions, DMB schedule. University of Surrey UK: School of Management LearnWare. Armstrong, M. (2001). A handbook of human resource management practice (8th ed.). London: Kogan Page. Backstro¨m, T., & Ladan, P. (2002). Knowledge matrix – a transformative organization, program for organisational development and learning. Stockholm: National Institute for Working Life. Balestrini, P., & Agathangelou, T. (2001). K-Cluster 4: Comparing means, DMB schedule. University of Surrey, UK: School of Management Knowledge Repository. BBC-News (2001a). US companies warn on profits. Retrieved September 18, 2001, from http:// news.bbc.co.uk/1/hi/business/1551433.stm BBC-News (2001b). US officially enters recession. Retrieved November 26, 2001, from http:// news.bbc.co.uk/1/hi/business/1677224.stm BBC-News (2002). Specter of recession haunts US economy. Retrieved October 25, 2002, from http://news.bbc.co.uk/1/hi/business/2361565.stm Bello, D., & Gilliland, D. (1997). The effects of output controls, process controls, and ﬂexibility on export channel performance. Journal of Marketing, 61, 22–38. Berenbeim, R. (1983). Operating foreign subsidiaries: How independent can they be? Report No. 836. The Conference Board, New York. Bethel, J., & Liebeskind, J. (1993). The effects of ownership structure on corporate restructuring [Special Issue]. Strategic Management Journal, 14, 15–31. Bolton, L. (2002). Types of organizational structures. Victoria: Swinburne University of Technology. Borgatti, S. (2001). Structures. Chestnut Hill, MA: Department of Organizational Studies Boston College. Bower, M. (2003). Organization: Helping people pull together. Retrieved June 3, 2003, from http://www.mckinseyquarterly.com/article_page.asp?ar=1308&L2=18&L3=29&srid=27&gp=0 Bowman, E., Singh, H., Useem, M., & Bhadury, R. (1999). When does restructuring improve economic performance? California Management Review, 41, 33–56.

The Relationship between Organizational Structures and Performance

203

Boyle, M. (2002). Rapid Growth in Rough Times; FORTUNE’s 100 fastest growers prove there is still life in the fast lane [online], Available from: http://www. fortune.com/indexw.jhtml?channel=artcol.jhtml&doc_id=208994 [Accessed 2 September 2002] Bradford, R. (2002). All aboard! (or not?): Align staff with your organization’s strategy. Nursing Management, 33, 14. Brumbach, G. (1988). Some ideas, issues, and predictions about performance management. Public Personnel Management, 17, 387–402. Campbell, J. (1990). Modeling the performance prediction problem in industrial and organizational psychology. Handbook of industrial and organizational psychology. Cambridge, MA: Blackwell. Candea, D. (1976). Organization structure; an approach to revision. Working Paper No. 407. Alfred P. Sloan School of Management, MIT, Cambridge, MA. Capon, N., Farley, J., Lehmann, D., & Hulbert, J. (1992). Proﬁles of product innovators among large U.S. manufacturers. Management Science, 38, 157–169. CGEY. (2000). Measuring the future: The value creation index. Cambridge, MA: Cap Gemini Ernst & Young Centre For Business Innovation. Chandler, A. (1966). Strategy and structure. Chapters in the history of the industrial enterprise. Massachusetts Institute of Technology, MA: The MIT Press. Chaudron, D. (2001). The balanced scorecard & performance improvement. San Diego, CA: Organized Change Consultancy. Cross, R., Nohria, N., & Parker, A. (2002). Six myths about informal networks – and how to overcome them. MIT Sloan Management Review, 43, 67–75. Cummings, J. (2004). Work groups, structural diversity, and knowledge sharing in a global organization. Management Science, 50, 352–364. Decker, P. (1996). Structures. University of Houston, Clear Lake: Healthcare Administration Program. Donaldson, L. (1987). Strategy and structural adjustment to regain ﬁt and performance: In defence of contingency theory. Journal of Management Studies, 24, 1–24. Duncan, W. (1981). Organizational Behavior (2nd ed.). Boston, MA: Houghton Mifﬂin. DuPont Heritage. (2002e). Pierre S. du Pont: In depth. Wilmington, DE: E.I. du Pont de Nemours and Company. Dyson, J. (2001). Accounting for non-accounting students (5th ed.). UK: The Financial Times Pitman Publishing. Eidleman, G. (1995). Z-scores. A guide to failure prediction. New York: N.Y. State Society of Certiﬁed Public Accountants. Ernst & Young. (1997). Measuring for the future. Cambridge, MA: The Ernst & Young Centre For Business Innovation. Fortune (2003). The FORTUNE 500; America’s largest corporations. Retrieved January 14, 2003, from http://www.fortune.com/lists/F500/index.html FT. (1999). Mastering strategy: New notions of organizational ‘‘fit’’. UK: Saı¨ d Business School and Financial Times. Galbraith, J. (1973). Designing complex organizations. Reading, MA: Addison-Wesley. Gertner, J. (2002). Wants to change the world. MONEY Magazine, 31, 118–124. Ghoshal, S., & Bartlett, C. (1999). The individualised corporation. A fundamentally new approach to management. London: William Heinemann. Grimm, L. (1993). Statistical applications for the behavioral sciences. New York: Wiley.

204

NICOLE AVDELIDOU-FISCHER

Guyon, J. (1997). Why is the world’s most proﬁtable company turning itself inside out? Royal Dutch/Shell looked at the future. Fortune Magazine, 137, 120–125. Harﬁeld, T. (1998). Strategic management and Michael Porter: A post-modern reading. Retrieved March 20, 2003, from http://www.mngt.waikato.ac.nz/depts/sml/journal/special/ harﬁeld.htm Harrington, A. (2003). Honey, I shrunk the proﬁts. Fortune Magazine, 147, 197–199. Harris, G. (1993). The post-capitalist executive: An interview with Peter F. Drucker. Harvard Business Review, 71, 116–118. Harvey, C. (2003). Hypertextual finance glossary. Retrieved May 13, 2003, from http:// www.duke.edu/charvey/Classes/wpg/glossary.htm Hill, C., & Jones, G. (2001). Strategic management theory, an integrated approach (5th ed.). Boston, MA: Houghton Mifﬂin Company. Hodgetts, R. (1991). Organizational behavior: Theory and practice. New York: Maxwell Macmillan International Pub. Group. Horgan, C., & Garnick, D. (2000). Structuring behavioral health care: Carved Out or integrated? Compensation & Benefits Management, 16, 39–47. Houston, M., Walker, B., Hutt, M., & Reingen, P. (2001). Cross-unit competition for a market charter: The enduring inﬂuence of structure. Journal of Marketing, 65, 19–35. Howell, D. (2002). Kmart restructures operations (cover story). DSN Retailing Today, 41, 1–2. Jawahar, I., & McLaughlin, G. (2001). Toward a descriptive stakeholder theory: An organizational life cycle approach. Academy of Management Review, 26, 397–415. Johnson, G., & Scholes, K. (1999). Exploring corporate strategy, text and cases (5th ed.). UK: Prentice Hall. Johnston, M. (2000). Delegation and organizational structure in small businesses. Group & Organization Management, 25, 4–22. Kahn, J. (1998). The world’s most admired companies. Fortune Magazine, 138, 206–226. Kerr, E. (2001). Electronic lecture 3: Analysing Company Accounts, Financial Management Schedule, School of Management Knowledge Repository, University of Surrey, UK. Kotter, J. (1995). A 20% solution: Using rapid re-design to build tomorrow’s organization today. New York: Wiley. KPMG. (1999). Unlocking shareholder value: The keys to success. Global Research Report on Mergers and Acquisitions, KPMG International, UK. Leavitt, H., & Lipman-Blumen, J. (1995). Hot groups. Harvard Business Review, 73, 113–116. Levinson, H. (1992). Thinking ahead: Appraisal of what performance? Harvard Business Review, 70, 70–76. McKenna, E. (2000). Business psychology and organisational behaviour, a student’s handbook (3rd ed.). UK: Psychology Press. Mintzberg, H. (1983). Power in and around organizations. Englewood Cliffs, NJ: Prentice-Hall. Mintzberg, H. (1994). The rise and fall of strategic planning. UK: Prentice Hall International. Moore, J., & Reichert, A. (1983). An analysis of the ﬁnancial management techniques currently employed by large U.S. corporations. Journal of Business Finance & Accounting, 10, 623–645. Motulsky, H., & Searle, P. (1998). InStat guide to choosing and interpreting statistical tests [Computer Software]. San Diego, CA: GraphPad Software Inc. Nayak, P., & Ketteringham, J. (1997). 3M’s Post-it Notes: A managed or accidental innovation? In: R. Katz (Ed.), The human side of managing technological innovation. A collection of readings (pp. 367–377). New York: Oxford University Press.

The Relationship between Organizational Structures and Performance

205

OECD. (1987D). Structure and organisation of multinational enterprises. France: Organisation for Economic Co-operation and Development. Papandreou, A. (1952). Some basic problems in the theory of the ﬁrm In: B. Haley (Ed.), A survey of contemporary economics (Vol. 2). Homewood: Richard D. Irwin. Pendlebury, M., & Groves, R. (2001). Company accounts: Analysis, interpretation and understanding (5th ed.). UK: Thomson Learning. Porter, M. (1980). Competitive strategy. Techniques for analyzing industries and competitors. New York: The Free Press. Ramesh, J. (2002). The global economy: A ‘‘synchronized’’ recession. Retrieved September 19, 2002, from http://www.worldpress.org/asia/0202indiatoday.htm RBR. (2001R). K-Cluster 10: Implementation of strategy, corporate strategy schedule. University of Surrey, UK: School of Management Knowledge Repository. Robbins, S., & Coulter, M. (2002). Controlling for organizational performance, PPT slide report prepared for the book: Management (7th ed.). USA: Prentice Hall. Rust, R., Moorman, C., & Dickson, P. (2002). Getting return on quality: Revenue expansion, cost reduction, or both? Journal of Marketing, 66, 7–25. Sablynski, C. (2002). Foundations of organization structure. Sacramento: College of Business Administration: Department of Organizational Behavior & Environment, California State University. CA. Sellers, P., & Barlyn, S. (1996). PepsiCo’s new generation. Fortune Magazine, 133, 110–118. Sellers, P., Mulcahy, A., & Notebaert, D. (2002). The new breed. Fortune Magazine, 146, 66–73. Stein, N. (2000). Measuring people power. Fortune Magazine, 142, 186. Styles, C. (1998). Export performance measures in Australia and the United Kingdom. Journal of International Marketing, 6, 13–19. Timmons, J. A. (1999). New venture creation. Entrepreneurship for the 21st century (5th ed.). Singapore: McGraw-Hill International Editions. Walker, K., & Bain, C. (1989). Sales volume forecasting: A comparison of management, statistical, and combined approaches. Journal of Management Accounting Research, 1, 119–136. Walker, O., & Ruekert, R. (1987). Marketing’s role in the implementation of business strategies. Journal of Marketing, 51, 15–33. Wheeler, W. (2000). Emerging organizational theory and the youth development organization. Applied Developmental Science, 4, 47–55. Whittington, R. (1993). What is strategy – and does it matter? New York: Routledge. Yunker, P. (1983). A survey study of subsidiary autonomy performance evaluation and transfer pricing in multinational corporations. Columbia Journal of World Business, 17, 51–64.

206

NICOLE AVDELIDOU-FISCHER

APPENDIX A. HANDLING EXTREME SCORES

Plots can aid in the validation of the assumptions of normality and linearity. Plots are also useful for detecting outliers, unusual observations, and influential cases. Simple boxplots show the median (heavy black line), outliers , and extreme values within a category. The box represents the interquartile range which contains the 50% of values. Whiskers are vertical lines extending from the box and ending in horizontal lines at the largest and smallest observed values that are not statistical outliers (SPSS, 2001). Q-Q Plots, plot the quantiles of a variable's distribution against the quantiles of any of a number of test distributions. They are generally used to determine whether the distribution of a variable matches a given distribution. If the selected variable matches the test distribution, the points cluster around a straight line. RoCE N=50

-30

RpE N=50

-20

0

-10

10

20

-1000000 1000000 3000000

5000000

7000000

9000000

Outliers: Cases with values between 1.5 and 3 box lengths from the upper or lower edge of the box. Extreme scores: Cases with values more that 3 box lengths from the upper or lower edge of the box. Normal Q-Q Plot of RoCE

1

-1

1

Expected Normal

0

0

-1

-2

Observed Value -20

-10

Normal Q-Q Plot of RpE

2

0

10

Expected Normal

2

Observed Value

-2 -2000000

0

2000000 4000000 6000000 8000000

Researchers have adopted the attitude that one extreme score is probably a fluke; therefore, they will consider discarding it (Grimm, 1993, p. 85). The boxplots have revealed the two most extreme scores; an interesting observation is that they both fall under the Functional category. After eliminating them the summary information is compared to the values before, in the tables below. N

Mean

Std. deviation

Skewness Std. error Kurtosis Std. error

Before Return on capital employed

50

3.866

7.185

-.973

.337

3.601

.662

Return per employee

50

251,849.139

1,359,453.952

7.027

.337

49.567

.662

After Return on capital employed

49

4.418

6.094

-.012

.340

1.013

.668

Return per employee

49

59,986.367

87,784.616

2.003

.340

3.935

.668

In general, a skewness value greater than 1 indicates a distribution that differs significantly from a normal, symmetric distribution. As a rough guide, a skewness more than twice its standard error is taken to indicate a departure from symmetry. The ratio of skewness and of kurtosis to its standard error is used as a test of normality; that is, normality can be rejected if the ratio is less than -2 or greater than +2. According to the results, RoCE does not differ significantly from a normal distribution, but RpE does. The histograms below display this graphically.

364

317

223

270

15.30

18.40

12.10

5.80

9.00

2.70

-0.50

-3.60

-6.80

0

-9.90

2

129

4

176

6

35

8

Return per Employee

82

10

22 20 18 16 14 12 10 8 6 4 2 0

-12

Return on Capital Employed

12

-59

14

PART IV: MERGERS AND ACQUISITIONS

This page intentionally left blank

DIFFICULTIES IN VALUE CREATION: TELECOM NEW ZEALAND’S ACQUISITION OF AAPT LTD Alireza Tourani-Rad and Zoltan Toth ABSTRACT We provide an overview of the Australian and New Zealand telecommunications markets through Telecom Corporation New Zealand’s (TCNZ) acquisition of AAPT Ltd in 2000, which amounted to more than NZ$2 billion. A few years later and after writing off approximately NZ$1 billion, TCNZ is considering a sell-off at a considerable loss. We discuss the strategic reasons behind the acquisition and explain how smaller telcos are struggling to compete with the incumbent telecom in Australia. We further conduct an event study to assess the impact of the acquisition on both TCNZ’s and AAPT’s share prices and look at some of the postacquisition issues.

INTRODUCTION In 2000, Telecom Corporation New Zealand (TCNZ), the largest listed company in New Zealand, acquired AAPT Ltd. AAPT was Australia’s third Value Creation in Multinational Enterprise International Finance Review, Volume 7, 209–227 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07009-9

209

210

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

largest telecommunication ﬁrm offering voice (local, national, and international), data, mobile, internet (broadband and dial up) and pay TV services nationwide to residential and corporate customers. At the time, TCNZ Chief Executive, Roderick Deane, said, ‘‘Telecom’s international strategies include growing our trans-Tasman business. AAPT provides Telecom with facilities-based participation in the fast-growing Australian market. Increasing our shareholding should enable both Telecom and AAPT to achieve key strategic goals sooner than if they were to proceed independently’’ (AAPT Ltd’s Australia Stock Exchange announcement, 1999f). By mid-2001, Telecom had to reduce the value of AAPT in its books by approximately NZ$1 billion. At the time of writing this paper, late 2005, Telecom was considering a potential sale of AAPT at a considerable loss. A number studies have examined the wealth impact of shareholders involved in cross-border mergers and acquisitions. For example, Doukas and Travlos (1988), Fatemi and Furtado (1988) and Markides and Ittner (1994) investigated international takeovers by the US companies where on average they found positive but rather negligible abnormal returns. Kang (1993) studied Japanese acquisitions of US ﬁrms. He found that these transactions created signiﬁcant wealth gains for shareholders of bidding ﬁrms. Corhay and Tourani-Rad (2000) found weak evidence that crossborder acquisitions by Dutch ﬁrms are wealth-creating activities. The available empirical research on cross-border transactions, as a whole, indicates little or no signiﬁcant gains for the bidding ﬁrms’ shareholders, and as Markides and Ittner comment, ‘‘The stock market, therefore, is not very enthusiastic about international acquisitions. It does not view them, however, as good news.’’ Whether international takeover creates value is far from settled and our aim in this paper is to add to this by examining in detail the case of a single and major international takeover. In the rest of the paper, we provide an overview of the telecommunication markets and the prevalent regulatory issues in Australia and New Zealand. We present the strategic reasons behind the acquisition and explain how smaller telcos are struggling to compete with the incumbent telecom in Australia. We further conduct an event study to assess the impact of the acquisition on share prices of both companies and look at some of the ﬁnancial issues involved. Finally, we report on the post-acquisition situation, summarising some of the difﬁculties faced by TCNZ in creating value in this transaction.

Difficulties in Value Creation

211

OVERVIEW OF THE NEW ZEALAND TELECOMMUNICATIONS MARKET TCNZ is New Zealand’s largest telecommunications company. It provides a full range of ﬁxed line voice and data, mobile, and internet services. TCNZ was separated from New Zealand Post Ofﬁce in 1987, as the ﬁrst step of sweeping market deregulation policies implemented in New Zealand in the late 1980s by the Labour Government. The company was subsequently sold to two US-based telecommunication companies, Bell Atlantic and Ameritech, for NZ$4,250 million. At the same time it got listed on New York, Australian, and New Zealand stock exchanges. A few years later they both sold their shares on the stock exchange and the shares have ultimately been acquired by a variety of institutional investors.1 The second largest company in the ﬁxed line market in New Zealand is TelstraClear, the Australian incumbent telco’s New Zealand subsidiary. Telstra entered the New Zealand market in 1996 and attained its present form by acquiring Saturn Communications and, later, Clear Communications Ltd. The company offers voice, broadband internet, and wireless services in New Zealand. By 2005, TelstraClear had acquired 35–40% of the corporate voice calling market (Harpur, 2006). TelstraClear has a 6000 km long ﬁbre optic network connecting major cities and regional centres. The network is linked to Telstra’s Australian network, providing an international extension to the New Zealand core network and allowing the company to handle increasing network trafﬁc and in the future, VoiP and a range of multimedia services. The company also has a wireless network which is utilised to provide broadband access to remote areas when DSL services are unavailable. Regulations in New Zealand required Telecom to open up its services to other players in the market. Telecom provides the vast majority of its ﬁxed line services for wholesale customers at a price determined by the regulator. Telecom operates New Zealand’s major national telephone network. More than 99% of the public switching network lines are connected to digital exchanges. In voice services, Telecom maintains its 95% share of the local access market and around 80% of the overall voice market (Hot Telecoms, 2005). The remainder of the market lies mainly with TelstraClear. Although there are now a number of other companies providing ﬁxed line service (local and international); their market share remains relatively small. Interconnection agreements have been signed between Telecom and the

212

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

other players and in some instances the companies have agreements with each other. Though there has been some development in competition for the smaller companies and margins improved through regulated deals regarding wholesale activity and interconnection, local loop unbundling still has not been permitted.2 In 2005, TelstraClear gave up its plans to build up an alternative local and nationwide network to compete with Telecom in the ﬁxed line business. Instead TelstraClear would remain a niche player concentrating on its residential markets in New Zealand’s second and third largest cities (Wellington and Christchurch), with its government and business customers looking after their trans-Tasman relations. This strategic decision strengthened further Telecom’s dominant position in the shrinking voice market. Telecom also owns a data services network (called Digital Services Transmission Network – DSTN) which enables it to offer data transmission services like private ofﬁce network, datalink, and ATM services. In 2004, Telecom announced its plans to implement a new Next Generation Network (NGN). The total cost of this investment is $1.4 billion and it is expected to be completed by 2012. The project aims to multiply the capacity of the core network by implementing an IP-based data and voice transmission infrastructure (Alcatel company announcement). In the wireless market, TCNZ and Vodafone are the key players. Vodafone’s predecessor, Bell South, established New Zealands’s ﬁrst mobile service to compete with Telecom. Vodafone acquired the company in 1998 and made large investments in its network and distribution system. As a result, the company managed to catch up with and then exceed Telecom’s market share and established its position as market leader. Through its wireless network, the company offers voice, data and value-added mobile services. More recently, Vodafone has lost a signiﬁcant portion of its market share due to Telecom’s successful 3G campaign and an attractive offer of low price texting (Harpur, 2006). Vodafone is operating its own 2G fully digital GSM network on 900 and 1800 Mhz, which covers 97% of the population. In 2005, they ofﬁcially launched a W-CDMA-based 3G network. The system enables them to offer video calling and initially covers major cities reaching around one third of the population. In the internet dial-up and broadband market, Telecom’s dominance is most important. Other ISPs (such as Ihug, Orcon, Maxnet, and CallPus) are resellers of Telecom’s services. New Zealand has been behind in the broadband uptake in an international comparison and the New Zealand government has developed plans to remedy this situation.

Difficulties in Value Creation

213

OVERVIEW OF THE AUSTRALIAN TELECOMMUNICATIONS MARKET The Australian telecommunications market has been dominated by the two large companies Telstra and Optus, sharing 75% of the market revenue, and a number of second-tier companies operating in a highly regulated and competitive environment. There are over 700 companies in the second-tier market; however, the distribution is uneven: the top 10 players share over 90% of all revenues (Australia Telecoms Industry – Overview and Statistics (2005/2006)). The market leader is Telstra Corporation owned by the Australian Government (51%). The company has a dominant position in the ﬁxed line telephone services, and is a market leader in mobile as well as in data (dialup and broadband) services. Optus is a wholly owned subsidiary of SingTel and is the second largest telco in the Australian market [Hot Telecoms – Country Proﬁle Australia (2004)]. Optus was established in 1991 and started to offer ﬁxed line services mainly by reselling Telstra’s national long distance and international telephone services. In 1992, the company won the second carrier license, enabling it to offer unrestricted local, national long distance, international, and mobile telephone services. AAPT was established in 1991, when the ﬁnancial and data services division was separated from the Australian Associated Press, Australia’s leading national news, information and communications group. The company started operating as the ﬁrst competitor to Telstra in long distance voice and data services. Throughout the years, AAPT has always been a signiﬁcant player in the market, challenging the big telcos, leading the innovation and generating growth by natural market share increase and an active acquisition policy which has made it attractive for potential buyers (for an overview of milestones in AAPT’s history, please refer to the appendix). Competition in Australia The regulatory environment in Australia has not been favourable to smaller telecommunication companies. Although the Australian government’s aim to encourage competition that drives new services and lower prices has shown some improvement over the years, in reality this is still not a classic free market where competitors have equal opportunities. In addition to the regulatory environment, the other restricting element for small telcos are the lack of infrastructure and the high level building up

214

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

their own network. Consequently, they do not have many viable options except to resell Telstra’s various products. Competition in the telecommunications sector began in the early 1990s, when the government decided to open up the market to other carriers and give permission for them to resell Telstra’s services, mainly long distance and international calling services. In 1992, Optus obtained the second carrier license and started to offer ﬁxed line and mobile services. The next major player in the mobile scene was Vodafone, who entered the market in 1993 offering services over its own mobile network. In 1997, the Australian Government introduced a new regulatory regime which allowed telecommunication companies to openly compete within all segments of the telecommunication industry. As a result, over 700 communication service providers have been established since then. Most of these companies are using Telstra’s wholesale services and competing in the retail market. Though this number is quite impressive in size, Telstra has managed to retain its dominant position across the market in the telecom sector. By June 2005, Telstra’s estimated retail market share was 73% in local calls, 62% in domestic long distance minutes, 51% in international long distance minutes, 45% in mobile services, 28% in dial-up internet services, 41% in broadband services and 60% in subscription television services (Telstra Australia Ltd, 2005).

ACQUISITION OF AAPT TCNZ’s Motivations From the early years on, Telecom has been planning a strategic Australian presence. Trans-Tasman telecommunication businesses have more than 100 years of history. Since international telecommunication lines were set up, both major companies (and their predecessors) have operated in both countries. It has become even more important since the two countries have developed into highly integrated economies and an increasing number of companies are operating on both sides of the Tasman. Driven by the liberalisation process during the late 1980s and early 1990s, both companies felt that it would be a perfect opportunity to increase their presence on the other side and attempt to take a chunk of the market which was still dominated by the domestic incumbent company.

Difficulties in Value Creation

215

TCNZ’s decision to purchase AAPT had several reasons. One of the most important aspects is the strategic considerations. Telecom’s presence in the Australian market was very limited. In order to service the increasing demand for Trans-Tasman connections from both sides, Telecom needed a strong and sizeable company to establish its presence. It is also very important to note that Telecom had some infrastructure capacity that had not been entirely utilised. A new investment with reasonable incremental volumes could ensure better utilisation of the infrastructure. Another aspect of this issue is that prior to the acquisition in the mid 1990s, Telstra had shown an increasing interest in the New Zealand market. The company had various plans to invest in and build up an alternative network which could have been developed into a wide range of competing services. This has been considered as a major risk by Telecom’s management and they felt that through the purchase of AAPT, this risk could be diversiﬁed. They also believed that the Australian market would provide a steady revenue growth that may offset the eventual reduction in domestic revenues. Some market analysts suggested that TCNZ’s management had plans to move the company’s head ofﬁce across the Tasman and create an Australian based fully integrated telecommunication company operating in both countries. This plan, however, has never been ofﬁcially conﬁrmed. Market analysts also speculated that the acquisition would have been the ﬁrst step of a mega merger by Telecom and Optus. Telecom and Optus would join forces to attack Telstra’s position in the Australian market. This rumour has never been conﬁrmed by either companies and subsequent events in the market have never indicated that this merger was ever going to take place. A possible explanation for this rumour may have been the fact that Optus had attempted to purchase AAPT in 1998 prior to Telecom but it was rejected by the ACCC (Australian Competition and Consumer Commission). Australian regulators believed that the merger would restrict market competition even more and narrow down the number of sizeable telecom companies to only two in the Australian telecommunications market. A classic element of merger motivations is often the bad performance of a company’s management. AAPT’s has shown similar characteristics as well; the market (and Telecom’s management too) believed that there was an opportunity for value creation by replacing the existing management and utilising the expertise of Telecom executives. What was not taken into account was that Telecom played a different role in its domestic market, being the incumbent company trying to save its own interests and maintain its high revenue from the New Zealand market. AAPT was in a drastically

216

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

different position, being the attacker in the Australian market. This requires attitude, management style and strategy of different kinds to meet the challenges that this situation implies. What both companies had to realise soon was that it is very difﬁcult to compete outside their respective domestic markets. In the other country, both telcos are just one of the other competitors and even though their investment capacity may be signiﬁcantly larger than that of any other player, they still have been hampered by strict regulations and, consequently, the high risk of building up their own network. Relevant Events Prior to TCNZ’s Takeover Prior to Telecom’s plans to acquire AAPT, Optus had attempted to take over the company. This takeover attempt failed due to the rejection of the ACCC. Originally, Optus offered AU$5 per share, which was strongly rejected by AAPT’s board and management. The chairman of the board, Mr Lee Casey, said at that time: ‘‘This opportunistic $5.00 bid substantially undervalues the shares of the Company. It ignores the Company’s growth potential and does not include any premium for control. We believe that AAPT represents a unique opportunity in the Australian market for CWO (Optus) and also for a number of international telecommunication companies with whom we are in discussion. The proposed CWO offer does not reﬂect AAPT’s strategic value’’ (AAPT Ltd’s Australia Stock Exchange announcement, 1999b). Shortly after the release of Optus’ offer, the AAPT management released a report from investment bankers, Grant Samuel & Associates, which valued AAPT Ltd in the range of $6.04–$7.01 per share, which is 20.8–40.2% above Optus’ conditional offer of $5 per share. Grant Samuel stated that ‘‘the valuation range of $6.04 to $7.01 per share reﬂects an expectation of substantial growth in AAPT revenues and earnings over the next few years ﬂowing from investment by AAPT in infrastructure and changes in the regulatory environment. In the context of explosive growth in internet usage and data trafﬁc, AAPT’s infrastructure and established customer base provide a platform for dramatic growth’’ (AAPT Ltd’s Australia Stock Exchange announcement, 1999d). Grant Samuel also stated that its valuation range may be conservative, adding: ‘‘In a competitive bidding situation, it is conceivable that industry participants seeking to establish a position or consolidate an existing position in the Australian telecommunications industry could offer higher prices to acquire the strategic number three competitor.’’

Difficulties in Value Creation

217

The above statements underline the fact that at the time, analysts gave credit to the high growth forecast and, as mentioned above, most of it was based on the expectation of further easing up the regulatory regime. Another possible reason for the original bid looking reasonably low in comparison is that we were in the middle of the dot.com boom, when technological companies showed an unprecedented growth in stock prices and analysts worldwide would tend to put a signiﬁcant emphasis on growth factor in their valuation models. Chronology of the Takeover 18 May 1999

31 May 1999

4 June 1999

15 September 1999

TCNZ Australia Investments Pty Ltd (wholly owned subsidiary of TCNZ) became a substantial shareholder in AAPT Ltd, with a relevant interest in the issued share capital of 29,588,430 ordinary shares (9.9%). Telecom paid $5.70 per share. Telecom indicated that ‘‘given Optus’ bid for AAPT, Telecom felt it prudent to take a minority position in order to preserve its strategic options’’ (‘‘Telecom tipped to expand into Australia,’’ 1999). Optus withdrew its offer to buy AAPT’s stake following ACCC’s refusal of clearance for the AAPT bid. TCNZ Australia Investments Pty Ltd and associates increased its relevant interest in AAPT Ltd from 29,588,430 ordinary shares (9.9%) to 59,149,859 ordinary shares (19.79%). TCNZ announced that its wholly owned subsidiary TCNZ Australia Investments Pty Ltd would make an A$5.10 per share takeover offer for AAPT Ltd. The offer price of A$5.10 per share valued AAPT at approximately A$1.5 billion. The offer was for all of the AAPT shares that TCNZ Australia did not already own. The share price of AAPT was $4.85 on 30 August, prior to the run-up of the last two weeks, which appears to have been prompted by speculation about Telecom’s intentions. The offer therefore represents a 5% premium.

218

16 September 1999

21 October 1999

25 October 1999

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

The Board of Directors of AAPT Ltd considered the proposed takeover bid by TCNZ Australia Investments Pty Ltd and believed that the bid of $5.10 was inadequate and undervalued the company. Accordingly, the Board recommended that shareholders reject the proposed offer. In making this recommendation, the Directors highlighted the following: Grant Samuel & Associates Pty Ltd assessed the value of the shares to be signiﬁcantly above TCNZ’s offer and accordingly has formed a view that the TCNZ offer is not reasonable. Grant Samuel considered the fair value of shares to be between $6.17 and $7.14 a share. TCNZ was prepared to pay $5.70 for 9.9% of AAPT in May 1999 and after subsequent discussions with AAPT, TCNZ made a nonbinding proposal that in addition to purchasing Optus’ stake at $5.20 a share, TCNZ would subscribe to a $300 million placement of AAPT’s shares at $6.00 a share. The TCNZ bid did not include a realistic premium for control of AAPT. The closing price of AAPT shares on ASX on the day prior to TCNZ’s announcement, 14 September 1999, was $5.15. The TCNZ offer reﬂected inadequate value for the signiﬁcant growth potential attributable to investments and acquisitions that the company continues to undertake, in particular the expansion of AAPT’s network. TCNZ Australia Investments Pty Ltd increased its relevant interest in AAPT Ltd from 59,154,859 ordinary shares (19.79%) to 65,298,490 ordinary shares (21.80%). TCNZ Australia Investments Pty Ltd increased its relevant interest in AAPT Ltd from 66,298,490 ordinary shares (21.8%) to 71,878,903 ordinary shares (24.03%).

Difficulties in Value Creation

1 November 1999

4 November 1999

4 November 1999

5 November 1999

219

TCNZ Australia Investments Pty Ltd increased its relevant interest in AAPT Ltd from 71,878,903 ordinary shares (24.03%) to 82,856,027 ordinary shares (27.71%). Optus announced that in the absence of a higher offer, it accepted TCNZ’s takeover offer of $5.10 for all its shares in AAPT Ltd. The company’s 10.6% stake was purchased earlier in 1999 for $4.85 per share in the course of the company’s proposed takeover of AAPT. Norman Gillespie, Deputy Chief Executive of Optus, said the company accepted the $5.10 per share being offered by TCNZ. ‘‘Our earlier interest in AAPT was due to its perceived strength in the SME and Government markets. When our bid for AAPT was blocked by the ACCC we concentrated our efforts on winning business in these markets, which has met with considerable recent success. We no longer require a strategic stake in AAPT, and selling to TCNZ gives us the opportunity to exit the shareholding at a proﬁt’’ (AAPT Ltd’s Australia Stock Exchange announcement, 1999a). TCNZ Australia Investments Pty Ltd increased its relevant interest in AAPT Ltd from 82,866,027 ordinary shares (27.71%) to 97,964,251 ordinary shares (32.76%). John Fairfax Holdings Limited said that AAPIS, in which it holds a 44% stake, had accepted, through its subsidiary, AAP Communications Holdings Pty Ltd, TCNZ’s offer of $5.10 per share pursuant to TCNZ’s bid for AAPT Ltd. Mr Fred Hilmer, Chief Executive Ofﬁcer (CEO) of Fairfax, said, ‘‘This decision is consistent with our objective to realise shareholder value by relinquishing our position in non-core assets’’ (AAPT Ltd’s Australia Stock Exchange announcement, 1999c).

220

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

6 November 1999

10 November 1999

TCNZ Australia Investments Pty Ltd closed its offer for the shares in AAPT Ltd after successfully obtaining close to 80% of the shares. AAPT’s Chairman, Mr Lee Casey, said: ‘‘We are obviously a little surprised that a number of our institutional shareholders sold at the low price offered. Telecom has bought very well and we now look forward to working with them to continue to deliver the exceptional returns that AAPT has delivered to date (AAPT Ltd’s Australia Stock Exchange announcement, 1999e). AAPT announced that Telecom’s CEO, Theresa Gattung, and Chief Financial Ofﬁcer, Mr Jeff White, joined the AAPT Board of Directors.

Empirical Study on Announcement Effects We applied the standard event study method and examined the abnormal returns (ARs) to shareholders of both TCNZ and AAPT for the period around several important announcements. We highlighted four key announcements in the process of the takeover and calculated ARs for both companies’ share performance. The AR is by deﬁnition a return on the share on a particular day above a benchmark of what investors require on that day. The benchmark we applied was ASX All Ordinaries index for Australia and NZALL for New Zealand. All the announcements and their corresponding dates are obtained from the Australian Stock Exchange. The parameters of the market model, applying the regression model, are calculated using 130 observations, starting 150 days before the 20 days before and after the considered events. Our ﬁndings are reported in Table 1. Most international studies indicate signiﬁcant positive returns for target shareholders and small negative or zero returns for bidders (Bruner, 2002). Studies conducted on New Zealand (Firth, 1997) and Australian (Walter, 1984) markets further conﬁrm these ﬁndings. The market reactions to Telecom’s acquisition show similar results in relation to bidder ARs. The daily ARs for the bidder company for all announcements are negative, ranging from 0.38% to 1.57%, though none of them are statistically signiﬁcant. These results support the international ﬁndings, namely, the markets do not value cross-border acquisitions from the bidder company’s perspective.

Difficulties in Value Creation

Table 1.

221

Event Study Results. AAPT

Events 18 May 1999 TCNZ purchase of 9.9% of AAPT shares 4 June 1999 TCNZ increased its stake to 19.79% 15 September 1999 TCNZ made a public offer to purchase the remaining shares of AAPT 5 November 1999 TCNZ becomes majority shareholder

TCNZ

AR (%)

t-Test

AR (%)

t-Test

0.52

0.23

1.57

1.45

3.58

1.58

0.38

0.35

0.40

0.18

1.14

1.06

0.68

0.30

1.03

0.95

Note: The estimation period is from 150 to 21 days prior to the announcements. t-stat is the heteroskedasticity-adjusted t-statistics for the average abnormal return (AR). It tests the null hypothesis that the average AR ¼ 0.

The results in relation to the target company also support earlier studies, though not that strongly. Speciﬁcally, AAPT’s ARs are distributed from 0.40% to 3.58%. On average, it indicates modest positive ARs which are in line with other studies but the magnitude is reasonably smaller than worldwide benchmarks. Post-Acquisition Performance Shortly after the takeover, Telecom established a working group comprising the Telecom and AAPT management, which identiﬁed the following areas of synergies: joint management and purchasing of international trafﬁc and bandwidth; operating cost savings and reducing capital expenditure through the coordination of domestic networks; joint marketing to major corporate businesses operating trans-Tasman by promoting a seamless service; alignment of respective mobile capabilities, including trans-Tasman roaming, handsets, operations, maintenance, and marketing; skills transfer between management teams, particularly in local, transTasman virtual private networks, xDSL, copper, and mobiles; and co-operation of internet and e-commerce initiatives and the creation of a content alliance.

222

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

As a result, during 1999 and 2000, Telecom consolidated many of AAPT’s activities and ultimately ended up fully integrating the two companies. As a consequence, AAPT, being active in reorganising itself, lagged behind Optus, which had taken the lead in competitive activities in the telecommunications market. At the same time, AAPT has replaced many of its top executives, including the CEO [AAPT Company Analyses – Paul Budde Communications Pty Ltd (2006)]. Telecom’s plan was to establish a genuine trans-Tasman business by bringing together the internet and mobile businesses of Telecom and AAPT under an integrated management team. Telecom’s new Chief Executive, Theresa Gattung, said that ‘‘expansion in Australia is a core strategy for the group, and we’re putting in place a strong platform for expansion through AAPT and TCNZ Australia, and the focused Internet and Mobile businesses’’ (AAPT company announcement, 2000, November 27). During 2000, AAPT and TCNZ achieved success in working together to secure telecommunications business on both sides of the Tasman for New Zealand corporates. AAPT won major contracts which represented total annual revenues of A$6 million. The contracts include the New Zealandbased forest, building, and paper products company Carter Holt Harvey, food group Heinz Wattie’s Australasia and Owens Group, a major transport and logistics business. Telecom was now able to deliver a full range of telecommunications services with the facility to consolidate their telecommunications arrangements. Previously, these companies had arrangements with Telstra and other third-party providers for their Australian operation. Despite some encouraging initial customer acquisition, the restructuring and change in management seemed to have had its backside. It became apparent that even the combined New Zealand and Australian market would not be sufﬁcient for future growth. In the meantime, other telecommunication providers have taken the opportunity to drive competition in Australia. Also, some of TCNZ’s problems seemed to have been exported to AAPT’s market as well, namely, a ﬂat national market with rapidly shrinking revenues and proﬁt margins, issues with the competition in the mobile area, non-compatible network technologies, etc. It has been proven that combining an incumbent player in one market with an attacker in the other is extremely challenging. By mid-2001, Telecom had to reduce the value of AAPT in its books by approximately NZ$1 billion. Another important factor has been that Telstra has always maintained wholesale prices at such a level that it was not motivating for other competing telcos to invest in building up a complete network that would have been able to operate independently. At the same time, wholesale prices were

Difficulties in Value Creation

223

not placed at a sufﬁciently low level in order to maintain an acceptable rate of return for smaller players in the market. TCNZ had realised the importance of a strategic change and laid down a new direction for AAPT, starting from late 2002, that involved a disintegration of the two companies and acknowledging of the fact that they operate in two different markets in completely different market positions. The new strategy aimed to focus on various niche markets where by concentrating forces AAPT could successfully challenge the more signiﬁcant players. The shift in strategy also meant that AAPT discontinued operating as a universal service provider and abandoned its broader residential market and began cross-selling between its various divisions such as ﬁxed line, mobile, internet, etc. The company set up speciﬁc sales teams focusing on selling packaged products to speciﬁc markets. The new initiatives in repositioning the AAPT brand was accompanied by an extensive cost-cutting project as well. While some of the resale margins continued to be remarkably low, bundling of the various products seemed to be a successful strategy. This strategy was supported by a $20 million advertising campaign. Despite the efforts of AAPT to maintain its proﬁtability and compete in the market, it has become evident that the current industry structure makes it enormously difﬁcult for second-tier companies to distinguish themselves and build up a genuinely different product range. Companies such as AAPT had to struggle with the fact that they can only sell what Telstra offers them at a price suitable for Telstra. Whilst several big players have ceased to exist, it was quite an achievement for AAPT just to be around and still be able to compete with other major telcos. In 2005, the Australian government decided to divest the remaining stake in Telstra (51%). This move has had a number of implications for other players in the telecommunications market, including AAPT. Before selling out Telstra’s shares, the government wanted to strengthen the company’s position in the market. This has resulted in an extreme situation where Telstra went as far as bringing down its retail price below the wholesale price (Paul Budde Communications Pty Ltd., 2006a). The immediate impact on AAPT was that in October 2005, Telecom decided that again it needs to restructure the operation by looking at options that include a potential sale of the business.

CONCLUSIONS This paper provides an overview of a cross-border takeover in the telecommunication sector, the TCNZ acquiring of AAPT in Australia. We analyse

224

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

the telecommunication markets in New Zealand and Australia including their regulatory environment. The paper highlights the main strategic decisions that led to the takeover and analyses the post-merger performance. It was found that the regulatory environment is very restrictive for market players competing with the incumbent telecommunication company, Telstra, and it had a signiﬁcant impact on the company’s performance subsequent to the takeover. We also conduct an event study and ﬁnd that market reactions are rather weak, which is in line with empirical ﬁndings of earlier cross-border studies.

NOTES 1. TCNZ’s website: www.telecom.co.nz/content/0,2502,200633-1548,00.html 2. In May 2006, the New Zealand government announced a decision to go ahead with unbundling from 2007. However, the legal framework and the exact details of this matter are yet to be determined.

REFERENCES AAPT Company Analyses – Paul Budde Communications Pty Ltd, 2006. AAPT company announcement (2000, November 27). TELECOM-AAPT integration progressed. Retrieved from www.aapt.com.au/news/content.asp?n=95 AAPT Company overview – Paul Budde Communications Pty Ltd, 2006. AAPT Ltd’s Australia Stock Exchange announcement. (1999a, November 4). CWO’s ann: Optus to accept Telecom NZ offer for AAPT. Retrieved March 1, 2006, from www.asx.com.au/asx/statistics/showSignalgDetail.do?issuerId=3829&announcementId=576180 AAPT Ltd’s Australia Stock Exchange announcement. (1999b, April 19). Directors reject CWO takeover bid. Retrieved March 1, 2006, from www.asx.com.au/asx/statistics/ showSignalgDetail.do?issuerId=3829&announcementId=576030 AAPT Ltd’s Australia Stock Exchange announcement. (1999c, November 5). FXJ’s ann: Statement by John Fairfax Holdings Limited on AAPT. Retrieved March 1, 2006, from www.asx.com.au/asx/statistics/showSignalgDetail.do?issuerId=3829&announcementId= 576187 AAPT Ltd’s Australia Stock Exchange announcement. (1999d, May 20). Grant Samuel values AAPT at $6.04 to $7.01 per share. Retrieved March 1, 2006, from www.asx.com.au/asx/ statistics/showSignalgDetail.do?issuerId=3829&announcementId=576057 AAPT Ltd’s Australia Stock Exchange announcement. (1999e, November 8). Media release: Telecom offer for AAPT closes. Retrieved March 1, 2006, from www.asx.com.au/asx/ statistics/showSignalgDetail.do?issuerId=3829&announcementId=576193 AAPT Ltd’s Australia Stock Exchange announcement. (1999f, September 15). Telecom New Zealand announces takeover offer for AAPT. Retrieved March 1, 2006, from www.asx.com.au/ asx/statistics/showSignalgDetail.do?issuerId=3829&announcementId=576106

Difficulties in Value Creation

225

Alcatel company announcement. Retrieved August 30, 2005, from www.home.alcatel.com/vpr/ vpr.nsf/DateKey/30082005uk Australian Competition and Consumer Commission. http://www.accc.gov.au/content/ index.phtml/itemId/475702 (Website accessed: March 2006) Australia Telecoms Industry – Overview and Statistics (2005/2006), Paul Budde Communications Pty Ltd. Bruner, R. (2002). Does M&A pay? A survey of evidence for the decision-maker. Journal of Applied Finance, Spring/Summer, 7–27. Company announcements on Australian Stock Exchange Website (www.asx.com.au). Corhay, A., & Tourani-Rad, A. (2000). International acquisitions and shareholder wealth: Evidence from the Netherlands. International Review of Financial Analysis, 9, 163–174. Doukas, J., & Travlos, N. G. (1988). The effect of corporate multinationalism on shareholders’ wealth: Evidence from international acquisitions. Journal of Finance, 43, 1161–1175. Fatemi, A., & Furtado, E. P. (1988). An empirical investigation of the wealth effects of foreign acquisitions. In: S. Kouri & A. Ghosh (Eds), Recent developments in international banking and finance, (Vol. 2, pp. 363–379). Lexington, MA: Lexington Books, D.C. Health. Firth, M. (1997). Takeovers in New Zealand: Motives, stockholder returns, and executive share ownership. Pacific-Basin Finance Journal, 5, 419–440. Harpur, P. (2006). Telecoms overview, statistics and analyses in New Zealand. Bucketty, Australia: Paul Budde Communications Pty Ltd. Hot Telecoms. (2004, November). Country proﬁle, Australia. Retrieved June 7, 2006, from www.hottelecoms.com Hot Telecoms. (2005, November). Country proﬁle, New Zealand. Retrieved June 7, 2006, from www.hottelecoms.com Kang, J. K. (1993). The international market for corporate control. Journal of Financial Economics, 33, 345–371. Markides, C. C., & Ittner, C. D. (1994). Shareholder beneﬁts from corporate international diversiﬁcation: Evidence from US international acquisitions. Journal of International Business Studies, 25, 343–366. Telstra Australia Ltd. (2005). Telstra Australia Ltd annual reports 2004–2005. ‘‘Telecom tipped to expand into Australia.’’ (1999, May 19). The Dominion (2nd ed., p. 20). Walter, T. S. (1984). Australian takeovers: Capital market efﬁciency and shareholder risk and return. Australian Journal of Management, 9, 63–95.

226

ALIREZA TOURANI-RAD AND ZOLTAN TOTH

APPENDIX: MILESTONES IN AAPT’S HISTORY [AAPT Company Overview – Paul Budde Communications Pty Ltd (2006)] 1991

1992 1993 1994 1995 1996

1997

1998

AAPT was established. Granted a Service Provider’s Class Licence in accordance with the Telecommunications Act. Negotiated an interconnect agreement with Telstra. Installed switches into selected regional centres. Established submarine cable links to USA. Offered mobile services through Vodafone’s network. Acquired 50% of ISP Connect Internet Solutions. Acquired 40% of Cellular One Communications. The company distributed its mobile services through over 120 dealer stores. It was also Vodafone Network’s largest independent service provider, accounting for almost 20% of Vodafone Network’s customer base. Awarded a carrier licence. Listed on ASX, raising $76 million for the public issue of new shares and becoming the ﬁrst Australian carrier to be listed on the ASX. Acquired the remaining 60% of Cellular One Communications. Purchased the outstanding 80.5% of corpTEL, a long distance voice provider. AAPT and Optus announced an agreement between the two companies for the provision of mobile communication services across the Optus network for AAPT’s customer base. Acquisition of 100% of Ozphone Pty Ltd, which gave AAPT control of its CDMA mobile network spectrum, which combined with the existing AAPT spectrum enabled coverage of over 50% of the Australian population. Acquisition of 100% of AT&T Easylink Services Australia Pty Limited, an e-commerce solutions provider. The acquired entity was re-branded and in conjunction with connect.com led the provision of e-commerce solutions to the high-growth business sector.

Difficulties in Value Creation

1999

2000

2001 2002

2003 2004

227

TCNZ acquired 80% of the company’s shares. Acquired the remaining shares in ISP Connect Internet Solutions. Fibre optic local-loop networks completed in the CBDs of the ﬁve major cities in Australia. TCNZ acquired the remaining 20% of the company. Voice and Data operations were split into two divisions. De-listed from ASX. Acquired national high-bandwidth network capacity from Optus. Formed a joint venture with America Online and Seven Network, AOL7. AAPT’s international telecommunications business transferred to TCNZ’s International division. AAPT Cellular One signed a three-year CDMA wholesale agreement with Telstra. AAPT Mobile launched its AAPT Mobile Wholesale Service. Divested its stake in AOL7.

This page intentionally left blank

ANALYSIS OF GLOBAL COMPETITORS’ REACTION TO MEGA MERGER ANNOUNCEMENTS BY AN MNC: THE CASE OF THE CITICORP– TRAVELERS MERGER Isaac Otchere and Suhadi Mustopo ABSTRACT We investigate global competitors’ reaction to the Citicorp–Travelers mega merger announcement and find that global competitors, especially banks in Europe and the US, reacted positively to the Citicorp and Travelers’ merger announcement. The uncertainties created by the investigations into the merger proposal had significant impact on the competitors’ stock price. The announcement that the merger had been consummated also elicited a significantly positive reaction from the rivals following the resolution of uncertainties emanating from the regulatory challenges. The positive reaction by competitors suggests that the merger was a wealthcreating event for the large firms in the financial services industry. The expected benefits outweighed any competitive effects resulting from the merger. The competitors’ reaction was, however, not homogenous. Value Creation in Multinational Enterprise International Finance Review, Volume 7, 229–254 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07010-5

229

230

ISAAC OTCHERE AND SUHADI MUSTOPO

Our cross-sectional analysis shows that the abnormal returns earned by the competitors were higher the larger the competitor. In addition, the abnormal returns were greater for the US rivals. That the global competitors reacted positively to the Citicorp–Travelers mega merger announcement is consistent with our assertion that the merger had ramifications that go beyond regulatory concerns in the US.

1. INTRODUCTION Deregulation and globalization have offered multinationals the opportunity to exploit economies of scale and scope on a global scale and to consolidate their activities to remain competitive. The ﬁnancial services industry, particularly in the US and Europe, has experienced a signiﬁcant rise in mergers and acquisitions activities. The value of bank mergers in the US increased dramatically from USD 160.4 million in 1980 to USD 37,049 million in 1997, while the number of bank mergers increased by 215% during the same period (Becher, 2000). In the insurance industry, known for its high-cost distribution and lack of price competition, the consolidation process has lagged behind that of the banking sector. Nevertheless, over the 1985–1997 period, mergers and acquisitions in the insurance industry accounted for 18.9% of the ﬁnancial intermediaries’ merger activities in the US and 18.6% for European domestic deals (Berger, Demsetz, & Strahan, 2000). Despite the increase in mergers and acquisition activities in these two sectors, there had not been any particularly noteworthy cross-sector merger between the banking sector and the insurance sector during the mergers and acquisition boom of the 1990s until 6 April 1998 when Citicorp and Travelers Group announced their USD 70 billion merger to form ‘Citigroup’, a merger that created the world’s largest bancainsurance company. Never before has the ﬁnancial services sector had a merger of such scale between two leading companies in their respective ﬁelds. The stock market reacted positively to this merger, with Citicorp shares rising by 26% from $150 to $187 and Travelers’ by 16% from $70 to $81, thus creating about $25 billion in value on the announcement date. At the time of the merger, Citicorp was the second largest US bank (after Chase Manhattan); it had operations in nearly 100 countries, and was the largest issuer of credit cards.1 Its merger partner, Travelers Group, was an integrated ﬁnancial services ﬁrm that owned leading Wall Street security ﬁrm Salomon Smith Barney and had a substantial market share in the life insurance business. The combined entity comprises three sub-sectors,

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

231

namely, commercial banking, investment banking, and insurance with assets in excess of USD 700 billion, net revenues of USD 50 billion, and market capitalization of USD 140 billion after the merger. Given the sheer size of the merger and the global presence of the combined entity, the merger had the potential to spark a new wave of mergers and acquisition not only in the US ﬁnancial services sector but also in the global ﬁnancial services industry, as competitors attempt to respond to the changing industry dynamics. Thus, this merger has the potential to generate positive valuation effects across the industry. The merger could also affect the competitiveness of not only the US rivals but also of the European and other foreign competitors. If investors expect the merger to create enormous competitive pressures on existing competitors, then the rivals would react negatively to the merger announcement. In this study, we analyse both the intra- and inter-industry effects of the Citicorp–Travelers mega merger announcement on global competitors in the banking and insurance sectors. The merger faced regulatory hurdles because of the Bank Holdings Company (BHC) Act. Hence, investors of the rival ﬁrms could react signiﬁcantly to any regulatory objections to, or approval of, the merger. Hence, in addition to examining the market reaction of the global competitors to the merger proposal and completion announcements, we also analyse the competitors’ reaction to other signiﬁcant merger-related events leading up to the consummation of the merger. Extant literature has examined the market reaction of competitors to bank mergers (Aharony & Swary, 1993; Carow, 2001; Eckbo, 1983; Wall & Peterson, 1990). The paper closest to ours is Carow (2001), which examines the impact of the Citicorp–Travelers merger on the probability of deregulation and the wealth effects for ﬁnancial service ﬁrms in the US. He found that life insurance companies and large banks in the US experienced signiﬁcantly positive abnormal returns on the announcement day, while the small banks, health insurers, and property/casualty insurers were not signiﬁcantly affected by the merger. He concluded that investors expected large banks and insurance companies to beneﬁt from deregulation. Our study is different from Carow’s in three ways. First, we examine the reaction of global competitors to this mega merger announcement. The Citicorp– Travelers merger deal has far-reaching ramiﬁcations beyond the US. In fact, the potential competitive effects of the Citicorp–Travelers merger on European ﬁrms compelled the European Union (EU) to investigate the merger. Given the signiﬁcant global presence of the Group in a number of countries, analysis of the reaction of non-US competitors will shed further light on the effects of this mega merger beyond regulatory consequences within the US.

232

ISAAC OTCHERE AND SUHADI MUSTOPO

Second, we examine the reaction of rival investment banks to the merger announcement. One would expect that leading Wall Street securities ﬁrms that are rivals of Salomon Smith Barney (a subsidiary of Travelers) would react signiﬁcantly to the Citicorp–Travelers merger as the merger would signiﬁcantly enhance the competitive position of the merged entity in the investment banking and underwriting services sector. Third, we examine the information effects of a number of regulatory events related to the merger. As mentioned previously, the merger faced regulatory hurdles because it was a direct challenge to the Bank Holdings Company (BHC) Act that prohibited mergers between commercial banks and insurance companies. As a result of the potential industry-wide or competitive effects, the rival ﬁrms could react signiﬁcantly to any potential regulatory objections to and/or approval of the merger. We determine whether rival ﬁrms that were affected by the merger proposal announcement (as a result of the perceived competitive effects or intra industry effects) might react to any regulatory challenges; for example, SEC investigation of the merger that had the potential to scuttle the merger proposal. We use a seemingly unrelated regression (SUR) technique to examine the impact of some regulatory events pertaining to the merger that occurred between the merger proposal announcement date and the merger completion date. By examining the reaction of competitors to other signiﬁcant merger-related events leading up to the completion date, we would understand the intra- and inter-industry effects much more comprehensively. Consistent with our conjecture that the Citicorp–Travelers merger has ramiﬁcations for the global ﬁnancial services ﬁrms, we ﬁnd that non-US competitors, especially rival European banks, reacted positively to the merger announcement. Also, the US rivals, particularly investment banks and large diversiﬁed banks, reacted positively to the merger announcement. We note that the regulatory investigations into the merger created uncertainties. The request by US law makers that the merger be delayed pending their investigations had positive effects on a number of the competitors in both the US and Europe. The positive reaction to the investigations perhaps reﬂects the potential competitive effects of the merger on the rivals. Consistent with this line of argument, we ﬁnd that the competitors reacted negatively to the US Federal Reserve’s approval of the merger. However, after the uncertainties surrounding the merger had abated, the announcement that the merger had been consummated had a signiﬁcantly positive impact on the rivals. The competitors’ reaction was, however, not homogenous. Our cross-sectional analysis shows that the abnormal returns earned by the competitors were greater the larger the competitor. In addition, the

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

233

returns were greater for the US rivals. We conclude that the positive wealth effects experienced by the competitors are consistent with the industry consolidation argument. Also, the fact that the global competitors reacted positively to the Citicorp–Travelers mega merger announcement is consistent with our assertion that the merger had ramiﬁcations that go beyond regulatory concerns in the US. The rest of the paper is organized as follows. In Section 2, we review the literature on mergers and acquisition in the ﬁnancial services industry. Section 3 presents the data selection procedure and methodology. The merger proposal announcement and completion effects are discussed in Section 4. The cross-sectional analysis is presented in Section 5. In Section 6, we discuss the SUR results showing the information effects of the other merger-related events. Concluding remarks are provided in Section 7.

2. LITERATURE REVIEW 2.1. Research on Banks Mergers A number of researchers have investigated mergers among banks and have found mixed results. Houston and Ryngaert (1994), Trifts and Scanlon (1987), and De Long (1998), among others, documented signiﬁcantly negative abnormal returns for bidders and positive abnormal returns for target ﬁrms involved in bank mergers. Neely (1987) and Cornett and Tehranian (1992) also found signiﬁcantly positive returns to bidders in bank mergers. Becher (2000) showed that bank mergers created positive wealth effects for both bidders and targets over a 36-day (30, +5) period. However, he found that the bidders’ returns were signiﬁcantly negative when measured over the 11-day (5, +5) window. In their study of European bank mergers, Cybo-Ottone and Murgia (1998) found that bidders realized positive but insigniﬁcant cumulative abnormal returns while targets experienced signiﬁcantly positive abnormal returns. Prior researchers have also shown that mergers involving banks and insurance companies generate positive wealth effects and have attributed the effects to synergistic beneﬁts. Saunders and Walters (1994) found that the market reacted favourably to mergers between banks and insurance companies because investors perceive such mergers to generate diversiﬁcation beneﬁts. Milbourn, Boot, and Thakor (1999) attributed the positive abnormal returns from such mergers to increased economies of scope, greater business diversiﬁcation, and reduced earnings uncertainty. Research has

234

ISAAC OTCHERE AND SUHADI MUSTOPO

also shown that mergers between large banks and insurers could boost the reputational advantage of the merged business in terms of the ‘‘Too Big to Fail’’ implicit guarantees (Kane, 1994). However, other researchers, including Dickens (1996) and Kane (1984), have argued that the merged entity’s stock price could drop if investors perceive that the merger would force regulators to remove some of the entry barriers and thus open the sector to other competitors. The foregoing discussion suggests that the Citicorp– Travelers (bancassurance) merger could generate synergistic beneﬁts for the combined entity. Perhaps, the strong positive stock market reaction to this merger that resulted in the creation of about $25 billion in value on the merger announcement day reﬂects synergistic beneﬁts. More recently, Carow (2001) has examined the effects of the Citicorp– Travelers merger on their peers in the US. Postulating that the merger challenged the barriers between banking and insurance, Carow argues that the mergers would affect peers because it increased the impetus towards regulatory changes and increased the probability of Congress passing the Financial Services Modernization Act of 1999, which would allow banks and insurance companies to among others, cross-sell bank and insurance products and take advantage of economies of scope (Carow, 2001, p. 1554). Carow ﬁnds that the merger had positive effects on the stock price of the US banks and life insurance companies. 2.2. Effects of Citicorp–Travelers Mergers on Global Competitors We present two alternative hypotheses with regard to the response of global competitors to the Citicorp–Travelers merger. On the one hand, the merger can enhance the operations of the merged entity. Berger, Humphrey, and Pulley (1996) argue that the improvement could emanate from scale and scope economies that lead to lower cost or greater revenues. The successful completion of the merger will enable Citigroup to apply its intangible assets to a larger scale of operations than would otherwise be possible by each party as a single entity. After the acquisition, many of the technological or marketing characteristics that have been internalized by each party in their own areas of comparative advantage would be incorporated into the services provided by the other party. Thus, the combined ﬁrm with its economic strength would be fundamentally set apart from its counterparts. The newfound economies of scale enjoyed by Citigroup may enable it to cross-sell insurance and banking products, reduce cost, undersell its rivals, increase its market share, and enhance its proﬁtability. Thus, the merger, with its attendant competitive threats, could disturb market relations and reduce the

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

235

short-term proﬁts of the competitors who choose to protect their long run market share by cutting margins.2 Based on the foregoing discussion, we conjecture that the merger will negatively affect the operations of the competitors. Therefore, competitors could react negatively to the Citicorp– Travelers merger announcement. However, they would react positively to any investigation that had the potential to terminate the merger proposal since any competitive effects that could result from the merger (should it be consummated) will not materialize. On the other hand, the Citicorp–Travelers merger could generate positive valuation effects for the rivals. First, the merger could lead to an increase in mergers among other large ﬁrms in the industry to take advantage of the implicit too big to fail guarantees that would be enjoyed by Citigroup (Carow, 2001; Kane, 1999). Second, the competitive threat to the rival ﬁrms’ market share could spur them to become more efﬁcient, as they respond to the competitive pressures resulting from the Citicorp–Travelers merger by improving their own services. Therefore, the rival ﬁrms’ stock market reaction to the merger announcement could be positive. We call this the kickin-the-arm effect. Also, the Citicorp–Travelers merger could beneﬁt the competitors because of the likelihood of further industry consolidation and increased probability that the rival ﬁrms would be involved in similar mergers. The industry consolidation hypothesis predicts that rival ﬁrms would react positively to the Citicorp–Travelers merger announcement and negatively to any challenge to this merger from regulators.

3. DATA AND METHODOLOGY 3.1. Data Selection Our sample consists of commercial banks, investment banks, and insurance companies from the G-7 countries (USA, Japan, Germany, Canada, France, UK, and Italy), Australia, and three Eurozone countries (Netherlands, Switzerland, and Spain) that have extensive global presence. The Australian, Canadian, Japanese, and Eurozone samples comprise the ﬁnancial service ﬁrms in the S&P Global 1200, while the US sample consists of ﬁnancial services ﬁrms in the S&P 500 Index as on 31st December 1998 (the ﬁnancial year of the merger). The total sample consists of 107 ﬁrms, 50% of which are US companies. Summary statistics of the sample are presented in Table 1. Panel A presents information on the US sample while Panel B contains statistics on non-US ﬁrms. The data show that the majority of the US ﬁrms

236

ISAAC OTCHERE AND SUHADI MUSTOPO

Table 1. USA

No. of Firms

Descriptive Statistics. Market Value – Mean (in USD Millions)

Market Value – Median (in USD Millions)

Panel A: Summary Statistics of US Rival Sample Regional banks 23 21,250.94 Money centre 2 76,664.90 Financial (diversiﬁed) 5 46,519.66 Investment banking 6 13,224.67 Savings & loan 2 20,879.20 Insurance Life/health 5 4,802.86 Multiline 2 89,702.07 Property/casualty 8 9,578.54 Total Country

5,577.80 89,702.07 7,189.78

24,072.70

53 Commercial Banks

16,268.05 76,664.90 49,798.41 9,324.77 20,879.20

No. of Firms

Mean Market Value (in USD Millions)

Median Market Value (in USD Millions)

Panel B: Summary Statistics of Non-US Sample Australia 1 – Canada 5 – France 2 2 Germany 5 1 UK 7 7 Italy 5 2 Japan 8 – Spain 2 – Switzerland 2 2 Netherlands 1 2

1 5 4 6 14 7 8 2 4 3

35,667.79 17,291.29 33,288.92 33,307.36 31,261.45 16,312.74 13,415.56 35,069.87 32,575.88 38,407.37

35,667.79 18,923.77 27,547.06 36,679.88 16,478.07 12,982.22 10,277.34 28,913.92 29,587.29 38,407.37

Total

54

26,111.86

20,404.04

38

Insurance Companies

13,484.88

16

Note: This table presents the distribution and summary statistics for 107 ﬁnancial services ﬁrms. The statistics are based on the market values of the ﬁrms as of 6 April 1998, the announcement date for the Citicorp–Travelers mergers. Panel A presents the statistics for the US competitors while Pane B shows the statistics for the non-US samples.

are regional banks.3 The non-US ﬁrms are larger in size than the US ﬁrms. This is not surprising because the non-US competitors are multinationals that compete with Citicorp and Travelers in different countries. The stock price data come from Datastream International database. The interest rate data used in the SUR model and data used in the cross-sectional regression as well as the stock price data come from Datastream International database. However, Datastream does not usually include ﬁrms that have been delisted or acquired.4 Hence, for the US and non-US ﬁrms that have

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

237

subsequently been delisted, we obtain the stock price data from the CRSP (Centre for Research in Security Prices) database and Bloomberg, respectively.5 3.2. Methodology In the ﬁrst part of the study, we use the market model to estimate abnormal returns realized by the competitors. The estimation period is from 1 July 1997 to 9 March 1998, the cut off date being 21 trading days prior to the announcement date. The market index in the respective countries is used as the benchmark index in the market model for ﬁrms in that country. Due to differences in time zones, we use an event window that lags the main event window by one day for markets that trade after the US market closes, i.e., the Australian and Japanese markets. Abnormal returns are computed as the difference between the observed returns and expected returns. The abnormal returns are cumulated over different event windows and their signiﬁcance is ascertained by dividing them by their standard errors. We also use Zellner’s (1962) SUR method to examine the effects of various merger-related announcements that occurred between the merger proposal date and the completion date. The SUR model is estimated using a system of n equations over the sample period starting from 1 October 1997 to 29 January 1999.6 For each country, the following set of regression was estimated for the rivals: R1t ¼ a1 þ b1 RMt þ b01 RMt1 þ f1 I ut þ R2t ¼ a2 þ b2 RMt þ b02 RMt1 þ f2 I ut þ

5 P

l1;j Di þ 1t

i¼1 5 P

l2;j Di þ 2t

i¼1

.. . Rnt ¼ an þ bn RMt þ b0n RMt1 þ fn I ut þ

5 P

ln;j Di þ nt

i¼1

where Rnt is the observed return for ﬁrm n on day t, Rmt is the return on the market index on day t and is used to control for general market movements, Rmt1 is the lagged market returns and is used to control for non-synchronous trading,7 fIut is the orthogonalized interest rate variable representing the residuals of the regression of the 10-year government bond returns on the respective market index returns and it represents the unexpected changes in interest rates. We include this interest rate variable in the regression

238

ISAAC OTCHERE AND SUHADI MUSTOPO

because Flannery and James (1984) and Yourougou (1990) found that unexpected changes in interest rates are a signiﬁcant determinant of the returns of banks and other ﬁnancial institutions even after controlling for market risks. Di is a dichotomous variable representing the merger-related challenging event and it equals 1 during the two-day (0, +1) period relative to the merger-related event date and 0 otherwise.8 The coefﬁcient ln,j represents ﬁrm n’s abnormal returns associated with the merger-related announcements j. The parameters of the regression are estimated using weighted least squares regression that assumes that the error terms are independently and identically distributed within each set of equation. Using the SUR method, two main tests of investors’ reactions to each of the merger-related events are carried out. The ﬁrst test is the ‘‘industrywide’’ valuation (information) effects test. The test involves using the Wald test procedure to ascertain whether the sum of the abnormal returns of all ﬁrms in the sample for each country is signiﬁcantly different from zero. The test is represented by the following restriction: l1,0+l2,0+l3,0+?+ln,0 ¼ 0. However, we acknowledge that positive and negative abnormal returns could cancel out each other; in such cases, the industry abnormal returns will be insigniﬁcantly different from zero just because of offsetting effects, hence we also test whether the announcements had differential effects on the sample. We perform the differential effects test by imposing the following restriction across the sample ﬁrms: l1,0 ¼ l2,0 ¼ l3,0 ¼ ? ¼ ln,0 ¼ 0. The information effects and the differential effects test statistic is chi-square distributed. The industry valuation effects and the differential effect test results are presented in Section 5.

4. RESULTS 4.1. Valuation Effects of Citicorp–Travelers Mega Merger Announcement In this section, we discuss the market reaction to the Citicorp–Travelers mega merger announcements using the market model. The excess returns earned by shareholders of the rival ﬁrms at the time of the Citicorp– Travelers merger announcement are presented in Table 2. Panel A shows the full sample results, Panel B presents the results for the European and other non-US ﬁrms, and Panel C contains the results for the US rivals. We ﬁnd that the merger proposal announcement had a positive impact on the full rival sample as indicated by the statistically signiﬁcant abnormal returns of 1.56% on the announcement date. The abnormal returns earned by the

Merger Proposal Announcement Period Abnormal Returns.

All, N ¼ 107

Event Window Mean

Banks, N ¼ 76 t-Stat

Mean

Panel A: Merger Proposal Announcement Effects for the Full Sample (2, 2) 0.0106 (2.11)** 0.0172 (1, 1) 0.0110 (2.02)** 0.0184 0 0.0156 (6.59)*** 0.0202 (0, 1) 0.0068 (1.34) 0.0136 (1, 5) 0.0039 (0.54) 0.0065 Event Window

Insurance, N ¼ 31 t-Stat

Mean

t-Stat

(2.93)* (2.98)* (6.73)*** (2.43)** (0.90)

0.0033 0.0062 0.0066 0.0060 0.0042

(0.34) (0.57) (2.15)** (0.55) (0.25)

Europe All, N ¼ 40

Banks, N ¼ 24

All Non-US Insurance, N ¼ 16

0.0052 (0.60) 0.0044 (0.46) 0.0004 (0.04) 0.0149 (4.83)*** 0.0093

239

Panel B: Abnormal Returns of the Non-US Sample at the Time of the Merger Announcement (2, 2) 0.0062 0.0089 0.0155 (0.59) (0.55) (0.94) (1, 1) 0.0128 0.0242 0.0127 (1.13) (1.35) (0.67) (0, 1) 0.0127 0.0225 0.0139 (1.12) (1.22) (0.73) (0) 0.0097 0.0123 0.0052 (3.47)*** (3.19)*** (1.24) (1, 5) 0.0191 0.0459 0.0163

N ¼ 54

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

Table 2.

240

Table 2. (Continued ) Event Window

Europe All, N ¼ 40

Banks, N ¼ 24

(1.18) Event Window

All

Regional Banks

Money Centre

N ¼ 53

N ¼ 23

N¼2

All Non-US

(1.93)* Diversiﬁed Financial N¼5

Investment Banking N¼6

Insurance, N ¼ 16

N ¼ 54

(0.52) Savings and Loan N¼2

Insurance Life N¼5

Multiline N¼2

Casualty N¼8

All N ¼ 15

0.0041 (0.14) 0.0151 (0.46) 0.0251 (0.85) 0.0081 (1.00) 0.0255 (0.81)

0.0054 (0.32) 0.0041 (0.23) 0.0083 (9.07)*** 0.0068 (1.00) 0.0130 (1.50)

0.0062 (0.79) 0.0063 (0.83) 0.0121 (2.59)** 0.0043 (0.78) 0.0175 (1.15)

0.0106 (0.83) 0.0068 (0.46) 0.0034 (0.26) 0.0189 (1.94) 0.0029 (0.19)

Note: This table shows the abnormal returns for 107 rival ﬁrms following the Citicorp–Travelers merger announcement. Abnormal returns are calculated using the market model. Panel A presents the returns of the combined rival ﬁrms’ sample. Panel B shows the abnormal returns of the European and other non-US ﬁrms, while Panel C contains the abnormal returns of the US ﬁrms. The ﬁgures in parentheses are t-statistics. The symbols ***, **, * represent signiﬁcance at the 1%, 5%, and 10%, respectively.

ISAAC OTCHERE AND SUHADI MUSTOPO

Panel C: Abnormal Returns Earned by the US Rivals at the Time of the Citicorp–Travelers Merger Announcement (2, 2) 0.0162 0.0095 0.0033 0.0492 0.0610 0.0013 (3.24)*** (1.58) (0.12) (2.21)** (5.54)*** (7.03)*** (1, 1) 0.0178 0.0128 0.0213 0.0491 0.0635 0.0077 (3.50)*** (2.94)*** (0.80) (2.86)** (3.05)*** (0.74) (0, 1) 0.0133 0.0088 0.0154 0.0398 0.0564 0.0081 (2.88)*** (2.18)** (0.60) (2.43)** (3.30)*** (6.51)*** (0) 0.0163 0.0091 0.0092 0.0393 0.0652 0.0075 (4.46)*** (3.27)*** (0.61) (2.16)** (4.91)*** (0.01) (1, 5) 0.0173 0.0191 0.0703 0.0019 0.0651 0.0223 (2.72)*** (2.25)** (7.44)*** (0.10) (3.51)*** (0.02)

(0.74)

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

241

sample ﬁrms in the ﬁve days surrounding the announcement date (day 2, +2) are also signiﬁcant at 5%. However, these returns are driven by the banks. The rival banks experienced relatively large abnormal returns of 2.02% on the announcement date (t-statistic ¼ 6.73), while the insurance companies registered only a return of 0.66%. These results suggest that investors expected Citigroup’s rivals to beneﬁt from this mega merger announcement, perhaps through further industry consolidation. Categorizing the sample into US and non-US competitors, we ﬁnd that the non-US competitors, on average, realized positive abnormal returns of 1.49% (signiﬁcant at the 1% level) on the event date. The European rivals reacted positively with an average abnormal return of 0.97%. Consistent with the full sample results, the abnormal returns of the European sample are driven by the banks. The European banks recorded an abnormal return of 1.23% (signiﬁcant at the 1% level) while the European insurance companies did not exhibit any signiﬁcant reaction. As shown in Panel C, the US rivals also reacted positively to the merger announcement, with shareholders of the rival ﬁrms earning statistically signiﬁcant abnormal returns of 1.63% on the announcement day. For the weeks following the merger announcement, the rival ﬁrms’ shareholders continued to earn signiﬁcantly positive abnormal returns. Both the diversiﬁed ﬁnancial service ﬁrms and the regional banks experienced statistically signiﬁcant positive abnormal returns of 3.93% and 0.91%, respectively, on the announcement day. The money-centre banks had a delayed reaction, with abnormal returns of 7.03% being recorded in the week following the announcement. The investment banking ﬁrms also reacted signiﬁcantly to the merger announcement, registering a return of 6.52% on the announcement date, with the positive returns continuing in the week after the announcement. The signiﬁcantly positive abnormal returns documented for the rival investment banks conﬁrm our conjecture that the investment banking sector was affected by this mega merger announcement. Lastly, the savings and loans sector experienced positive abnormal returns of 0.8% (signiﬁcant at the 1% level) over the two-day (0, +1) event window. Consistent with the results documented for the European sample, we ﬁnd that the announcement day abnormal returns of 1.63% documented for the US sample are also driven by the banks. For the insurance companies, we ﬁnd that only the casualty insurance sub-sector exhibits statistically signiﬁcant abnormal returns of 1.21% over the two-day event window. However, the positive returns are offset by the negative, albeit insigniﬁcant return documented for the life insurance and multiline insurance sectors. Contrary to the predictions of the competitive hypothesis, the positive abnormal

242

ISAAC OTCHERE AND SUHADI MUSTOPO

returns documented for the rivals suggests that further industry consolidation is expected in the US banking industry. Except for the results of the regional banks, our results for the US sample are generally consistent with those of Carow (2001). However, we document additional evidence of valuation effects than that documented by Carow in the form of merger completion announcement effects. 4.2. Valuation Effects of Citicorp–Travelers Merger Completion Announcement The Citicorp–Travelers merger faced a number of challenges from regulatory bodies and politicians who wanted to examine the possible competitive effects of the merger. Since these challenges had the potential to scuttle the merger proposal, we examine whether the completion of the merger generated additional returns for shareholders. The abnormal returns earned by the competitors at the time of the merger completion announcement are presented in Table 3. Panel A shows the results for the full sample, Panel B presents the results for the European and other non-US competitors, and Panel C shows the results for the US rivals. We ﬁnd that the merger completion announcement had signiﬁcantly positive effects on the full sample, with the competitors earning abnormal returns of 2.51% over the two-day (0, 1) window and continuing to earn positive returns in the week following the merger completion. The rival banks experienced positive abnormal returns of 1.34% on the merger completion date and 5.29% in the week following the consummation of the merger. The insurance companies also realized 5.25% abnormal returns in the week following the merger completion. Thus, the consummation of this inter sector merger had positive valuation effects on the insurance industry. The non-US sample recorded abnormal returns of 0.96% in the two-day (0, 1) window; and in the week following the consummation of the merger, they realized signiﬁcantly positive abnormal returns of 5.23%. The European rivals did not experience any signiﬁcant reaction on the merger completion date or the day after, albeit we document evidence of differential reaction to the merger completion announcement. The lack of a signiﬁcant reaction from the European sample is due to offsetting effects. The banking sector’s positive cumulative abnormal returns of 2.13% is largely offset by the 1.54% abnormal return earned by the insurance companies (both are signiﬁcant at the 5% level). Despite the lack of signiﬁcant reaction by the combined European sample on the merger completion announcement date, we ﬁnd that the European banks and insurance companies recorded positive

Event Window

Abnormal Returns Around the Merger Completion Date.

Full Sample, N ¼ 107 Mean

t-Statistics

Banks, N ¼ 76 Mean

Insurance, N ¼ 31

t-Statistics

Panel A: Abnormal Returns Earned by the Rivals Around the Merger Completion Date (10, 10) 0.0178 (1.8524)* 0.0318 (2.7813)*** (2, 2) 0.0451 (5.9266)** 0.0563 (6.0708)*** (1, 1) 0.0189 (3.3409)*** 0.0308 (4.5081)*** 0 0.0064 (1.3229) 0.0134 (2.3388)** (0, 1) 0.0251 (4.4742)*** 0.0363 (5.5706)*** (1, 5) 0.0522 (7.7265)*** 0.0529 (6.4773)*** Event Window

Mean

t-Statistics

0.0204 0.0195 0.0098 0.0106 0.0010 0.0525

(1.2463) (1.5513) (1.1583) (1.2594) (0.1043) (4.2163)***

Europe All N ¼ 40

Banks N ¼ 24

Insurance N ¼ 16

N ¼ 54

0.0011 (0.0437) 0.0549 (3.0336)*** 0.0080 (0.6952) 0.0154 (2.1337)*** 0.0131 (1.6284) 0.0517

0.0373 (2.4678)** 0.0846 (7.9180)*** 0.0191 (2.4260)** 0.0096 (1.7753)* 0.0003 (0.0693) 0.0523

243

Panel B: Abnormal Returns of the Non-US Sample Around the Merger Completion Date (10, 10) 0.0319 0.0492 (1.7885)* (1.9933)** (2, 2) 0.0887 0.1112 (7.6703)*** (8.1143)*** (1, 1) 0.0111 0.0230 (1.3738) (2.1634)** (0, 1) 0.0065 0.0213 (0.9489) (2.2873)** (0) 0.0048 0.0006 (0.9142) (0.0910) (1, 5) 0.0662 0.0762

All Non-US

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

Table 3.

244

Table 3. (Continued ) Event Window

Europe All N ¼ 40

Banks N ¼ 24

(5.0673) *** Event Window

All N ¼ 53

Regional Banks N ¼ 23

Money Centre N¼2

All Non-US

(4.4485)*** Financial N¼5

Investment Banking N¼6

N ¼ 54

Insurance N ¼ 16 (2.4798)*** Savings and Loan N¼2

(4.9234)*** Insurance

Life N¼5

Multiline N¼2

All N ¼ 15

0.0416 0.0328 (2.08)** (1.71)* 0.0010 0.0155 (0.06) (0.83) 0.0054 0.0315 (0.33) (1.53) 0.0275 0.0634 (1.21) (2.20)** 0.0055 0.0289 (0.46) (1.27) 0.0552 0.0726 (5.14)*** (5.21)***

Note: This table shows the abnormal returns of 107 rivals at the time of the completion of the Citicorp–Travelers merger. Abnormal returns are based on the market model. Panel A presents the returns of the combined rivals sample. Panels B shows the returns for the European and other non-US ﬁrms, while Panel C exhibits the abnormal returns of the US counterparts. The ﬁgures in parentheses are t-statistics. The symbols ***, **, * represent signiﬁcance at the 1%, 5%, and 10%, respectively.

ISAAC OTCHERE AND SUHADI MUSTOPO

Panel C: Abnormal Returns of the US Competitors at the Time of the Completion of the Citicorp–Travelers Merger (10, 10) 0.0021 0.0131 0.0858 0.0897 0.0421 0.0718 0.0714 0.0205 (0.19) (0.74) (1.15) (3.05)*** (1.86)*** (30.43)*** (1.36) (0.43) (2, 2) 0.0048 0.0248 0.0151 0.0555 0.0461 0.0525 0.0479 0.0061 (0.63) (0.49) (0.43) (2.47)** (1.14) (0.55) (2.66)** (0.13) (1, 1) 0.0188 0.0105 0.0136 0.0603 0.0893 0.0362 0.0249 0.0432 (2.26)** (1.06) (0.30) (4.29)*** (2.28)** (0.37) (0.77) (1.33) (0, 1) 0.0408* 0.0044 0.0515 0.0734 0.1521 0.0547 0.0008 0.0015 (4.30) (2.45)*** (1.19) (2.55)** (4.59)*** (0.84) (0.02) (0.04) (0) 0.0132 0.0078 0.0116 0.0380 0.1298 0.0008 0.0023 0.0400 (1.57) (1.20) (0.50) (1.76)* (5.35)*** (0.04) (0.05) (5.59) (1, 5) 0.0521 0.0628 0.0034 0.0802 0.0177 0.1386 0.0321 0.1458 (6.12)*** (6.22)*** (0.02) (3.05)*** (0.96) (1.48) (1.51) (3.48)***

Casualty N¼8

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

245

abnormal returns of 7.62% and 5.17%, respectively, in the week following the completion of the merger. The results presented in the last column of Panel B show that, generally, the global non-US sample reacted positively to the consummation of the deal. These results reinforce our ﬁnding that the merger was value enhancing for not only the US rivals but also for the European competitors. We also note that after the regulatory hurdles had been cleared and the merger had been consummated, the US competitors reacted positively to the merger completion announcement, earning abnormal returns of 4.08% over the two-day (0, +1) event window. Most of the positive reaction is driven by the large ﬁnancial services and investment banking ﬁrms, with these ﬁrms recording abnormal returns of 3.8% and 12.98%, respectively (signiﬁcant at 10% and 5%, respectively). The US insurance sector was also positively affected by the merger completion announcement, with the rival insurance companies registering abnormal returns of 6.34% and 7.26% over the twoday (0, 1) event window and the week following the merger completion, respectively. The positive abnormal returns earned by the insurance companies were contributed largely by the casualty and multiline sub-sectors. Thus, the announcement that the merger has been consummated following the resolution of the regulatory challenges and the uncertainties surrounding the merger elicited positive stock price reaction from not only the US ﬁrms but other competitors as well. We conclude that the largely positive abnormal returns documented for the sample indicate that industry consolidation beneﬁts outweigh any competitive pressures from Citigroup. The results are consistent with our assertion that the merger had ramiﬁcations that go beyond regulatory concerns in the US.

5. INFORMATION EFFECTS OF THE MERGERRELATED ANNOUNCEMENTS USING SUR MODEL As mentioned previously, the Citicorp–Travelers merger faced regulatory hurdles because it was a direct challenge to the Bank Holdings Company (BHC) Act that prohibited mergers between commercial banks and insurance companies. Because of the potentially positive industry-wide information effects or competitive effects associated with the merger, and the generally positive reaction to the merger proposal announcement, the rival ﬁrms could react signiﬁcantly to any regulatory objections to, or approval of, the merger. We expect that rival ﬁrms that reacted positively (negatively) to the merger proposal announcement would react negatively (positively) to

246

ISAAC OTCHERE AND SUHADI MUSTOPO

Table 4. Dummies

Merger-Related Events.

Date

D1 D2 D3

9 April 1998 24 June 1998 26 June 1998

D4

22 July 1998

D5

23 September 1998

Event US Lawmakers want the merger to be delayed EU clears merger of Citicorp–Travelers Citicorp–Travelers merger under Federal Reserve Bank of New York’s scrutiny Citicorp, Travelers’ shareholders approve the merger The US Federal Reserve approves the merger

Note: This table shows selected merger-related events that occurred between the merger announcement date and the merger completion date.

any merger-related investigation such as the SEC investigation that had the potential to scuttle the merger proposal. In this section, we examine the information effects of ﬁve merger-related announcements listed in Table 4 that occurred between the merger announcement date and completion date.9 We employ the SUR technique to estimate abnormal returns associated with these merger-related events. We use the SUR model to enhance the efﬁciency in the estimation process because the residual correlation of the sample ﬁrms’ returns is likely to be non-zero. The two-day [0, 1] event window is used to capture any excess returns associated with these mergerrelated events. The regression results relating to both industry-wide valuation effects and differential valuation effects are presented in Table 5. Panel A shows the results for the European and other non-US ﬁrms and Panel B presents the results for the US ﬁrms. We ﬁnd that globally, investors reacted signiﬁcantly to Congress’ decision to investigate the anti-competitive nature of the merger. The Canadian and European rivals reacted positively to the request by the law makers that the merger be delayed pending their investigation.10 Similarly, the US rival banks reacted positively to the announcement of US lawmakers’ investigation of the merger with average abnormal returns of 0.69%. Regional banks and money-centre banks reacted signiﬁcantly to this regulatory challenge, earning small but signiﬁcant returns of 0.35%. We expected this announcement to have negative effects on US banks since the initial reaction to the merger announcement was positive. Perhaps the banks reacted this way because of the possibility that this mega merger, with its potentially strong implications for competition, could be scuttled. Consistent with this line of argument, we ﬁnd that the smaller rivals reacted negatively to the US Fed’s approval of the merger. The Federal Reserve’s approval of the merger also

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

Table 5. Country

Abnormal Return Associated with Citicorp–Travelers MergerRelated Events Based on SUR. Hypotheses

D1

D2

Panel A: Two-Day Abnormal Returns Based on SUR Canada Differential effects 0.0181 0.0012 (0.4269) (0.7940) Industry effects 0.0724** 0.0050 (0.0425) (0.8782) Japan Differential effects 0.0228 0.0099 (0.5835) (0.2044) Industry effects 0.1827 0.0790 (0.1036) (0.4809) European (total) Differential effects 0.0080** 0.0015 (0.0139) (0.5975) Industry effects 0.3203*** 0.0610 (0.0327) (0.6709) Banks Differential effects 0.0089*** 0.0028 (0.0296) (0.8188) Industry effects 0.0212* 0.0670 (0.0773) (0.5626) Insurance Differential effects 0.0065 0.0007 (0.2126) (0.4563) Industry effects 0.0849 0.0092 (0.1184) (0.8614) USA

247

Hypotheses

D1

D2

D3

D4

D5

0.0062 0.0001 0.0133 (0.4022) (0.3870) (0.5256) 0.0247 0.0003 0.0531*** (0.3190) (0.8264) (0.0753) 0.0036* 0.0072 0.0173* (0.0002) (0.9652) (0.0006) 0.0289 0.0576 0.1385 (0.7961) (0.6076) (0.2195) 0.0009 0.0033 0.0017*** (0.8305) (0.1511) (0.0000) 0.0364 0.1336 0.0682 (0.8030) (0.3534) (0.6459) 0.0000 0.0039 0.0082*** (0.8471) (0.2065) (0.0000) 0.0004 0.0933 0.0197* (0.9973) (0.4230) (0.0968) 0.0031 0.0026 0.0073** (0.9186) (0.2532) (0.0108) 0.0408 0.0334 0.0955* (0.4392) (0.5265) (0.0756) D3

Panel B: The Two-Day Abnormal Returns for the US Sample Based on All Differential effects 0.0138* 0.0068 0.0068 (0.0860) (0.3122) (0.7800) Industry effects 0.6919*** 0.3380 0.3380 Banks (0.0063) (0.1639) (0.9328) Regional banks Industry effects 0.3505*** 0.1497 0.0532 (0.0019) (0.1351) (0.7577) Money centre Industry effects 0.0702** 0.0181 0.0049 (0.0325) (0.5804) (0.8802) Financial Industry effects 0.0352 0.0052 0.0111 (0.1950) (0.9042) (0.7087) Invest. banking Industry effects 0.1932** 0.0016 0.0279 (0.0171) (0.9847) (0.7298) Savings and loan Industry effects 0.0012 0.0319 0.0159 Insurance (0.9724) (0.3514) (0.6420) All Industry effects 0.0207 0.0173 0.0100 (0.9736) (0.6960) (0.3765) Life Industry effects 0.0141 0.0020 0.0025 (0.7298) (0.9610) (0.9519)

D4

D5

SUR Model 0.0072*** 0.0072*** (0.0009) (0.0000) 0.3622 0.3583 (0.1561) (0.1373) 0.1859 0.2279 (0.1488) (0.2649) 0.0585* 0.0694** (0.0746) (0.0348) 0.0702** 0.1258** (0.0165) (0.0550) 0.1591** 0.1607** (0.0500) (0.0487) 0.0630* 0.0686** (0.0661) (0.0459) 0.0268 0.0870** (0.3880) (0.0379) 0.0242 0.0921** (0.5533) (0.0249)

248

ISAAC OTCHERE AND SUHADI MUSTOPO

Table 5. (Continued ) USA

Hypotheses

Multiline

Industry effects

Casualty

Industry effects

D1 0.0099 (0.7409) 0.0114 (0.1769)

D2 0.0170 (0.5715) 0.0304 (0.5846)

D3

D4

0.0200 0.0012 (0.5042) (0.9690) 0.0554 0.0264 (0.3400) (0.6428)

D5 0.0824*** (0.0063) 0.1020* (0.0509)

Note: This table shows the reaction of the Citicorp–Travelers rivals to the merger-related challenges and investigations. The market reaction is estimated using the seeming unrelated regression (SUR) method. Panel A presents the SUR results for the Canadian, Japanese, and European sample, while Panel B presents the results for the US sample. The ﬁgures in parentheses are p-values, and the symbols ***, **, * represent signiﬁcance at the 1%, 5%, and 10%, respectively.

had negative effects on most of the other rivals including the Canadian and European samples that reacted positively to the initial investigations. Contrary to expectations, the European sample did not signiﬁcantly react to the news of the EU’s approval of the merger; neither did the investors in the US, although the share price of Citicorp and Travelers increased following the approval. As indicated by the differential effects test results, the reaction of competitors to the merger is not homogeneous. Although the Citicorp–Travelers shareholders’ approval of the merger did not have signiﬁcant effects on the combined US sample, the approval was well received by shareholders of the large competitors including the investment banks and diversiﬁed ﬁnancial services ﬁrms. This differential reaction generated a signiﬁcantly high test results (the p-value of the differential effects test is 0.0009, see Panel B). Similarly, the global competitors’ reaction to these events was not homogeneous across the board. For example, the hypothesis that the reaction of the European competitors to the merger was homogenous is rejected at the 5% level for the two events that had the most signiﬁcant impact on the sample, namely the initial investigation by US Congress and the US Federal Reserve Bank’s approval of the merger. The behaviour of the sample ﬁrms’ stock returns around the time of the merger is presented in Fig. 1. We note from the graph that most of the merger-related news that had positive effects on Citicorp also had positive effects on the US and European rivals. Not surprisingly, the merger announcement had the most signiﬁcant effect on Citigroup.11 The US Fed’s approval of the merger also seems to have caused a signiﬁcant spike in the return behaviour of Citigroup. Consistent with the results presented in Table 5, shareholders’ approval of the merger did not signiﬁcantly affect the

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

249

0.25

Percentage Returns

0.2

6/4 Acquisition Announcement

23/9 US Federal Reserve approves merger

0.15

9/10 Merger complete

0.1

0.05

24/6 EU clears merger

9/4 US Law makers want merger delayed

22/7 Shareholders approve merger

0 23Mar-0.05 98

23Apr98

23May98

23Jun98

23Jul98

23Aug98

23Sep98

23Oct98

26/6 Federal Reserve Bank's scrutiny

-0.1

-0.15

Citicorp

(a)

US Rivals (All)

US Rivals (Exc. Insurance)

0.25

0.2

6/4 Acquisition Announcement 23/9 US Federal Reserve approves merger

s Percentage Return

0.15

9/10 Merger complete

0.1

0.05

24/6 EU clears merger

9/4 US Law makers want merger delayed

22/7 Shareholders approve

0 23Mar-0.05 98

23Apr98

23May98

23Jun 98

23Jul98

23Aug98

23Sep98

23Oct98

26/6 Federal Reserve Bank's scrutiny

-0.1

-0.15

(b)

Fig. 1.

Citicorp

All Rivals

European Rivals (All)

Comparable Returns: (a) Citicorp vs. US Rivals; (b) Citicorp vs. Rivals.

returns of the combined rival ﬁrms’ sample. The rivals’ returns, however, appear to have fallen following the US Federal Reserve’s approval of the merger. The evidence conﬁrms our hypothesis that this US-based mega merger affected not only the US rivals but also the European and other competitors.

250

ISAAC OTCHERE AND SUHADI MUSTOPO

6. DETERMINANTS OF THE COMPETITORS’ REACTION TO THE MERGER ANNOUNCEMENT The results presented in the preceding sections show that although the competitors’ reaction to the Citicorp–Travelers merger was generally positive, the reaction was not homogeneous. In this section, we examine the determinants of the competitors’ reaction to the merger announcement using the following regression: AR0i ¼ ai þ bi ln TotAsseti þ gi MV=BVi þ li ROEi þ di DUS þ ei where AR0i is the merger proposal announcement day returns of rival ﬁrm i, TotAsseti is the total assets of ﬁrm i, MV/BVi is the market to book ratio for ﬁrm i, ROEi is the return on equity of ﬁrm i. DUS is a dummy variable that takes on a value of 1 if the competitor is a US ﬁrm and 0 if it is European or other non-US rival. The independent variables are based on the most recent ﬁnancial statement available prior to the merger announcement date. We submit that the sheer size of the merged entity and the attendant implications of the merger for competition in the industry could spur large banks to become efﬁcient. Furthermore, Travelers has substantial investment banking activities through its subsidiary, Salomon Smith Barney; therefore, banks that are active in the investment banking sector (large rivals banks) could react negatively to the merger announcement since the consummation of the merger would automatically lead to the entry of a large and competitive player into the industry. However, since this inter-sector merger elicited positive returns for both the bidder and the target ﬁrms, investors may expect other large competitors to engage in similar mergers and to realize positive abnormal returns. We include total asset in the regression as a proxy for size to ascertain whether the reaction to the merger is a function of the size of the competitor. The ratio of market value of equity-to-book value of equity (MV/BV) is used as a proxy for growth opportunities. If investors expect ﬁrms with high growth opportunities to follow Citicorp–Travelers decision to merge, then given that the stock price of Citicorp and Travelers appreciated following the merger announcement, one would expect such ﬁrms to be positively affected by the merger. On the other hand, ﬁrms with low MV/BV ratio are attractive candidates for takeover because of the potential to extract large beneﬁts from the target. We include this variable in the regression to ascertain whether the rival ﬁrms’ reaction to the merger is a function of their growth potentials. Lastly, return on equity (ROE) is used to proxy for

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

Table 6.

251

Determinants of Citigroup Rivals’ Reaction to the Merger Announcement.

Coefﬁcients

All Samples

Banks

Insurance Firms

Intercept

0.11 (2.69)*** 0.002 (3.07)*** 0.0004** (0.92) 0.0001 (0.02) 0.011 (2.31)**

0.156 (2.59)** 0.009 (2.84)*** 0.0001 (0.02) 0.0006 (0.21) 0.017 (2.66)***

0.02 (0.33) 0.0003 (0.12) 0.0002 (0.40) 0.0007 (0.167)* 0.010 (1.36)

97 0.1284 0.0901 3.39 0.012

67 0.1660 0.1120 3.07 0.022

30 0.2113 0.0850 1.67 0.187

TotAssets MV/BV ROE USD

N R2 Adjusted R2 F-statistic Prob (F-stat)

Note: The table presents OLS regression results of the determinants of the rivals’ reaction to the Citicorp–Travelers merger announcement. The dependant variable is the merger proposal announcement day abnormal returns. Total asset is a proxy for size; MV/BV is the ratio of market value of equity to book value of equity and is used as a proxy for growth opportunities; ROE is the return on equity and USD is a dummy variable that takes on a value of 1 if the rival ﬁrm is a US ﬁrm and 0 otherwise. The independent variables are obtained using data from the most recent ﬁnancial statements prior to the merger in 1998. Due to data limitations, only 97 ﬁrms, consisting of 67 banks and 30 insurance companies, were included in this regression. The t-statistics are based on White heteroskedasticity-consistent standard errors. The symbols ***, **, * indicate that the coefﬁcient is signiﬁcantly different from zero at 1%, 5%, and 10%, respectively.

proﬁtability of the ﬁrms. Proﬁtable ﬁrms become targets in future mergers and acquisition activities and could exhibit strong reaction to the Citicorp– Travelers merger. Alternatively, the competitive pressures resulting from the merger and the potential loss of proﬁts could outweigh the beneﬁts of further industry consolidation, thus the merger could negatively affect proﬁtable ﬁrms. We run the regression for the combined sample and for the banks and insurance sub-samples separately. The results are reported in Table 6. We note from the combined results that size signiﬁcantly explains the competitors’ reaction to the merger announcement. The sign of the coefﬁcient suggests that the larger the rival bank, the greater the abnormal returns. The dummy variable is also signiﬁcant at 5%, suggesting that the merger announcement period abnormal returns were greater for the US rivals. Consistent

252

ISAAC OTCHERE AND SUHADI MUSTOPO

with the event study results, we ﬁnd that most of the abnormal returns we document for the combined sample are driven by the banks. We note from the sub-sample results that the US banks earned signiﬁcantly larger abnormal returns than the non-US (European) rival banks.12 The regression results for the insurance companies are insigniﬁcant, except that ROE is marginally negatively related to the announcement day abnormal returns. This result suggests that proﬁtable insurance companies suffered possibly from competitive effects resulting from the merger. The explanatory power of the regressions is relatively good, with an adjusted R2 of 9.01%, 11.27%, and 21.13% for the combined sample, the rival banks, and insurance ﬁrms’ regressions, respectively.

7. SUMMARY AND CONCLUSION In this paper, we examine global competitors’ reaction to the Citicorp– Travelers’ mega merger announcement and other merger-related events. We ﬁnd that non-US competitors, particularly European banks, reacted positively to the Citicorp–Travelers’ merger announcement. The US banks also reacted positively to the announcement. The request by US law makers that the merger be delayed pending their investigation had a positive effects on a signiﬁcant number of the competitors, including the European rivals. The reaction perhaps reﬂects the potential competitive effects of the merger on the competitors. Consistent with this line of argument, we ﬁnd that the competitors reacted negatively to the US Federal Reserve’s approval of the merger. However, we observe that after the uncertainties surrounding the merger had been abated, the announcement that the Citicorp–Travelers merger has been consummated had a signiﬁcantly positive impact on the competitors, in particular the rival banks and insurance companies in the US and Europe. The fact that most of the competitors of Citigroup reacted positively to both the merger proposal and merger completion announcements indicates that the information effects overcompensates for any competitive effects arising from the merger. Our cross-sectional analysis also shows that the abnormal returns earned by the competitors were greater the larger the competitor. In addition, we ﬁnd that the reaction of the US rivals to the merger was signiﬁcantly greater than that of the global competitors. That the global competitors also reacted positively to the Citicorp–Travelers mega merger announcement is consistent with our assertion that the merger had ramiﬁcations that go beyond regulatory concerns in the US. A search of the Securities Data Corporation Mergers and Acquisition database showed that 35 of the 107 ﬁrms included in the study were involved in

Global Competitors’ Reaction to Citicorp–Travelers Merger Announcement

253

mergers and acquisition activities after the Citicorp–Travelers Group merger had been consummated. In 1998, three of the ten largest mergers and acquisition in US history (namely, Bank of America–Nations Bank, Banc One– First Chicago, and Norwest–Wells Fargo mergers), occurred in the banking industry after the completion of the Citicorp–Travelers Group merger. Also, in 1999 USB and Swiss Bank Corp (SBS) merged and in 2000 JP Morgan and Chase Manhattan merged, while in Europe Allianz acquired Dresdner bank. This evidence corroborates our industry consolidation argument.

NOTES 1. Citicorp’s website, www.citigroup.com 2. See Williamson (1986) and Dunning (1989) for further discussion. 3. The investment banking category includes only four banks. However, for the purpose of this study, we classify two banks from the ‘‘Financial (Diversiﬁed)’’ category, namely JP Morgan and Morgan Stanley Dean Witter, as investment banks. Furthermore, we include in the sample two other investment banks, namely, Donaldson Lufkin and Jeanrette (DLJ) and Paine Webber, that are not part of the S&P 500 index but are widely regarded as ‘‘blue-chip’’ Wall Street investment banks. 4. The only exception is the German-based Dresdner bank. 5. The stock prices are adjusted for stock splits, dividends, and rights issues. 6. Studies that have used the SUR model to analyse the valuation effects of corporate events include Otchere and Chan (2003), Carow (2001), Amoako-Adu and Smith (1995), Schipper and Thomson (1983), and Saunders and Smirlock (1987). 7. The sample comprises primarily blue chip companies and therefore nonsynchronous trading is not likely to be a problem. 8. We examine the information effects of ﬁve merger-related events that occurred between the merger proposal date and the merger completion date (see Table 4 for a list of these events). 9. We analysed the effects of announcements that were likely to be informative. 10. The Australian sample contained only one ﬁrm and was therefore excluded from this analysis. 11. The Citigroup return is the equally weighted return for the two parties. We also use the value weighted returns but the results are qualitatively similar. 12. All but two of the global competitors included in the cross-sectional regressions are European ﬁrms.

REFERENCES Aharony, J., & Swary, I. (1993). Contagion effects of bank failures: Evidence from capital markets. Journal of Business, 56, 305–322. Amoako-Adu, B., & Smith, B. (1995). The wealth effects of deregulation of Canadian ﬁnancial institutions. Journal of Banking and Finance, 19, 1211–1236.

254

ISAAC OTCHERE AND SUHADI MUSTOPO

Becher, D. A. (2000). The valuation effects of bank mergers. Journal of Corporate Finance, 6, 189–214. Berger, A., Demsetz, R. S., & Strahan, P. E. (2000). The consolidation of the ﬁnancial services: Causes, consequences, and implications for the future. Journal of Banking and Finance, 23, 135–194. Berger, A., Humphrey, D., & Pulley, L. (1996). Do consumers pay for one stop banking? Evidence from an alternative revenue function. Journal of Banking and Finance, 20, 1601–1621. Carow, K. (2001). Citicorp–Travelers Group merger: Challenging barriers between banking and insurance. Journal of Banking and Finance, 25, 1553–1571. Cornett, M. M., & Tehranian, H. (1992). Changes in corporate performance associated with bank acquisitions. Journal of Financial Economics, 31, 211–234. Dickens, R. (1996). Contestable markets theory, competition, and the United States commercial banking industry. New York: Garland. Dunning, J. (1986). Japanese participation in British industry: The impact of Japanese affiliates on their competitors, CROOM HELM. Flannery, M., & James, L. (1984). The effects of interest rate changes on the common stock returns of ﬁnancial institutions. Journal of Finance, 39, 1141–1153. Houston, J., & Ryngaert, M. (1994). The overall gains from large bank mergers. Journal of Banking and Finance, 18, 1155–1176. Kane, E. (1984). Technological and regulatory forces in the developing fusion of ﬁnancial services competition. Journal of Finance, 39, 759–772. Kane, E. (1994). What is the value added large US banks find in offering mutual funds? Working Paper, Boston College. Kane, E. (1999). Implications of superhero metaphors for the issue of banking power. Journal of Banking and Finance, 23, 663–673. Milbourn, T., Boot, A., & Thakor, A. (1999). Megamergers and expanded scope: Theories of bank size and activity diversity. Journal of Banking and Finance, 23, 195–214. Neely, W. (1987). Banking acquisitions: Acquirer and target shareholder returns. Financial Management, 16, 66–74. Otchere, I., & Chan, J. (2003). Intra-industry effects of bank privatization: A clinical analysis of the privatization of the Commonwealth Bank of Australia. Journal of Banking and Finance, 27, 949–975. Saunders, A., & Smirlock, M. (1987). Intra- and inter-industry effects of security market activities: The case of discount brokerage. Journal of Financial and Quantitative Analysis, 22, 467–482. Schipper, K., & Thomson, R. (1983). The impact of merger-related regulations on the shareholders of acquiring ﬁrms. Journal of Accounting Research, 21, 184–221. Trifts, J. W., & Scanlon, K. P. (1987). Interstate bank mergers: The early evidence. The Journal of Financial Research, 10, 305–311. Wall, L., & Peterson, D. (1990). The effects of Continental Illinois failure on the ﬁnancial performance of other banks. Journal of Monetary Economics, 26, 77–99. Williamson, P. J. (1986). Multinational enterprise behaviour and domestic industry adjustment under import threat. Review of Economics and Statistics, 3, 359–368. Yourougou, P. (1990). Interest rate risk and the pricing of depository ﬁnancial intermediaries’ common stock: Empirical evidence. Journal of Banking and Finance, 14, 803–820. Zellner, A. (1962). An efﬁcient method of estimating seemingly unrelated regressions and test for aggregate bias. Journal of American Statistical Association, 57, 348–368.

A COMBINED CASCADE MODEL TO EXPLAIN INDUSTRIAL CONSOLIDATION: THEORY AND AN APPLICATION TO STEEL1 Huaichuan Rui ABSTRACT Expansion through mergers and acquisitions (M&As) continues to be a viable international strategy utilised by industrial firms. A striking feature of this is that global giant firms lead the M&A wave and generate an unimaginable impact on relatively small and weak firms across sectors and even nations. There seems to be a kind of ‘cascade effect’ between the industrial consolidations in these areas. A combined cascade model developed in this paper explains that, the power imbalance caused by the degree of consolidation of the players within a firm’s value system determines the movement and direction of the ‘cascade effect’. With the existence of such effect, M&A will be a mutually interdependent, dynamic, reversible and endless process among industries.

Laws are like spider webs. If some poor weak creature comes up against them – it is caught. But the bigger one can break through and get away —Greek philosopher Solon (ca. 630–555 BC)

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 255–281 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07011-7

255

256

HUAICHUAN RUI

1. INTRODUCTION The period since the 1990s has seen an unprecedented intensity of mergers and acquisitions (M&As), resulting in high levels of concentration in a wide range of sectors. Global M&As rose from less than $300 billion in 1992 to around $3,500 billion in 2000 (The Editor, 2001, pp. 23, 39). Although the scale and speed of global M&As have been slowed down since 2000, the concentration process is still inexorable and even speeded up again since 2004. The value of global deals during the period from January to September 2005 has reached $1,800 billion compared with about $1,300 billion in the ﬁrst three quarters in 2004 (FT, 26 Sep 2005, p. 26). This M&A cycle is predicted to be ‘still nascent’ and to have lot of ‘legs’ before it runs out of steam (FT, 26 Sep 2005, p. 26). M&As permeate almost every sector. Firm after ﬁrm shed non-core business to focus on the areas in which the ﬁrm could compete globally, which has resulted in a small number of ﬁrms accounting for over half of the global sales in numerous sectors, from the extremely high-tech sector of aerospace which is dominated by Boeing and Airbus, to the less high-tech sector of coal whose international market is dominated by BHP Billiton, Rio Tinto and Anglo American. The United Nations World Investment Report 2000, concludes that cross-border M&As are the drivers behind the globalisation trend and their ultimate objective is the concentration of market power in a few but mighty hands. Global giant ﬁrms are the most obvious drivers of this concentration process. Possessing a combination of leading brands and technologies, multibillion-dollar procurement budgets and superior human resources, these giants command the M&A wave and generate an unimaginable impact on ﬁrms and industries which are compelled to follow the similar strategy of these leading ﬁrms so as to consolidate themselves. For example, the consolidation of the leading automobile ﬁrms has led to similar consolidation in the pattern and pace of the world’s auto part sector,2 and M&A of leading retailers has led to consolidation of bottlers and even logistics companies. Moreover, consolidation of ﬁrms normally initiated in the headquarter country of the multinational companies (MNCs) has extended into the countries in which they have invested or outsourced (Nolan, 2001). In this paper, such a phenomenon is described as the ‘cascade effect of industrial consolidations’ (referred to as ‘cascade effect’ henceforth), referring to the close connection between consolidations across sectors and even nations. The question of how to explain this ‘cascade effect’ therefore arises. There are various explanations of the logic behind the M&A decision, including

A Combined Cascade Model to Explain Industrial Consolidation

257

the most convincing transaction cost theory (Williamson, 1975) and the resource-based view (Wernerfelt, 1984). However, there are no dedicated theories to explain the logic behind the M&A cascade effect. This paper seeks to ﬁnd such an explanation by developing a combined cascade model which absorbs elements from both industry structure and value chain/value system models and the resource-based view and empirical research evidence on industrial consolidations in the automobile and steel industries in the context of both China and the world. Moreover, this paper aims not only to demonstrate the existence of the cascade effect but also to explain such an effect. The second section presents the developing process of the combined cascade model. The third section provides empirical evidence for the model by demonstrating the cascade consolidations in the global context from steel’s value chain industries (mainly concerned with the auto industry in this research) to the steel industry. The fourth section provides empirical evidence for the model in the Chinese context by showing that the consolidation of global manufacturing industries has compelled the consolidation of China’s manufacturing industries and further to China’s steel industry. In the ﬁnal section, it is concluded that due to the ‘cascade effect’ no ﬁrm and industry seem to be able to ‘escape’ from the new wave of world industry consolidation with the result that the best way to cope with it is to face it and participate in it with competitive advantage.

2. THE COMBINED CASCADE MODEL DEVELOPED TO EXPLAIN THE CASCADE EFFECT The transaction cost theory (Williamson, 1975) justiﬁes acquisition with minimising the production and transaction costs. The resource-based view (Wernerfelt, 1984) suggests that through acquisition a ﬁrm can maximise value by gaining access to the target ﬁrm’s valuable resources (Madhok, 1997; Ramanathan, Seth, & Thomas, 1997). Although motivations of M&A explained by these theories are different, ﬁrms are generally assumed to make their own decisions independently. Therefore, these existing theories, though providing foundations to explain the individual case of M&A, are unable to explain why ﬁrms have to shift their strategy and structure (towards M&A) to follow the leading ﬁrms. Porter’s industry structure model (1980) is widely applied to analyse a ﬁrm’s strategy choice and strategy shift but is less recognised as an equally powerful tool to analyse industrial consolidation. Porter’s value chain model (1985) is used to diagnose each value activity of the ﬁrm so as to gain a

258

HUAICHUAN RUI

competitive advantage, but the fact that the value chain analysis is embedded in a much broader value system analysis is largely neglected. It is impossible to analyse a ﬁrm’s value chain without analysing its value system. In fact, it is the value system analysis that explains in greater depth the cascade effect of industrial consolidation. In the following paragraphs, I will demonstrate how Porter’s industrial structure and value chain/value system models together with the resource-based view can satisfactorily explain the cascade effect of industrial consolidation. In daily business life, a ﬁrm could shift its strategy and structure (towards M&A) by any of a number of causes, even an unexpected social event. However, from a long-term and fundamental view, whether a ﬁrm shifts its strategy and structure is rationalised by the judgement as to whether its competitive advantage and, in turn, its proﬁtability are threatened. According to Porter, the ﬁrst fundamental determinant of a ﬁrm’s competitive advantage and proﬁtability is industry attractiveness, which is embodied in ﬁve competitive forces: threat of new entrants, threat of substitutes, bargaining power of suppliers, bargaining power of buyers and the rivalry among existing competitors (Fig. 1-1 in Porter, 1980, p. 4). The collective strength of these ﬁve competitive forces (or, more simply, three forces, if we view potential entrants as part of the ﬁrm’s rivals and substitutes as part of suppliers) determines the ability of ﬁrms in an industry to earn. In other words, a ﬁrm’s potential proﬁtability is, in reverse proportion, determined by the strengths of the three simpliﬁed forces: buyers, suppliers and competitors. This is why an industry’s consolidation generates a huge threat to its related industries, because the consolidation implies increased bargaining power of the industry due to the consequent reduction in the number of customers in the industry, suppliers with better service, or better control of quantity, quality, price and so on. For example, the increased buying power could inﬂuence the prices ﬁrms can charge and also inﬂuence cost and investment because powerful buyers demand costly service. The second determinant of a ﬁrm’s competitive advantage and proﬁtability is the ﬁrm’s competitive position in its industry. Positioning determines whether a ﬁrm’s proﬁtability is above or below the industry average (Porter, 1985, p. 11). If a ﬁrm is unfortunately located in a fragmented industry, it may still enjoy the above-average proﬁt in its industry provided it has a competitive advantage on cost or differentiation. Porter (1985, p. 33) introduces the value chain as the tool to analyse the source of a ﬁrm’s competitive advantage. The value chain disaggregates a ﬁrm into its strategically relevant activities. A ﬁrm gains competitive advantage by

A Combined Cascade Model to Explain Industrial Consolidation

259

performing the strategically important activities more cheaply or more efﬁciently than its competitors. However, a ﬁrm’s value chain is embedded in a large stream of activities called the value system. Working effectively in the value chain must imply working effectively in the entire value system in which the ﬁrm is embedded. If it is a single-industry ﬁrm, its value chain is embedded in a value system, including the supplier value chains, the buyer value chains and the channel value chains (see Fig. 2-1 in Porter, 1985, p. 35). For example, suppliers have value chains (upstream value) that create and deliver the purchased inputs used in a ﬁrm’s chain. Suppliers also inﬂuence a ﬁrm in many other ways such as reforming a ﬁrm’s inventory system by providing ‘just-in-time’ service. A ﬁrm’s product eventually becomes part of its buyer’s value chain. The ultimate basis for differentiation is a ﬁrm’s and its product’s role in the buyer’s value chain, which determines buyer needs. Most importantly, therefore, as Porter (1985, p. 34) points out, gaining and sustaining a competitive advantage depends on understanding not only a ﬁrm’s value chain but also how the ﬁrm ﬁts into the overall value system. This is why a firm’s consolidation generates a huge threat to its competitors, because the consolidation implies the possible reduction of cost and improved differentiations for the firm so that it can fit into the value system better and gain a competitive advantage over its competitors. We can now better understand the cascade effect of industry consolidation. To minimise costs across the whole value chain, leading global ﬁrms, acting like systems integrators within its value system, penetrate the respective value chains deeply both upstream and downstream. They are closely involved in business activities that range from long-term planning to meticulous control of day-to-day production and delivery schedules. As concentration among systems integrators has proceeded since the 1980s, it has created intensiﬁed pressure from the ﬁrms that sit at the centre of the ‘spider’s web’ that constitutes the global value system. Systems integrators practice their own form of ‘industrial policy’, selecting capable ﬁrms from their supply base that can meet their global needs. This has created an explosive ‘cascade’ effect of institutional change across the whole supply chain and, thereafter, the whole value system. To meet the needs of their giant customers with a worldwide production system, supplier ﬁrms have themselves needed to increase in size and invest in the necessary technologies and minimise costs through large-scale procurement from their own suppliers. To do so, they need a consolidated industry so as to have more bargaining power over both suppliers and customers, to have competitive advantages over competitors so as to sustain

260

HUAICHUAN RUI

their position in the global value system and to increase their proﬁt margins so as to be able to spend more on research and development (R&D). It is at this point that Porter’s industrial analysis model can be best incorporated with the resource-based view. While the former emphasises a ﬁrm’s industrial structure and location in a competitive position towards gaining a competitive advantage, the latter emphasises that a ﬁrm gains competitive advantage by possessing and developing valuable, rare, imperfectly imitable and non-substitutable resource (Barney, 1991), which in turn depends both on the ﬁrm’s endowment and the endless effort in improving its R&D, quality human resources and so on. However, these two modes are not contradictory but supplementary. In fact, Porter’s value chain theory already contains the rationale of the resource-based view. A ﬁrm working effectively on each value-adding activity obviously includes making the best use of the ﬁrm’s resource and core competence. It cannot be denied that a ﬁrm could sustain a competitive advantage more effectively and efﬁciently if it could deploy both means, especially when competition becomes so ﬁerce that failure to make any progress in any of the two would put itself at a disadvantage. At the beginning of the 21st century, this is particularly true while ﬁrms face unprecedented ﬁerce competition with global big business revolution. With this background, it is unimaginable that a global leading ﬁrm can maintain its leading position without possessing and developing its own core competence. But it is equally important to ask whether a ﬁrm is able to develop or worthy of developing core competence all by itself, and even though the ﬁrm is capable of doing so, whether it is effective and efﬁcient enough compared with outsourcing or simply acquiring the core competence through M&A. We have to realise that the dynamic industrial structure and ﬁerce competition brought by the big business revolution have forced ﬁrms to strengthen core competence by combining methods of both M&A and resource-based building up and, probably even more importantly, by whatever means can assist them to win competitive advantage.3 Facing intense competition, a ﬁrm’s M&A decision may not be simply based on its own interests, rather the decision must be based on a consideration of the interests of the entire industry, various relations within the value system and, further, on the whole global environment to which the given industry and value system are connected. It is, therefore, believed that the conventional theories to explain M&A such as the transaction cost and resource-based view alone may not be sufﬁcient to explain the current M&A cascade effect. These theories need to be developed to be a model which combines Porter’s models and the resource-based view, so as to explain the

A Combined Cascade Model to Explain Industrial Consolidation

261

current trend in M&A and to guide ﬁrms to analyse why and how to shift their strategy. Put simply, this combined model tries to explain the cascade effect of industrial consolidation as that, the power imbalance caused by the degree of consolidation of the players within a firm’s value system, which mainly combines the firm, its suppliers and customers, determines the movement and direction of the ‘cascade effect’. Those in a stronger position push the weaker ones to consolidate as the former have more bargaining power and the latter have to consolidate to achieve a better position. Once the weaker ones strengthen themselves after consolidation, they can reverse the direction of the cascade effect and push the movement either backwards or forwards to ﬁrms in their value chains. From this perspective, M&A will be a mutually interdependent, dynamic, reversible and endless process among industries. More precisely, the following two sections provide empirical evidences of how consolidations within a steel ﬁrm’s value system compel the consolidations of the world’s steel industry and, furthermore, of the Chinese steel industry. Taking the world steel industry as an example, the power imbalance caused by the degree of consolidation of the major customers such as auto ﬁrms, major suppliers such as iron ore suppliers and major competing steel ﬁrms determines the movement and direction of the ‘cascade effect’ (Metallurgy Management, 2004). Table 1 is developed here to illustrate the framework. Fig. 1 presents ‘the combined cascade model’, which is developed on the key arguments of existing models. The key arguments and their explanations to the cascade effect are listed in the second and third columns, whereas empirical evidence is given in the fourth and ﬁfth columns of the table. Fig. 1 illustrates the key points of the framework developed in this paper.

3. EMPIRICAL EVIDENCE (1): CASCADE EFFECT ACROSS SECTORS FROM AUTOMOBILES TO STEEL 3.1. Consolidation Within the Steel’s Value System The steel industry has traditionally been thought of as highly fragmented with a low level of technology. However, the impact of its value chain ﬁrms’ consolidation, such as automobile ﬁrms, has compelled intense pressures for industrial concentration on steel ﬁrms. Steel is an important industry, and is likely to remain so, far into the future. Although it accounts for only about 0.5 per cent of the world’s gross domestic product, customers of steel companies are responsible for nearly

The Combined Cascade Model to Explain the Industrial Consolidations. Key Arguments on Firm’s Acquiring Competitive Advantage

How Can the Key Arguments Explain the Cascade Effect

Empirical Evidence in the Global Context

Empirical Evidence in the Chinese Context

Industry structure model

Positioning relative to the ﬁve competitive forces: buyers, suppliers, competitors and so on

Value chain/value system model

By outperforming its competitors. This depends on working effectively in the entire value system, including the supplier’s, the buyer’s and the channel’s value chains

By possessing and developing valuable, rare, imperfectly imitable and non-

The consolidation of steel’s consuming and supplying industries such as auto and iron ore, have compelled the consolidation of world’s steel industry due to the much higher bargaining power of the former Among the world’s top 10 steel ﬁrms, there is intense competition in speed, scale and scope to M&A, to form joint ventures and strategic alliances in order to serve global auto and other customers better and enter emerging markets successfully Despite possessing core competence in making high quality steel, global steel ﬁrms still

The consolidation and relocation of global manufacturers have compelled the consolidation of China’s manufactures, including steel ﬁrms Baosteel and other Chinese steel ﬁrms are active in pursuing M&As and forming joint ventures, in order to sustain domestic market share and access global supply chains

Resource-based view

An industry’s consolidation generates threat to its related industries due to the potentially consequent increase in its bargaining power over other forces A ﬁrm’s consolidation generates threat to its competitors due to the potentially consequent reduction of cost and improved differentiation so as to outperform competitors in their value systems Firms face ﬁerce competition. It is unimaginable that a global leading

The Chinese steel ﬁrms are aware that they lack core competence in

HUAICHUAN RUI

Models

262

Table 1.

The combined cascade model

By developing its own core competences, but also grasping every opportunity to gain core competence by M&A, forming joint ventures and strategic alliances and other possible mechanisms

ﬁrm does not possess core competence. But it is equally important to consider whether a ﬁrm is worthy of developing core competence all by itself To meet the global needs of giant customers, the supply ﬁrms have themselves needed to increase in size and possess core competence. This creates an explosive ‘cascade effect’ of institutional change across sectors and nations

actively seek opportunities of acquiring/cooperating to develop core competences so as to match the innovative pace of customers and confront the bargaining power of giant suppliers Even a few years ago the world’s auto ﬁrms still had a much stronger bargaining power over steel ﬁrms. However, after a series of M&A, the steel ﬁrms have had a better control of the steel supply and its price. This furthermore stimulates the consolidation of not only their customers but also Chinese steel ﬁrms

making high quality steel. This urged them to undertake M&A, to form joint ventures and strategic alliances to acquire core competence quickly and easily The Chinese steel ﬁrms have been compelled to concentrate. However, the latter are still pursuing further concentration in order to serve demanding customers and to have bargaining power over global iron ore suppliers

A Combined Cascade Model to Explain Industrial Consolidation

substitutable resources and core competences

263

264

HUAICHUAN RUI effect on first tier suppliers

Suppliers value chains

(2)

System integrator’s value chains

effect on first tier buyers

Buyers value chains

(1) effect on second tier suppliers effect goes beyond

Second tier suppliers value chains

Other tier suppliers value chains

(3)

Fig. 1. Key Points of the Combined Cascade Model to Explain the Cascade Effect of Industrial Consolidations. Notes: (1) To minimise costs across the whole value chain, global leading ﬁrms, acting like systems integrators among its value system, deeply penetrate the respective value chains both upstream and downstream. As concentration among systems integrators has proceeded since the 1980s, it has created intensiﬁed pressure of concentration across the whole value system. (2) To meet the needs of their giant customers with production systems spread across the world, supplier ﬁrms have themselves needed to increase in size and invest in the necessary technologies and minimise costs through large-scale procurement from their own suppliers. It is at this point that this elective model also incorporates with the resource-based view of ﬁrms to explain the ‘cascade effect’. (3) The combined cascade model developed in this paper explains that the power imbalance caused by the degree of consolidation of the players within a ﬁrm’s value system determines the movement and direction of the ‘cascade effect’. Industrial consolidation will be a mutually interdependent, dynamic, reversible and endless process among industries.

20 per cent of the global output. In the wake of World War II, steel was regarded as a strategic industry and was widely protected by the state, leading not only to a low level of industrial concentration but also to many chronic problems such as surplus capacity. In 1998, the world’s top 10 steel ﬁrms produced only 22 per cent of the global steel output by weight. The world’s largest steel ﬁrm, Posco, produced just 3per cent of the total output (Table 2). Steel has an important place in the supply chain of many industries, with customers in construction, defence, machinery (including automobiles, shipbuilding, railways, etc.) and home appliances. Its own supply chain

Rank 1 2 3 4 5 6 7 8 9 10 Top 10 total (a) World total (b) a/b (%)

1995 Nippon Posco British Steel Usinor Riva Group US Steel NKK Arbed Kawasaki Sumitomo

1998 27.8 23.4 15.7 15.5 14.4 12.1 12 11.5 11.1 10.7 154.2 752 20.5

Posco Nippon Arbed Usinor LNM Group British Steel Thyssen Krupp Riva Group NKK USX

2002 25.6 25.1 20.1 18.9 17.1 16.3 14.8 13.3 11.5 11 173.7 777 22.3

Arcelor LNM Nippon JFE Posco Bao-steel Corus Thyssen Krupp US Steel Nucor

2004 44 34.8 29.8 28.9 28.1 19.5 16.8 16.4 14.4 12.4 245.1 902 27.1

Mittal Arcelor Posco Nippon JFE Nucor US Steel Thyssen Krupp Baosteel Corus

57 43 32 32 30 20 20 20 20 19 293 1,000 29.3

A Combined Cascade Model to Explain Industrial Consolidation

Table 2. Steel Firms’ Rank and Their Output 1995–2004 (Mt).

Sources: FT, 29 Sep (2004), International Iron and Steel Institute (2003, 2004), annual reports of individual companies (various years), Osiris (2004).

265

266

HUAICHUAN RUI

includes iron ore, coke, coking coal, water and logistics. Since the 1990s, most of these industries have consolidated rapidly, internationalised and achieved high levels of technical progress. For example, in the automobile industry, which is one of the main customers for steel, intense competitive pressures have stimulated consolidation among systems integrators.4 These include the merger of Daimler and Chrysler, Ford’s acquisition of Volvo Motors and Jaguar, GM’s acquisition of Saab and Renault’s acquisition of a controlling share of Nissan. By 2003, the top ﬁve automotive manufacturers accounted for over 60 per cent of global production. On the upstream side of the steel industry, in the iron ore industry, intense consolidation meant that by 2003 three giant ﬁrms (Rio Tinto, BHP Billiton and CVRD) produced almost half of the world’s total iron ore output and accounted for 70 per cent of the global seaborne iron ore trade. Finance is an essential part for any industry to create and add value. Steel ﬁrms also strive to attract funding for investment. However, the world’s ﬁnancial market has also been highly globalised (Smith, 2002) and ﬁnancial resources tend to ﬂow to those ﬁrms with the best returns. From 1989 to 2000, the world’s steel industry’s average total shareholder return was negative. It was also lower than the total shareholder returns of other basicmaterials industries (Stikova & Maug, 2004). As a result, steel ﬁrms were relatively starved of external sources of ﬁnance. This was an important stimulus to consolidation in the steel industry, which it was hoped would improve returns and attract more ﬁnancial resources to the industry. Steel ﬁrms cannot escape the pressure of the cascade effect from either the upstream or the downstream side. Since the 1990s, they have faced intensifying pressures to adapt to the changed situation. They have been forced to interact with a diminishing number of larger customers and suppliers, each with massive bargaining power, adopting new methods of procurement and logistics organisation. Since the 1980s, there has been large scale movement of manufacturing into low- and middle-income countries. Of which, the most striking feature is that all the major automobile producers not only have entered China but also had a large scale of investment there, mainly of the type of joint ventures with local best partners, as shown in the next section. Steel ﬁrms have faced the choice of either following their traditional customers or losing them (Morooka, 2002). Steel ﬁrms can only join and retain their place in the world’s leading manufacturers’ supply chains by meeting their intense and remorselessly advancing technology, price and service requirements: Competitive pressure is leading manufacturers to seek out ways to optimise added value, reduce costs and enhance their offer to meet different requirements around the world. To this end, purchasing is becoming increasingly centralised, with outsourcing based on

A Combined Cascade Model to Explain Industrial Consolidation

267

close relationships with sub-contractors, and an accelerating pace of technical innovation through partnerships. (Arcelor, 2003, p. 35)

The only counterweight to the enormous leverage of corporations who make up the buyers is to ‘think big, merge or buy out others, thereby expanding capacity and expunging competition’ (Wheelan, 1999). As will be seen below, this is the fundamental reason for the fact that the top 10 steel ﬁrms in recent years have entered a period of intense M&As as well as forming strategic alliances with both customers and suppliers. 3.2. Consolidation of Steel Firms Under intense pressure from the cascade effect arising from global industrial concentration, between 1995 and 2004, the world’s steel industry entered a period of intense M&As. No fewer than 8 new names appeared in the list of the world’s top 10 steel ﬁrms, each of which resulted from mergers and acquisitions. The share of the top 10 steel ﬁrms rose from 21 per cent in 1995 to over 29 per cent in 2004 (Table 2). Of the top 10 steel ﬁrms in 1995, only Nippon Steel had an output of over 20 Mt, but by 2004, 9 of the top 10 ﬁrms had an output of over 20 Mt. And there is a consensus of opinion in the world’s steel industry that the consolidation process will continue unabated. The chief executives of the world’s top two steel ﬁrms Arcelor and Mittal Steel forecast that within a decade, there will emerge one or two giant ﬁrms with a capacity of 100–120 Mt and several ﬁrms with a capacity of 50–100 Mt. This handful of ‘truly global players’ will leave their footprints in all the major regions (Mittal, 2004; Steel Success Strategies, 2003). The recent Mittal’s hostile bid for Arcelor in May 2006 clearly indicated such serious intention. Moreover, it is thought likely that the new global players will move towards vertical integration of the upstream raw material supply with steel production (Mittal, 2004). Even more signiﬁcant than the increased levels of industrial concentration in terms of tonnage of steel is the fact that the leading steel ﬁrms account for an even larger share of global markets in terms of value of sales revenue (Table 3). The total market for global steel rose from around $250 billion in 1995 to $300 billion in 2003, with the share of the top 10 ﬁrms rising from almost 51 per cent to over 55 per cent in the same years, compared with their 27 per cent share of world total in terms of output weight. This indicates that the world’s leading steel ﬁrms manufacture products with higher unit values, due mainly to the fact that their products are specialised and technologically advanced to meet the demands of leading global customers. As a

268

HUAICHUAN RUI

Table 3.

Share of the 10 Steel Producers in Market Value (Rank by Output) (Million US$).

Rank

1995

2003

Firms

Market Value

Firms

Market Value

1 2 3 4 5 6 7 8 9 10

Nippon Steel Posco British Steel Usinor Riva Group US Steel NKK Arbed Kawasaki Steel Sumitomo Metal

22,969 11,181 11,032 15,719 4,474 6,872 18,711 8,979 12,064 14,830

Arcelor LNM Nippon JFE Posco Baosteel Corus US steel ThyssenKrupp Nucor

31,836 12,000 16,479 22,672 14,930 14,548 14,640 9,458 22,488 6,266

A B C

Top 10 total World total A/B

126,831 250,000 50.7%

165,317 300,000 55.1%

Notes: (1) Nippon and ThyssenKrupp have a relatively large proportion of sales revenue from nonsteel sale, the ﬁgures in this table referred to the sales value of their steel and steel-products only. (2) The data for Riva Group refer to that of 1997. The data of US Steel refer to 1996. Sources: International Iron and Steel Institute (2004); annual reports and press releases of each company in this table.

result, the world’s steel industry is fast becoming bifurcated, one section producing low quality, low value-added products, typically for local ﬁrms and the other producing high value-added, high proﬁt margin products for global ﬁrms. 3.3. Transmitting Consolidation Pressure Across Sectors and Nations More importantly, the world’s giant steel ﬁrms have not only become larger in terms of their volume of output but also enhanced their competitive capability, which is obviously reﬂected in their rising bargaining power and ability to pass consolidation pressures forward to customers, suppliers and competitors. In 2003–2004, the prices of many grades of steel increased by 50–100 per cent in most countries, and the share prices of the publicly quoted steel companies outperformed stock markets by almost 50 per cent (FT, 29 September 2004). In part, this was caused by the boom in the Chinese

A Combined Cascade Model to Explain Industrial Consolidation

269

demand for steel, but, in part, it is also thought to have been stimulated by the preceding period of intense consolidation, which created the bargaining power of large steel ﬁrms with their main customers. A major motive for the steel industry’s drive towards consolidation was the consolidation process in the surrounding value chain, especially in the auto and iron ore industries. In a symbiotic fashion, the subsequent consolidation process in the steel industry and the soaring steel price have helped to drive further consolidation in the surrounding industries. In general, the more consolidated a particular sector is, the higher the industry’s bargaining power with the steel companies. For example, the household appliance sector is considered among the most exposed to the higher steel prices ‘since the industry is highly competitive, with few companies capable of passing on increases in raw material prices to consumers. y In contrast, in sectors such as automotive and construction, companies that buy steel are in a better position to put up prices’ (FT, 24 September 2004). We have seen that the consolidation process has resulted in a small group of immensely powerful steel producers who dominate the production of a range of high-technology, high value-added steels. Meanwhile, M&A strengthens the world’s leading steel ﬁrms so that they are able to expand overseas more quickly and easily. The multinational giant steel ﬁrms are rapidly expanding their production capabilities in fast-growing developing country markets to serve not only the local market but also (possibly more importantly) their global giant customers which are building production bases across the world. Steel production relocation could supply customers worldwide from a relatively short distance and ensure just-in-time supply of high-quality steel. This is a major challenge for steel ﬁrms from developing countries, as the relocation has destroyed the conventional value system of domestic steel ﬁrms and exploited high margin from high-proﬁle customers, as demonstrated in China’s case. China is the world’s largest steel market, consuming 285 Mt in 2004, which was predicted to rise to 300 Mt by 2006 (Luo, 2004). As the world’s fastest industrialising country, China’s high demand for steel and, particularly, for high-quality steel as the global leading manufacturers’ relocation destination has stimulated a rush of multinational steel ﬁrms to follow each other into China, which has been accentuated by the impact of China’s membership of the World Trade Organisation (WTO) on both tariff reduction (which was from 10.87 per cent to 8.07 per cent on steel imports by 2005) and relaxed non-tariff policy. As will be seen later, China already imports a large fraction of its high value-added steel from the world’s leading steel ﬁrms. However, each of the

270

HUAICHUAN RUI

Table 4. The Formation of the Top 10 Firms and Their Presence in China by 2004. Rank Resulting Firm Merged or Acquired Firms M&A Year 1

Mittal

LNM (MNCs), Ispat (MNCs), ISG (US)

2004

2

Arcelor

2001

3

Posco

Unisor (France), Arbed (Luxembourg and Belgium), Aceralia (Spain) Share-swap with Nippon

1999

4

Nippon

Share-swap with Posco

1999

5

JFE

2003

6

Nucor

7

US Steel

8

Baosteel

9

ThyssenKrupp

NKK (Japan), Kawasaki (Japan) BHP (Australia), Trico (US), Birmingham (US) US Steel (US), National Steel (US) Shanghai Metallurgy Holdings (China), Meishan Group (China), etc. Thyssen (Germany), Krupp (Germany) British Steel (UK), Hoogovens (Netherlands)

1999

10

Corus

2002, 2003 2003 1998

1999

Presence in China Green-ﬁeld project, potential JV with Wugang, Valin, etc. JVs with Baosteel, with Baosteel and Nippon

15 JVs located in all over China JVs with Baosteel, with Baosteel and Arcelor JV with Guangzhou Iron & Steel n.a. Actively seeking partners to form JV JVs with ThyssenKrupp, Nippon and Arcelor

JVs with Baosteel, Angang and plants in Wuhan, Dalian n.a.

Note: All top 10 ﬁrms have various cooperative relations with each other or with their value chain ﬁrms, which have not been demonstrated in this table due to limited space. JV, joint venture; n.a., not applicable. Sources: Financial Times and annual reports of these companies.

world’s top 10 steel companies has given high priority to expanding or starting their production operations in China (Table 4). The view of the largest investor, Posco, is representative of the attitudes of these ﬁrms: China is a formidable steel power, accounting for a ﬁfth of the world’s production and a quarter of its consumption. This as well as close proximity to Korea and rapidly growing local demand for high-quality steel products makes the Middle Kingdom a key market for us as we continue to shift our focus to value-added steel grades. By establishing

A Combined Cascade Model to Explain Industrial Consolidation

271

Posco China Holding Corporation, we aim to fully integrate ourselves with Chinese society as a partner for national growth. (italics added, Posco, 2003, p. 144)

Along with Posco, almost all the top 10 global leading ﬁrms, such as Germany’s ThyssenKrupp, Japan’s Nippon and JFE and Mittal, are building a comprehensive production system in different parts of China based on joint ventures with different local partners to produce exclusively high value-added steel products. These large-scale foreign investments have brought major competitive challenges to China’s indigenous steel ﬁrms. This, together with the growth of the global giant manufacturing ﬁrms’ production systems in China, has not only increased demand for steel products but also totally changed the value system that the Chinese steel ﬁrms used to be familiar with and even enjoyed and has, therefore, forced Chinese steel ﬁrms to consolidate in order to compete with the global giant ﬁrms.

4. EMPIRICAL EVIDENCE (2): CASCADE EFFECT ACROSS NATIONS 4.1. Consolidation Within the Chinese Steel’s Value System It is apparent that China has become a world workshop and manufacturing relocation centre during the last two decades. However, the consequent impact on the Chinese manufacturing ﬁrms which are forced to undergo the responsive transformation of strategy and structure and the difﬁculty in doing this is much less apparent. Although consolidation in almost all the Chinese manufacturing sectors has been encouraged by the government as early as the 1980s, such a requirement has never been felt so urgent and forceful before China joined the WTO in 2001. It was believed that the large scale and fast speed of entry of global manufacturers into China are important factors to stimulate the consolidation of Chinese manufacturers (China’s Social Science Academy [CSSA], 2004). On the one hand, entry of the global giants challenges Chinese domestic ﬁrms which feel the urgent need to consolidate to gain economy of size and to gather core competence in order to compete with the global giants. On the other hand, the global giants, once entered, select their competent suppliers and integrate their value chain and value system, which speeds up the consolidation among global and domestic ﬁrms.

272

HUAICHUAN RUI

Steel ﬁrms, as major suppliers of many manufacturing ﬁrms, have witnessed the fast pace and degree of consolidation in their important customer sectors. Taking the steel ﬁrms’ main customer, the automobile ﬁrms, as an example, China’s auto industry, especially after joining the WTO, has undergone a sea change. Under WTO commitments, quotas on cars will be eliminated until mid-2006. Tariffs on automotive imports were cut from 70 to 80 per cent before 2001 and will ultimately come down to 25 per cent by mid-2006. More signiﬁcant than the resulting surge in imports is the large scale of entry into China by the global auto giants due to the relaxed policy on foreign investment in China’s auto industry. By 2002, the world’s top 15 auto giants had picked up their Chinese partners via M&A, share swap, joint ventures and other methods (CSSA, 2004, p. 281). Of those joint venture auto ﬁrms, the most active ones are the passenger car makers. For years, passenger car makers have been the major source of growth in China’s automobile industry. It is not surprising that all the major auto producers are present in China and the top six have invested or plan to invest heavily in China’s auto market. They have deeply penetrated China’s auto industry and even ‘have completed their strategy positioning in China’ (China’s Auto Industry Consultancy Company [CAICC], 2004, p. 70). The important strategies include making China their car-manufacturing centre for the Asian region, selecting the best local partners to form joint ventures and strategic alliances and procuring inputs worldwide and also bringing auto component makers into China. In a single month of 2002, eight auto component multinationals set up their regional headquarters in Shanghai, China (CAICC, 2004, p. 285). China’s auto industry faces challenges from both surging imports and the entrance of global players. It has a widespread production without economies of scale. There are over 120 assemblers, but only 3 have an annual output of over 200,000 vehicles with others averaging 1,000. It is said to be lacking in core competence, especially weak in R&D, marketing and distribution. It is also handicapped by the backwardness in the component, steel and other supporting industries, which is even more widespread and lacking in economies of scale, supplying to only one assembler, with higher prices compared with international ones and lagging behind in the technologies and qualities. Since the very beginning of China’s reform, China’s auto industry, concerned about the competitive capacity of domestic ﬁrms especially after joining the WTO and eager to build up indigenous global competitive ﬁrms, was encouraged to consolidate and to be globally competitive (Nolan, 2001). China’s July 2001 blueprint for the auto industry declared the decision to restructure its 13 major auto manufacturers into three or four groups and

A Combined Cascade Model to Explain Industrial Consolidation

273

to close the hundreds of smaller, inefﬁcient companies engaged in car manufacturing and assembly. Consequently by 2003, the national market share of the top four automobile makers had been increased to 58 per cent (from 38 per cent in 1992), while the market share of the top four passenger car makers reached 54 per cent (CAICC, 2004, pp. 47, 56). However, the Chinese also hoped to build up their global competitive giants through ‘learning by doing’, i.e., by improving itself through the technology transfer of the core competence of joint venture partners. The consolidation between domestic and foreign ﬁrms was dramatically fast while that between competitive domestic ﬁrms (qiang qiang lianhe) was painfully slow. As a result, China’s top auto companies have all been picked up and tied with foreign companies. In 2003, the total national passenger car production was 2,018,875, of which 1,971,601 or 98 per cent were produced by joint ventures (CAICC, 2004, p. 56). China’s auto industry has been consolidated, despite the fact that domestic companies disappeared in the large-scale movement of M&A and joint ventures led by the global auto ﬁrms. No matter whether it is a home car maker or a joint venture or a foreign-invested company, they are all now in a much better position to force the Chinese steel and auto component ﬁrms to provide better quality of products and service. Foreign-invested ﬁrms and joint ventures have already brought their steel and parts suppliers into China to maintain the high quality of their branded cars, as shown in the last section. They do increase procurement from domestic suppliers such as Baosteel, as shown in the next section, but this has no doubt increased pressure on domestic suppliers as they have to compete with global suppliers at a very low base. As for the domestic auto ﬁrms, limited by ﬁnancial resources and industrial policy (which discourages imports), they even more desperately search for higherquality and lower-priced steel and parts suppliers from domestic ﬁrms. We have seen that the consolidated Chinese auto ﬁrms are demanding highquality and competitive steel price from Chinese steel ﬁrms. Compared with the global leading steel ﬁrms, Chinese ﬁrms have competitive disadvantages in one or more of a range of attributes, including product quality, product diversity, faster distribution channels, reliable service and spending on R&D. Consequently, Chinese steel ﬁrms are much less competitive than the global players, as the president of the world fourth largest steel ﬁrm JFE recently commented: As for China, people say they’ve turned into a net importer to exporter. In a sense, this is true, but only for semi-ﬁnished steel products, such as slab. To make high quality steel for cars and for electrical appliances, they still don’t have the know-how and will not have the capability until at least 2008’ (FT, 13 April 2005).

274

HUAICHUAN RUI

It is, therefore, not surprising that the increased entry of international steel ﬁrms and rising imports generate a direct threat for domestic steel ﬁrms to lose traditional as well as prospective customers. This sort of losing out to foreign competitors is reﬂected at the national level by the fact that in 2003 around nine tenths of China’s steel imports consisted of high value-added products (Table 5), and moreover, the share of domestic supply in the total supply of high value-added steel products is declining year after year (Table 6), despite the surge in domestic steel output. We have also seen that global steel ﬁrms have entered China as new suppliers of these customers. Meanwhile, suppliers to the steel ﬁrms such as iron ore ﬁrms have also been very concentrated: China, as a country short in iron ore supply has to purchase almost half of its demand for this commodity from the three major global suppliers introduced in the second section. Chinese buyers are at an obvious disadvantage when thousands of

Table 5. China’s Imports of High Value Steel Products in 2003 (Mt). Items (High Value Products)

Quantity

Cold rolled sheets Hot rolled sheets Galvanised sheets/coated sheets Cold rolled silicon sheets Stainless sheets Alloy steel sheets Other high-quality products Sub-total (high-value products) Total import Share of import by high-value products

9.65 6.98 7.60 1.58 2.75 0.75 3.77 33.08 37.17 89%

Sources: Wu (2004, p. 5), Jia (2004).

Table 6.

2001 2002 2003

Change of Market Share in China’s Steel Market by Home Products (per cent).

All Steels

Medium Sheet

Hot-rolled Sheet

Cold-rolled Sheet

Galvanised Sheet

Tin Plate

Silicon Sheet

90.06 88.41 86.31

98.37a 95.2 89.8

78.41 74.16 57

62.17 59.55 51.82

46.47 37.88 29.54

75.57 71.72 66.82

– 64.59 53.76

Source: Wang (2004, pp. 18–22). a This was the ﬁgure of 2000.

A Combined Cascade Model to Explain Industrial Consolidation

275

small ﬁrms negotiate with the only three global suppliers, especially when the demand was pushed so high before 2004 when the economy was overheated and there was a shortage of iron ore worldwide. We can, therefore, understand how the impact of all these changes has been difﬁcult for Chinese steel ﬁrms to overcome, because the main ‘players’ within its value system – customers, suppliers and competitors – have all undergone a sea change through M&A or other consolidation methods such as forming joint ventures. This has resulted in a severe power imbalance between Chinese ﬁrms and their competitors and suppliers, compelling industrial concentration among Chinese steel ﬁrms. 4.2. Consolidation of Chinese Steel Firms Compared with the high-speed consolidation of the global industries, including the world’s steel industry, the Chinese steel industry has remained highly fragmented, with thousands of medium and small steel ﬁrms ﬁercely competing in the low-value segment of the market. Diseconomies of small scale production are the key reason for the fact that the technology and productivity of the industry are among the lowest in the world. Responding to the explosion in demand, the Chinese steel industry has undertaken enormous investments in capacity expansion, reaching no less than 140 billion yuan (US$17 billion) in 2003. However, a large proportion of the investments was in small ﬁrms producing low value added steel. In sharp contrast with the surge in industrial concentration among the world’s leading steel ﬁrms, the share of China’s top 10 steel ﬁrms in total national steel output fell from 49 per cent in 1999 to 37 per cent in 2003 (Table 7). ‘Small-scale has inevitably led to excessive segmentation and excessive competition, irrational resource allocation and low investment in R&D’ (Jia, 2004). As a result, most Chinese steel ﬁrms lack core competence. China’s steel industry, its policy makers and the companies involved have endeavoured to achieve industrial consolidation. The number of ‘large-scale’ steel companies fell from 1,669 in 1994 to around 260 in 2003. The central government strongly encourages steel ﬁrms, especially leading ones, to expand their scale through both vertical and horizontal integration. The Baoshan Iron and Steel Group Co. Ltd (‘Baosteel’), the only Chinese steel ﬁrm listed in FORTUNE 500 by 2003, is regarded by both Chinese and international analysts as by far the most competitive large Chinese steel ﬁrm. Baosteel has important competitive advantages over its rivals, including both domestic ﬁrms and some international ones. It is the only large steel ﬁrm to have been entirely equipped with the state-of-the-art imported machinery and permitted to use entirely imported high-grade iron ore. China’s

276

Table 7.

HUAICHUAN RUI

Change of Market Share of the Top 10 Chinese Steel Firms 19992003(Mt).

Rank

1 2 3 4 5 6 7 8 9 10 Top 10 total National total Top 10 share(%)

Firm

Output

Rank

Firm

Baosteel Angang Wugang Shougang Bengang Tanggang Hangang Pangang Baogang Hualing

2003 19.88 10.18 8.43 8.17 7.20 6.08 6.06 5.33 5.25 5.19

1 2 3 4 5 6 7 8 9 10

Baosteel Angang Shougang Wugang Bengang Magang Tanggang Pangang Hualing Baogang

81.77 222.34 36.78

Output 2002 19.48 10.07 8.17 7.55 6.21 5.38 5.07 5.00 4.91 4.80

2001 19.14 8.79 8.25 7.09 4.91 4.77 4.60 4.70 4.10 4.20

2000 17.70 8.80 9.00 6.70 4.20 3.90 3.20 3.60 2.80 3.90

1999 16.70 8.50 7.30 6.20 3.90 3.60 3.20 3.30 3.00 3.90

76.64 192.18 39.88

70.55 160.68 43.91

63.80 131.46 48.53

59.60 121 49.25

Sources: China Steel Industry Association (2004), Yin (2003, p. 5), Jia (2004).

large steel ﬁrms all began life before 1976. This left them all with a large ‘legacy’ of problems in terms of old equipment and huge workforces, which was the case also for some of the leading international steel ﬁrms. Uniquely among large Chinese steel ﬁrms, the productivity of the listed Baosteel matches that of the world’s leading steel companies. From the start, Baosteel was designed to focus on high technology and high value-added steel to substitute for imports. It has been the sole or the major domestic supplier of specialist steel for sectors such as automobiles, shipbuilding, pipelines and beverage cans. Despite these advantages, even Baosteel faces great challenges from the internationalisation of China’s steel industry in its efforts to secure a leading position in the high value-added segments of the steel industry. Over many years, it faced intense competition in high value-added steels from imports. For example, joint venture car makers used mainly imported steel. However, as we have seen, alongside Baosteel’s increased capability to supply such steels, it faced a growing challenge both from rising imports and surging production within China by the multi-plant operations of the international steel giants. This threatens Baosteel. As its chairwoman Xie (2004) acknowledged, ‘There are obvious gaps between Baosteel and leading world steel ﬁrms, in terms of core competence, R&D capacity, new technology, human resources and international management capability’.

A Combined Cascade Model to Explain Industrial Consolidation

277

Acknowledging the gap dividing it from the global steel ﬁrms, Baosteel pursues a strategy of tying up with both upstream and downstream global giants r to secure its position within the global value chain. Based on the belief that ‘only those companies which win a secure place within the industrial value chain can survive the current intense competition’ (Li, 2004), Baosteel has become the most active steel ﬁrm in China to strengthen itself by the taking up of mergers, acquisitions, joint ventures and strategic alliances. As early as in 1998, the state-owned Shanghai Metallurgy Holdings and Shanghai Meishan Group were merged into the BaoSteel Group, creating 14 per cent of China’s total steel output. This $12 billion merger enabled Baosteel to be the largest steel ﬁrm in China and one of the largest in the world. However, while it is keen to follow the examples of Nippon and Posco to catch up with the global leaders, Baosteel is regrettably unable to enjoy a consolidated domestic market in the same way that Nippon and Posco enjoyed in their early development period and even now. Despite the progressive consolidation of the domestic steel market, Baosteel as the largest steel ﬁrm in China still only had 8.9 per cent of the Chinese market share in 2004, declining from 14 per cent share in 1998 and can do nothing to control the industry’s overcapacity or prevent the ﬂuctuations in steel prices. It has to struggle for material procurement, infrastructure and price cut with thousands of domestic ﬁrms as well as for high proﬁle customers with global suppliers. It has been suggested that speeding up the consolidation should be supported by the government with integrated policies on M&A, trade, foreign investment and R&D, in the same way that the Japanese government did in the past. Currently, due to the non-existence of such integrated policies, leading Chinese ﬁrms have all been selected by leading global ﬁrms to build up joint ventures, but M&A among large indigenous ﬁrms has failed to take place. Multinational steel giants are taking advantage of forming joint ventures with leading Chinese ﬁrms to organise their own integrated production systems involving multi-plant operations across the whole of China. Some analysts have expressed the fear that top indigenous ﬁrms will be defeated one by one (ge ge ji po) by the global competitors. Consolidation of the indigenous Chinese steel industry has been painfully slow. Local protectionism has prevented the large-scale M&A of steel ﬁrms in different regions. The central government admits that in the market economy it can only play the role of an ‘advisor’ for a ‘free marriage’ but not act as the ‘parent’ for a ‘forced loveless marriage’. There is the sharpest possible contrast between the painful progress of steel industry consolidation in China and the explosive process of consolidation globally.

278

HUAICHUAN RUI

China’s leading steel ﬁrms face the prospect of high-speed growth of production capabilities of the global giants within China which focus on high technology and high value-added products. Moreover, the Chinese government has also signalled its clear intention to gradually sell off the state’s ownership in the country’s leading ﬁrms. This raises the possibility of China’s leading steel ﬁrm, Baosteel, being acquired by one of the global industry leaders. It raises also the possibility of the acquisition of less successful large Chinese steel ﬁrms by leading international steel companies, which might then radically downsize employment, upgrade management skills and receive the infusion of technology that international steel ﬁrms are so reluctant to transfer to their joint venture partners.

5. CONCLUSION This paper has demonstrated that there is a cascade effect between industrial consolidations determined by the global value chain cascade effect connecting ﬁrms, industries and nations, making them interdependent. The combined cascade model explains the cascade effect as that, the power imbalance caused by the degree of consolidation of the players within a ﬁrm’s value system, which mainly combines the ﬁrm, its suppliers and customers, determines the movement and direction of the ‘cascade effect’. Due to the existence of the cascade effect, M&A will be a mutually interdependent, dynamic, reversible and endless process among industries. This presents profound implications for business and management. There has been an unprecedented degree of industrial consolidation since the 1990s, the most easily visible part of the structure of the industrial consolidation being the well-known ﬁrms with powerful, globally recognised brands and/or powerful technologies. Included in those industries that are heavy consumers of steel, are such ﬁrms as GM, Ford and Toyota in automobiles, Whirlpool, Electrolux, Bosch-Siemens (BSH), GE and Maytag in household appliances, Coca-Cola, Pepsico, Anheuser Busch, Interbev and SAB Miller in canned beverages. These constitute the ‘systems integrators’ at the apex of extended value chains. As they have consolidated their leading positions, they have exerted intense pressure across the whole value chain to minimise costs and stimulate technical progress. The ‘cascade effect’ across the global value chain has stimulated a period of unprecedented consolidation in the world’s steel industry, resulting in the rapidly increasing size and power of leading steel ﬁrms such as Mittal, Arcelor, JFE, Nippon Steel and Posco. This,

A Combined Cascade Model to Explain Industrial Consolidation

279

furthermore, challenges much weaker ﬁrms in developing countries, which have to consolidate in order to compete with global giant ﬁrms. The reality of the existence of the cascade effect of the global value chain on industrial concentration and the predicted ‘endless’ M&As by the combined cascade model present immense challenges for ﬁrms across all sectors and nations, especially ﬁrms in developing countries. No ﬁrm or industry seems to be able to escape from the impact of industrial consolidation. Firms, therefore, must shift their strategy and structure to cope with it, no matter whether they intend to do so or not. This, unfortunately, might be the ‘law’ commanding all ﬁrms and industries with the current globalisation trend.

NOTES 1. Unless otherwise indicated, the source of this paper comes from my case studies and interviews. Industrial consolidation includes styles of not only M&A but also forming strategic alliances, joint ventures and many others, but M&A is the major focus in this paper for the sake of simplicity. The gathering speed of forming strategic alliances since the 1990s has also been extraordinary, see Swaminathan and Moorman (2004). 2. For example, it is well known that about seven leading automobile makers dominate world’s automobile market, but it is not well known that the component makers are also highly concentrated. Each segment of the automobile industry is dominated by only two or three of giant component makers, including complete systems for brakes, transmission, electrical circuits, temperature controls, audio, glass, seats and exhausts (Nolan, 2001). There were 14 of them in total listed in the FORTUNE 500 (Fortune, 25 July 2005) 3. Regarding the fact that managers want to grow through acquisitions, or prefer organic growth, research shows that companies that managed to sustain growth best didn’t adhere to one approach to the exclusion of the other. Rather, they expanded through a mixture of acquisitions and organic growth and balanced the two over time. This suggests that the two modes of growth are not substitutes for one another but complementary (Vermeulen, 2005; Vermeulen & Barkema, 2001). 4. These include pressures to spend greater amounts on R&D (to increase safety and well-being, reduce weight and fuel consumption) and on marketing, to introduce new models more rapidly, to take advantage of lower costs achieved through larger volumes of procurement and to reduce costs per unit by producing larger volumes of each model and its variants.

ACKNOWLEDGEMENT I am deeply grateful for the ﬁnancial support of Dr Paul Aiello of the Morgan Stanley Dean Witter and the Isaac Newton Trust of the Cambridge

280

HUAICHUAN RUI

University between 2002 and 2005, and the enduring education and support of Professor Peter Nolan of Cambridge University for many years. The research, which this paper is based, would be impossible without their support. Special thanks are also going to all of my case studied companies and visited government departments, including the joint venture of Shanghai Automobile Industrial Corporation (SAIC) and Voxwagon, Baosteel Group and its joint venture with ThyssenKrupp, the State Reform and Development Commission, China’s Auto Industry Association, and China’s Steel Industry Association, for their great assistance and support during my ﬁeldwork in 2004.

REFERENCES Arcelor. (2003). Arcelor: A portrait. Retrieved March 20, 2003, from http://www.arcelor.com. Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120. China Steel Industry Association. (2004). The rating of domestic steel firms among China Top 500 firms. Retrieved October 10, 2004, from http://www.chinaisa.org.cn/news/rdjd/ rdjd.htm. China’s Social Science Academy (2004). China’s Industrial Development Report 2003. Beijing, China: Economy Administration Press. China’s Auto Industry Consultancy Company, (2004). Operation review and situation analysis of auto industry, memo. FT (2005). How to gain the upper hand. Retrieved April 13, 2005, from http://www.ft.com/cms/ s/85a85810-abba-lld9-893c-000e2511c8,ft_acl=.html International Iron and Steel Institute. (2003). Steel statistical yearbook 2003. Brussels, Belgium: IISI. International Iron and Steel Institute. (2004). Steel ranks. Retrieved July 17, 2004, from http://www.indiainfoline.com/sect/stee/rank.html. Jia, Y. (2004). Interview, 15 August, Beijing. Li, R. (2004). The potential danger of industrial chain can be seen. China’s steel ﬁrms face live and death question. Xinhua News Net. Retrieved March 21, 2004, from http://www.rednet.com.cn. Luo, B. (2004). Speech at The 2004 first news report conference of China’s Steel Industry Association, February 26, 2004. Retrieved February 27, 2004, from http://www.chinaisa. org.cn/news/rdjd/rdjd1.htm. Madhok, A. (1997). Cost, value and foreign market entry mode: The transaction and the ﬁrm. Strategic Management Journal, 18(1), 39–61. Metallurgy Management (2004). World steel industry needs super-large size of merger. Metallurgy Management, 36(1), 51. Mittal, L. (2004). Preface. Forging Ahead. Retrieved August 20, 2004, from http://www. thelnmgroup.com. Morooka, Y. (2002). Presentation by Yasukazu Morooka, President Sumitomo Metal USA Corp. at the Association of Women in the Metal Industries on March 13, 2002 in Cleveland, OH.

A Combined Cascade Model to Explain Industrial Consolidation

281

Nolan, P. (2001). China and the global business revolution. Basingstoke: Palgrave. Osiris (2004). Public companies worldwide. Retrieved August 18, 2004, from http:// osiris.bvdep.com/cgi/template.dll?product=20&user=CambridgeUser-&pw=xl21Ili0O. Porter, M. (1980). Competitive strategy: Techniques for analysing industries and competitor. New York: Free Press. Porter, M. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press. Posco (2003/2004). Posco Annual Report. Retrieved March 20, 2003, from http://www. posco.com.kr. Ramanathan, K., Seth, A., & Thomas, H. (1997). Explaining joint ventures: Alternative theoretical perspectives. In: P. W. Beamish & J. P. Killing (Eds), Cooperative Strategies: Vol 1, North American Perspectives (pp. 51–85). San Francisco, CA: New Lexington Press. Smith, C. R. (2002). Globalization of financial markets. Retrieved August 20, 2004, from http:// media.wiley.com/product_data/excerpt/10/04712292/0471229210.pdf. Steel Success Strategies. (2003). Consolidation won’t level competition, summary of the 18th Annual Conference, held in June 2003 by World Steel Dynamics and American Metal Market, New York. Stikova, J., & Maug, E. (2004). Arcelor: Creation of a European Steel Company [Memo]. Institute for Konzern management, Humboldt Universita¨t zu Berlin, Germany. Swaminathan, V., & Moorman, C. (2004). Value creation in hi-tech alliances: The role of resource strength, resource fit, and resource transfer [Speech]. The Katz Graduate School of Business, University of Pittsburgh, Pittsburgh, PA. The Editor (2001). 10-year global merger completion record 1991–2000. Merger and Acquisitions, (2), 23, 39. Vermeulen, F., & Barkema, H. (2001). Learning through Acquisitions. Academy of Management Journal, 44, 457–476. Vermeulen, F. (2005). How acquisitions can revitalize companies. Sloan Management Review, 46(4), 44–52. Wang, J. (2004). Comments on supply and demand in domestic steel market in 2003. China Steel (3), 18–22. Wernerfelt, B. (1984). A resource-based view of the ﬁrm. Strategic Management Journal, 5(2), 4–12. Wheelan, S. (1999). British Steel and Hoogovens merge. Retrieved July 16, 2004, from http:// www.wsws.org/articles/1999/jun1999/stee-j23.shtml. Williamson, O. E. (1975). Markets and hierarchies: Analysis and antitrust implications. New York: Free Press. Wu, X. (2004). To develop steel industry in a comprehensive, coordinate and sustainable way. Metallurgy Management, 2, 4–7. Xie, Q. (2004). Ascend to FORTUNE 500 to compete with world giants: Exclusive interview with Xie Qihua. Retrieved October 11, 2004, from biz.163.com/2004w08/12642/ 2004w08_1092274037866.html. Yin, R. (2003). The achievement, issues, and prospects of China’s steel industry. Metallurgy Management, 5, 4–11.

This page intentionally left blank

PART V: FINANCE AND GOVERNANCE

This page intentionally left blank

MEASURING AND MANAGING THE FOREIGN EXCHANGE EXPOSURE OF CHINESE COMPANIES Patrick J. Schena ABSTRACT This paper explores the sensitivity of Chinese stock returns to changes in trade-weighted indexes of the renminbi (RMB) and the currencies of China’s trading partners from 1999 to 2003. It analyses this exposure elasticity cross-sectionally using accounting variables to proxy for size and costs of financial distress. It finds that internationally oriented Chinese companies have experienced exchange exposure particularly against the yen. It also finds that, against a trade-weighted index, there is no empirical evidence that Chinese firms are engaged in hedging activities. However, when exposures are measured in yen terms, it finds that Chinese firms, particularly exporters, engage in active currency hedging.

Since 1994, China has ofﬁcially maintained a managed ﬂoating exchange rate system, de facto pegging its currency to approximately 8.25 renminbi (RMB)/USD in 1995 (Wang, 2004; Eichengren, 2004). Along with the peg, China’s capital account has remained tightly controlled. Maintenance of Value Creation in Multinational Enterprise International Finance Review, Volume 7, 285–306 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07012-9

285

286

PATRICK J. SCHENA

such a currency regime has become the cause of considerable consternation among private and public sector interests in the USA, who view China’s management of the RMB as a source of competitive, perhaps even unfair, advantage. While the implications of China’s currency regime are important from a macro-economic standpoint, no less signiﬁcant is its impact on Chinese corporate ﬁnance, from the dual perspectives of exchange rate exposure and foreign exchange risk management. Exchange rate variability is a source of cash ﬂow risk for ﬁrms with foreign-denominated assets and liabilities (whether exporters or importers), as well as for ﬁrms with overseas operations. Additionally, ﬁrms without foreign revenues, costs, or operations might be indirectly affected by exchange rate change through its impact on the effects of foreign competition or broader macro-economic conditions (Parsley & Popper, 2006). A peg, or quasi-ﬁxed exchange rate, seemingly offers a degree of currency stability while eliminating a potential source of cash ﬂow variability. Yet because countries maintain multiple trading partners, while pegging against a single currency, they inevitably retain a degree of exchange exposure against the currencies of these countries. That is, a currency peg may reduce, but does not eliminate, foreign exchange exposure. Less benignly, a peg might also be considered a form of implicit guarantee against exchange rate risk. Under such conditions, the peg could actually encourage risk taking by discouraging currency risk management, as Moreno, Pasadilla, and Remolona (1998) suggest occurred during the Asian Crisis, or encouraging selective hedging (Allayannis, Brown, & Klapper, 2001). Moreover, a peg, combined with capital controls, can stiﬂe or retard the development of on-shore markets for currency derivatives (see Sawyer, 2002). In contrast, ﬁrms in countries with a liberalized capital account and whose currencies ﬂoat freely will have more experience hedging foreign exchange risk and a broader choice of less expensive on-shore and off-shore products with which to do so (Parsley & Popper, 2006). The broader questions of corporate foreign currency exposure and hedging have been examined in some detail in the empirical literature over the last 20 years. From single-factor OLS models, current technology employs a variant of the market model to measure ﬁrm value (in the form of stock returns) as a function of market returns and exposure elasticity (i.e. changes in determinant exchange rates). Most empirical analyses are dominated by studies of US multinationals and have found only limited evidence of signiﬁcant exposure (see Bodnar & Wong, 2000; Jorion, 1990, 1991; Pritamani, Shome, & Singal, 2004, etc.).1 With respect to the present analysis, however, two studies stand out in their broader geographical scope, contextual relevance, and

Foreign Exchange Exposure of Chinese Companies

287

methodological innovation. Parsley and Popper (2006) conduct a thoughtful analysis of the foreign exchange exposure of ﬁrms in nine Asian countries (excluding China) and examine the relationship between the currency regime, speciﬁcally a currency peg, and the exchange rate exposure of individual ﬁrms. He and Ng’s (1998) study of Japanese multinationals utilizes a twostep estimation procedure developed by Jorion (1990) to ﬁrst estimate the foreign exchange exposure of a 171-ﬁrm sample and then explain the exposure coefﬁcient as a function of ﬁrm-speciﬁc accounting variables that proxy for the ﬁrm’s propensity to hedge its exchange rate exposure. My analysis beneﬁts from the prior work of both papers and attempts to extend this literature by conducting an assessment of the foreign exchange exposure of Chinese companies. Its motivation is to understand the impact of changes in the value of the RMB on the value of Chinese companies in the context of China’s capital controls and its policy of pegging the RMB against the dollar. The study estimates the sensitivity of the individual ﬁrm’s stock returns to changes in a trade-weighted index of the RMB, as well as against the currencies of China’s major trading partners, over the ﬁve-year period from 1999 to 2003. It then analyses this exposure elasticity in a cross-sectional model using average accounting variables to proxy for the size and costs of ﬁnancial distress. My review suggests that there are no prior studies of the exchange rate exposure of Chinese ﬁrms, nor studies that attempt to examine hedging practices among Chinese corporates.2 In addition to the empirical value of such a study, it also helps to further understand the structural problems that confront Chinese ﬁrms resulting from the non-convertibility of the RMB and, from a dimension not yet fully explored empirically, the circumstances that complicate and constrain Chinese ﬁnancial policy. The remainder of the paper is organized in the usual fashion. Section 1 presents a brief review of the relevant empirical literature on measuring and managing corporate foreign exchange exposure. Section 2 develops three key hypotheses to be tested. Section 3 describes the methodology employed in this study and explains the manner in which data are selected and tested. Section 4 presents the empirical results, while Section 5 concludes with a summary of the key ﬁndings and a brief discussion of the major limitations of the current study and areas for further research and analysis.

1. LITERATURE REVIEW Jorion (1990) credits Bernard Dumas, among others, with deﬁning economic exposure to foreign exchange volatility as the elasticity between

288

PATRICK J. SCHENA

random changes in an exchange rate and the real value of a ﬁrm, i.e. what Bodnar and Wong (2000) refer to as total exposure. Among the factors complicating total exposure estimates has been the inﬂuence of changes arising from the interaction of domestic economies and foreign markets, i.e. the so-called ‘‘dual-effect’’ dilemma (Pritamani et al., 2004). Because exchange rate effects may be a function of both direct and indirect inﬂuences, simply understanding the form of a ﬁrm’s international engagement may not be sufﬁcient to interpret the impact of exchange rate changes on ﬁrm value, particularly among exporters. For instance, exporting ﬁrms are thought to beneﬁt from a depreciation of the domestic currency, but in fact may suffer if the currency depreciation is accompanied by reduced domestic demand. For importing ﬁrms, the circumstances are less ambiguous as local and foreign market effects are reinforcing such that a change in the value of the local currency enhances (appreciation) or detracts from (depreciation) the value of the importer (see Pritamani et al., 2004). Using simple univariate models of exposure elasticity with a tradeweighted index as the single independent variable, early studies of total exposure often reported only marginally signiﬁcant statistical and economic levels of exposure elasticity. Bodnar and Wong (2000) trace such marginality to model speciﬁcation issues that fail to control for value-relevant marketwide factors – macroeconomic effects – that may be correlated with the exchange rate. Parsley and Popper (2006) further suggest that model speciﬁcation issues can extend to the choice of currency index (i.e. the second right-side variable), emphasizing that individual currency effects may be masked by a trade-weighted index particularly if the home currency is not fully convertible and does not ﬂoat freely. Bodnar and Wong (2000) study the exchange rate exposure of a broad sample of US ﬁrms from 1977 to 1996 and estimate several different model structures – both univariate and multivariate – employing a market proxy to capture macro-economic inﬂuences on equity returns in the latter. They distinguish between the ‘‘total’’ exposure estimates of the univariate models and the more stable and statistically relevant ‘‘residual’’ exposure estimates of the market model. They demonstrate empirically that effectively controlling for market-wide factors in the latter has a discernable impact on the estimation of exposure elasticity and its statistical signiﬁcance. However, and importantly, in measuring residual exposure, the introduction of the market proxy requires interpretive care (Bodnar & Wong, 2000; Pritamani et al., 2004). In a two-factor regression model on ﬁrm-level equity returns against an exchange rate variable and the market variable, the coefﬁcient of the market proxy does not measure the absolute (or total) foreign exchange

Foreign Exchange Exposure of Chinese Companies

289

exposure of the ﬁrm. Rather, it indicates the deviation of the ﬁrm’s exposure from that of the market portfolio. Accordingly, a statistically insigniﬁcant coefﬁcient does not necessarily suggest that the ﬁrm is not exposed to exchange rate risk, but instead that it has the same exposure as the market portfolio (Bodnar & Wong, 2000; Pritamani et al., 2004).3 With respect to the issue of choosing an exchange rate proxy, Parsley and Popper (2006) study the relationship between currency arrangements and ﬁrm-level exchange rate exposure among nine Asian countries and speciﬁcally examine whether currency exposures are masked by trade-weighted indexes in the context of known pegged currencies. They test for both total and residual exposure and use returns on individual exchange rates, in favor of that on a trade-weighted index, to disaggregate and capture the exposure effects of individual currencies. They isolate individual exposure elasticities under both pegged and non-pegged regimes and ﬁnd that despite dollarbased pegs, ﬁrms exhibit signiﬁcant foreign exchange exposure (particularly, in their empirical reference, to the yen). Given the complexity of estimating foreign exchange exposure, it is reasonable to question the economic value of such estimates and whether they justify actionable ﬁnancial choices, such as active programs to manage foreign exchange risk. Some researchers (see Jorion, 1991) have concluded that foreign exchange risk, in a portfolio context, is diversiﬁable. Thus, active hedging programs will not affect a ﬁrm’s cost of capital. Conversely, ﬁnancial theory suggests that if a hedging policy is to affect the value of the ﬁrm, it will do so through tax effects, bankruptcy effects such the cost of ﬁnancial distress, and agency cost effects (Nance, Smith, & Smithson, 1993; Smith & Stulz, 1985). Certainly, there is ample evidence to suggest that ﬁrms actively manage foreign exchange exposure either through ﬁnancial derivatives or structurally via currency decisions concerning the denomination of corporate investments and/or capital acquisitions (Nance et al., 1993). However, as Choi and Kim (2003) show, exposure effects of asset deployments in Asia are complex and interactive with a ﬁrm’s ﬁnancial strategy. Allayannis, Ihrig, and Weston (2001) in fact ﬁnd that operational hedging is not an effective substitute for the use of ﬁnancial derivatives and report that operational hedging strategies only beneﬁt shareholders when used in conjunction with ﬁnancial hedges. This suggests that well-developed, liquid markets for currency derivatives play a critical role in the efﬁcacy of corporate foreign exchange risk management. Having established the broad dimensions of the joint issues of exchange rate exposure and management, it becomes important, from a

290

PATRICK J. SCHENA

methodological perspective, to design an effective approach to identifying and explaining exposure elasticity. Jorion (1990) proposes a two-step estimation procedure in which he ﬁrst estimates exposure from time series regressions of ﬁrm-level stock returns against market returns and a tradeweighted exchange rate, then uses the coefﬁcient of the exchange variable (exposure) as the dependent variable in a cross-sectional regression to be explained by a ﬁrm’s foreign involvement, deﬁned as its ratio of foreign to total sales. He ﬁnds signiﬁcant cross-sectional differences in ﬁrm value and the exchange rate among US multinationals, i.e. exposure was determined to be positively correlated with the degree of a ﬁrm’s foreign operations. He and Ng (1998) extend Jorion’s two-step procedure in their study sample of 171 Japanese multinational ﬁrms during the period from 1979 to 1993. Testing a hedging hypothesis, they suggest, in line with hedging theory, that ﬁrms more exposed to ﬁnancial distress or agency costs have a greater incentive to hedge. They argue that exchange rate exposure, as a function of stock returns, is correlated with a ﬁrm’s export ratio (positively) and with ﬁrm size. They further argue that ﬁrms more exposed to ﬁnancial risk will have a greater propensity to hedge foreign exchange exposure. Consistent with this hypothesis, they ﬁnd that long-term debt, as a proxy for ﬁnancial risk, is negatively correlated with exposure, suggesting that ﬁrms with greater ﬁnancial risk have less exposure and so are more likely to be hedging. They also ﬁnd that accounting measures of liquidity – dividend payout and the quick ratio – are, respectively, negatively and positively correlated with exposure, suggesting that less liquid ﬁrms with higher ﬁnancial risk have less foreign exchange exposure, again implying that they are actively hedging this exposure.

2. HYPOTHESIS DEVELOPMENT A key motivation for this study is to understand the impact of changes in the value of the RMB on the value of Chinese companies in the context of China’s capital controls and its policy of pegging the RMB against the dollar. Often overshadowed, but no less important for the present analysis, is the relationship between the RMB and the Hong Kong dollar (HKD). Here, it is important to note that Hong Kong’s currency board also ‘‘pegs’’ the HKD to the USD, thus de facto pegging the RMB against the HKD. In the context of China’s bilateral trade, this three-way currency arrangement is signiﬁcant. Table 1 lists China’s largest trading partners as of the ﬁrst quarter of 2001. While Japan was China’s largest trading partner,

Foreign Exchange Exposure of Chinese Companies

Table 1. Country Japan USA Hong Kong South Korea Taiwan Germany Singapore Russia UK Netherlands

291

China’s Top Trade Partners (Total Trade, $ million). Q1 of 2001

% Total PRC Trade

Rank 2000

20,559 17,262 11,884 8,102 6,999 5,154 2,689 2,394 2,280 2,174

18.1 15.2 10.4 7.1 6.1 4.5 2.3 2.1 2 1.9

1 2 3 4 5 6 7 16 8 10

Source: PRC General Administration of Customs, China’s Customs Statistics; US–China Business Council at www.uschina.org/press/tradetable.html

representing some 18% of China’s total trade, the USA and Hong Kong combined represented over 25% of China’s external trade. This is indicative of the magnitude of the peg and its impact on the value of the RMB when expressed in terms of a trade-weighted index of the currencies of individual trading partners. Chart 1 plots returns in the RMB nominal, effective exchange rate (NEER), as well as the RMB expressed in units of the currencies of its major trading partners, exclusive of the USA and Hong Kong.4 The indexes of the individual currencies generally track each other, while the NEER exhibits much less variability owing presumably to the inﬂuence of the USD and HKD pegs. Because of the pegs, one would expect little correlation in the value of the RMB against these currencies with its value against the currencies of China’s other major trading partners. Ceteris paribus, a currency peg should reduce aggregate ﬁrm-level exchange rate exposure. Accordingly, I expect that Chinese ﬁrms will not display signiﬁcant exposure elasticity against the RMB expressed as an NEER. H1. When benchmarked against a trade-weighted index, internationally oriented Chinese companies will exhibit little foreign exchange exposure. From a corporate ﬁnance perspective, what practical implications will the peg arrangements have for Chinese corporations? As noted above, while the peg offers a degree of currency stability and mitigates at least one source of cash ﬂow variability, the maintenance of multiple trading partners precludes the elimination of exchange rate risk altogether, even if a portion of bilateral trade is conducted in the pegged currencies. In fact, because of the de facto

292

PATRICK J. SCHENA

160.00

140.00

120.00

100.00

80.00

60.00 NEER euro gbp yen won ntd

40.00

20.00

M 12 M 19 2 98 M 199 4 9 M 199 6 9 M 199 8 9 M 19 10 99 M 19 12 9 9 M 19 2 99 M 200 4 0 M 200 6 0 M 200 8 0 M 20 10 00 M 20 12 0 0 M 20 2 00 M 200 4 1 M 200 6 1 M 200 8 1 M 20 10 01 M 20 12 0 1 M 20 2 01 M 200 4 2 M 200 6 2 M 200 8 2 M 20 10 02 M 20 12 0 2 M 20 2 02 M 200 4 3 M 200 6 3 M 200 8 3 M 20 10 03 M 20 12 0 20 3 03

-

Chart 1.

RMB Exchange Rate Indices.

dollar peg, the RMB should inherit exposures commensurate with those of the appreciation or depreciation of the USD against those currencies. I expect therefore that the value of Chinese companies will be sensitive to changes in individual exchange rates and that the direction of such exposures will be empirically determined based on the international linkages of each company. H2. Chinese companies will exhibit exchange rate exposure to varying degrees when the RMB is measured against non-pegged currencies. Finally, I return again to the issue of hedging. I reiterate the Allayannis, Ihrig, and Weston (2001) ﬁnding that operational strategies are not alone sufﬁcient. Thus, for internationally oriented ﬁrms to effectively hedge they must have access to ﬁnancial derivatives that allow them to structure hedging programs in such a way as to reduce exposure elasticity. This presumes that active, liquid markets exist for such instruments. In the case of China, capital controls and regulatory stagnation have constrained the development of on-shore markets for currency derivatives available to Chinese companies, leaving them nascent and thin (see Sawyer, 2002). In fact, it was not until the autumn of 2003 that Chinese regulators completed the drafting

Foreign Exchange Exposure of Chinese Companies

293

of derivatives regulations, which until then had left banks offering derivatives on-shore exposed to elements of counter-party risk (see Sawyer, 2003). Off-shore, a market in RMB non-deliverable forwards (NDFs)5 is growing steadily and currently dwarfs that of the on-shore market such that RMB NDFs represent 90% of the estimated combined turnover of on-shore forwards and off-shore NDFs. There is currently no onshore market for currency swaps (Ma, Ho, & McCauley, 2004). The ready availability of derivative instruments, notwithstanding, it is nonetheless conceivable that the USD–HKD pegs obviate the need (or reduce the incentive) for Chinese ﬁrms with dominate USD or HKD transactions to hedge. Thus, vis-a`-vis a trade-weighted index, I expect to ﬁnd little evidence of active hedging by Chinese companies. Against individual currencies, I expect hedging effects to be empirically determined in a manner consistent with hedging theory and, thus, related to a ﬁrm’s proximity to ﬁnancial distress.6 H3. Because of disincentives to hedge USD and HKD exposures owing to the peg and subject to the availability of currency derivatives for hedging, Chinese companies will exhibit little evidence of effective foreign exchange hedging when exposures are measured relative to a tradeweighted index.

3. METHODOLOGY, EMPIRICAL MODELS, AND DATA SELECTION To test my hypotheses, I employ the two-step methodology of Jorion (1991) and He and Ng (1998) with some modiﬁcation required by my data. First, to measure foreign exchange exposure, I use a two-factor OLS market model, which measures ﬁrm-level foreign exchange exposure via a currency index variable (see also Bartov & Bodnar, 1994; Pritamani et al., 2004). Second, I analyze exposure elasticity, by cross-sectionally regressing average ﬁrm-level accounting variables on the coefﬁcient of the currency index from the ﬁrst regression. The ﬁrst-step OLS regression takes the basic form: Ri ¼ a þ bRfx þ cRm þ e The dependent variable (Ri) is calculated as the monthly adjusted stock return of ﬁrm i. Explanatory variables include Rfx, the monthly return on a

294

PATRICK J. SCHENA

currency index, which expresses the RMB in foreign currency units, and Rm, the monthly return on a market index of Chinese A-shares.7 The second cross-sectional model takes the form: b ¼ a þ bLTA þ cQUICK þ dDEBTEQ þ e Here, the dependent variable b is the coefﬁcient on the exchange rate variable estimated in the ﬁrst regression. Independent variables include proxies for size and proximity to ﬁnancial distress based on leverage and liquidity. Size is proxied by the log of a ﬁrm’s total assets.8 To capture leverage effects, I use a ﬁrm’s ratio of debt to equity. Liquidity effects are measured using a ﬁrm’s quick ratio. For each variable, I use its 1999–2003 average (i.e. consistent with the time series tests). I also include six industry dummies to ﬁlter international linkages that may be correlated with a ﬁrm’s industry. Admittedly, empirical studies of Chinese corporate ﬁnance suffer from limitations of data availability and reliability. To mitigate such concerns, I begin with a recognized data source – Worldscope – and supplement available fundamental data with adjusted monthly closing stock price data from Bloomberg. For currency pricing data, I use daily RMB exchange rates from www.oanda.com to construct indexes of monthly returns. In addition, I employ the monthly RMB NEER index distributed by the IMF’s International Financial Statistics to capture trade-weighted currency effects. As a market index, I use the Dow-China 88 Index, which consists of the 88 largest and most liquid stocks in the Dow-China Total Market Index selected on the basis of market capitalization and trading liquidity, as measured by average daily turnover.9 The index is quoted in RMB. In designing this study, I set a minimum requirement of ﬁve years of monthly returns. Furthermore, based on measures of China’s international trade since 2000 (see Tables 1–3) and my interest in testing Chinese ﬁrm-level exposure to the non-pegged currencies of China’s major trading partners, I focused my testing on the Japanese yen, the Korean won, the New Taiwan dollar (NTD), the British pound (GBP), and the euro (for which data are available beginning on December 15, 1998). I include the euro for continuity and to capture currency effects against the European Community more generally. My data set then runs from January 1999 through December 2003. With regard to sample selection, Bartov and Bodnar (1994), referencing previous studies of US corporations that failed to document a signiﬁcant correlation between stock returns and dollars ﬂuctuations, note that such ﬁndings may be the result of noise introduced into these studies through, among other things, the inclusion of ﬁrms with limited linkages to

Foreign Exchange Exposure of Chinese Companies

Table 2.

295

China’s Top Exports ($ million).

Commodity Description Electrical machinery and equipment Power generation equipment Apparel Footwear and parts thereof Iron and steel Mineral fuel and oil Toys and games Inorganic and organic chemicals Furniture Plastics and articles thereof

H1 2000

H1 2001

% Change

20,132 12,316 14,537 4,887 4,476 3,042 375 3,328 3,243 3,094

22,974 15,564 14,534 4,996 4,074 3,972 3,832 3,808 3,452 3,220

14.1 26.4 0.02 2.2 9 28.1 2.9 14.4 6.5 4.1

Source: PRC General Administration of Customs, China’s Customs Statistics.

Table 3.

China’s Top Imports ($ million).

Commodity Description Electrical machinery and equipment Power generation equipment Mineral fuel and oil Plastic and articles thereof Iron and steel Inorganic and organic chemicals Medical equipment Man-made ﬁlaments and staple ﬁbers Vehicles other than railway or tramway Copper and articles thereof

H1 2000

H1 2001

% Change

22,469 15,718 8,719 6,508 4,874 4,598 3,195 3,207 1,656 2,196

26,575 19,229 9,387 7,483 6,476 5,035 4,371 3,091 22,338 2,198

18.3 22.4 7.7 15 32.9 9.5 9.5 3.6 34.9 0.1

Source: PRC General Administration of Customs, China’s Customs Statistics.

international conditions. They prescribed developing effective selection criteria to reduce or eliminate this noise. In this light, recent studies have attempted to ﬁlter sample sets by including only listed ‘‘multinationals’’ or other listed ﬁrms ranked above some threshold of the export ratio, the ratio of net foreign sales to total sales, etc. Because of data limitations, I was not able to use such ﬁltering criteria, but instead leverage China’s listing market infrastructure into an alternative selection method. In 1999, China had a total of 949 listed companies. China’s domestic stock markets (Shanghai and Shenzhen) are structurally segmented into two share types. A-shares trade in RMB and are available only to local residents.

296

PATRICK J. SCHENA

B-shares trade and pay dividends in either USD (Shanghai) or HKD (Shenzhen) and were available only to overseas investors until February 2001, when domestic investors with foreign exchange were ﬁrst permitted to buy B-shares. In 1999, there were 82 ﬁrms with listed A- and B-shares and another 26 ﬁrms with only listed B-shares (Wang & Jiang, 2004). In addition to local market listings, Chinese ﬁrms also list on a number of the world’s stock exchanges, most particularly those in Hong Kong, where in 1999, 19 ﬁrms with A-shares also had listed H-shares. In 1999, Worldscope’s coverage of Chinese companies consisted of precisely 127 ﬁrms, which when cross-referenced against the Bloomberg database were found consistent with the 127 listed A–B, A–H, or B-share ﬁrms noted above. Based on their international orientation as exhibited by their cross-listing, these ﬁrms have demonstrated a desire, perhaps a necessity, to attract foreign capital and, if paying dividends, have foreign currencydenominated cash ﬂows (see Chakravarty, Sarkar, & Wu, 1998; Sun & Tong, 2000). As such I argue that, through their cross-listing, they establish, in the Chinese context, clear international linkages. Accordingly, using the Worldscope Chinese company data set for 1999, I selected those ﬁrms with A- and B-shares and excluded ﬁrms with A- and H-shares in order to avoid any residual inﬂuences on returns that might be introduced by the overseas listing. After collecting all relevant pricing and fundamental data for the period from 1999 to 2003, my ﬁnal data set consists of 70 Chinese companies across six industry segments (i.e. approximately 85% of all the companies in this category trading in 1999). All stock returns are calculated using A-share adjusted closing prices, i.e. a market value established by local investors in RMB.

4. TEST PROCEDURES AND EMPIRICAL RESULTS To effectively employ the two-step regression procedure to analyze exchange rate exposure in the context of a pegged currency regime, I run individual time series regressions for each of the ﬁrms in my sample, estimating exposure elasticity using the RMB NEER as a benchmark. Next, to test for individual currency effects, I rerun the ﬁrm-level regressions using a vector of individual monthly currency returns. In both instances, I test for residual exposure and so include the market variable in each set of time series regressions. The results for the ﬁrst-step regressions are given in Table 4. Of the 70 ﬁrms analyzed, in only 9 cases, or 12.9% of the sample, were exchange rate

Ticker 600587 600822 600613 600698 000031 000020 600617 000007 600604

CH CH CH CH CH CH CH CH CH

% Sample

Signiﬁcant Exposure Estimates: RMB NEER.

Company

Exchange

Variable

Shandong Xinhua Medical Shanghai Materials Trading Centre Shanghai Wing Sung Co. Ltd Jinan Qingqi Motorcycles Co. Ltd Shenzhen Baoheng (Group) Co. Shenzhen Huafa Electronics Holding Shanghai Lain Hua Fibre Co. Ltd Shenzhen Seg. Dasheng Corp. Ltd Shanghai Erfangji Textile Machinery

Shanghai Shanghai Shanghai Shanghai Shenzhen Shenzhen Shanghai Shenzhen Shanghai

NEER NEER NEER NEER NEER NEER NEER NEER NEER

Coefﬁcient

t-Statistic

Probability

Adjusted R2

(2.229) 2.038 2.316 (5.142) 2.367 2.689 2.810 1.947 2.081

(1.936) 1.922 1.846 (1.823) 1.805 1.787 1.772 1.750 1.750

0.058 0.060 0.070 0.074 0.076 0.079 0.082 0.086 0.086

0.077 0.279 0.085 0.099 0.194 0.084 0.107 0.368 0.173

Foreign Exchange Exposure of Chinese Companies

Table 4.

12.9

Note: Results of 70 times series regression of the exposure model Ri ¼ a+bRfx+cRm+e, where Rfx is equal to the monthly return on the renminbi nominal, effective exchange rate (RMB NEER).

297

298

PATRICK J. SCHENA

coefﬁcients statistically signiﬁcant and then at only between the 5% and 10% levels, suggesting that few Chinese companies exhibit currency exposure relative to the market when measured against a trade-weighted index. Furthermore, adjusted R2 values were rather low. This ﬁnding is consistent with the inﬂuence of the peg mitigating some degree of foreign exchange exposure and also with evidence of the NEER masking exposure against non-pegged currencies. Interestingly, when the same sample was retested using the vector of individual exchange rates, 24 (or 34.3%) of the ﬁrms tested exhibited signiﬁcant exposure to one or more of the currencies.10 Among the 5 currencies tested, 15 (or over 62% of the sample) displayed signiﬁcant exposure to the Japanese yen, 29.2% to the euro, 25% to the won, 16.7% to the GBP, and 8.3% to the NTD (see Table 5). As is evident from Tables 2 and 3, China’s manufacturing base exhibits a signiﬁcant value-added production component whereby Chinese ﬁrms import primary or intermediate products and export ﬁnished goods. This is discernable from the patterns of imports and exports where the same product sectors make up 60% of China’s trade sectors. Thus, I am unable conclusively to explain the signs of the coefﬁcients, but rather forced to accept them as empirically determined. Nevertheless, while ﬁrms demonstrated variously both signiﬁcant positive and negative exposures to the euro, GBP, won, and NTD, the sign associated with signiﬁcant yen exposure was exclusively negative. This predominance suggests that the Chinese ﬁrms in the sample overwhelmingly beneﬁted (or suffered) from a depreciation (appreciation) of the RMB against the yen as would ﬁrms with an export orientation. Taken together, these tests indicate that the peg may contribute to reduced foreign exchange exposure for Chinese companies, but Chinese corporations nonetheless remain sensitive to ﬂuctuations in the individual currencies of China’s trading partners against which the RMB is not pegged. This is particularly evident in the case of the Japanese yen, where a negative exposure elasticity predominates.11 The second-step regressions are intended to explain the estimated exposures from the ﬁrst step using a multivariate cross-sectional model that includes regressors capturing size, leverage, liquidity, and industry effects. If hedging, by mitigating variance in ﬁrm value, reduces the probability that a ﬁrm will become ﬁnancially distressed, then ﬁrms with greater leverage and lower liquidity (proxies for ﬁnancial distress) will have more incentive to hedge their foreign exchange exposure (Smith & Stulz, 1985). Thus, following He and Ng (1998), I suggest that for ﬁrms that hedge, balance sheet measures of long-term debt will be negatively correlated

Foreign Exchange Exposure of Chinese Companies

Signiﬁcant Exposure Estimates: Yen. Table 5.

299

300

PATRICK J. SCHENA

Foreign Exchange Exposure of Chinese Companies

301

with exposure, indicating that despite their greater ﬁnancial risk, they have less exposure. Similarly, I expect that accounting measures of liquidity – quick ratio – will be positively correlated with exposure, implying that less liquid ﬁrms with higher ﬁnancial risk and who hedge will have less foreign exchange exposure. While one might propose that small ﬁrms are more susceptible to ﬁnancial distress, transactional economies of scale may provide cost-efﬁciencies to hedging to larger ﬁrms (Nance et al., 1993). The latter may be difﬁcult to argue in China’s case owing to the underdevelopment of local derivatives markets. Thus, I expect the sign of the size variable to be positive, indicating hedging activity by small ﬁrms. The results for the second-step tests are found in Table 6a. Here, I test for determinants of RMB NEER exposure. In addition, based on the results of the ﬁrst-step tests and the dominance of yen exposure, I also test for Table 6a.

C LTA QUICK DEBTEQ MANU RETAIL WHOLE FIN TRANS R2 Adjusted R2 Obs: 67

Cross-Sectional Test of RMB NEER Exposures. RMB NEER

Yen

12.94055 5.707981 0.796681 5.369703 0.555692 2.810865 0.000975 0.504254 0.466938 1.047823 0.377094 0.34335 0.209945 0.318806 0.22013 0.301844 0.086849 0.138519

1.868727 1.626169 0.086978 1.15655 0.143968 1.436688 0.001909 1.947017 0.062481 0.276608 0.224263 0.402842 0.119057 0.356668 0.320206 0.866211 0.283666 0.892569

0.436331 0.358583

* *

**

0.198762 0.088247

Note: Results of cross-sectional regression: b ¼ a+bLTA+cQUICK+dDEBTEQ+e. t-statistics in italics: *, **, *** connote 1%, 5%, 10% probabilities, respectively. RMB NEER, renminbi nominal, effective exchange rate.

302

PATRICK J. SCHENA

determinants of yen-based exposure elasticity. I note at the outset that for both RMB NEER and yen exposures, I tested with and without industry dummies and found that the coefﬁcients on the industry variables were largely not signiﬁcant, added little explanatory power to the tests, and did not substantively change the structure of my results. I, nonetheless, report them for completeness. Furthermore, as an alternative size proxy, I also tested using the log of a ﬁrm’s net revenues and found consistently that this regressor produced nearly identical results to those reported using the log of total assets. In tests for RMB NEER exposure, only the size and liquidity proxies were signiﬁcant, both at the 1% level. Interestingly, the sign of the liquidity coefﬁcient was negative and that of the leverage coefﬁcient was positive in each case, suggesting no evidence of hedging. The coefﬁcient of the size proxy was negative, also as expected in the absence of hedging. In tests for yen exposure, only the leverage proxy was signiﬁcant at conventional levels. However, the signs of all variables switch, suggesting consistently, though weakly, some evidence that Chinese ﬁrms hedge yen exposure. Because the exposure estimates used as dependent variables have both positive and negative signs, it is important to examine whether hedging proxies display consistent effects on both positive and negative coefﬁcients. To do so, I construct additional regressors by interacting each of the accounting variables with a dummy variable equal to 1 if the coefﬁcient is negative (DNEER and DYEN) and a second dummy equal to 1 minus each of these dummies (D1N and D1Y, respectively; see He & Ng, 1998). I rerun each cross-sectional regression, replacing the constants with the generated dummies and all interacted accounting variables, and report these results in Table 6b. Again, for completeness, I include industry dummies, which similarly do not materially affect my results. For both currency indexes, all signs remain consistent. Signiﬁcance levels, however, do vary based on the sign of the exposure coefﬁcient. For RMB NEER exposures, explanatory variables for all observations with negative exposures were not signiﬁcant, whereas for observations with positive exposures (suggesting import sensitivity) the signiﬁcance of the size variable held. Though weak, these results again suggest no evidence of hedging in the context of the peg. Conversely, for yen exposures, explanatory variables for all observations with positive exposures were not signiﬁcant, while for observations with negative exposures (i.e. export sensitivity) both the size and the leverage variables were signiﬁcant. These ﬁndings are consistent with Chinese exporters actively hedging yen exposures.

Foreign Exchange Exposure of Chinese Companies

Table 6b.

Cross-Sectional Test of Yen Exposures.

RMB NEER DNEER DNEERLTA DNEERQUICK DNEERDEBTQU D1N D1NLTA D1NQUICK D1NDEBTEQ MANU RETAIL WHOLE FIN TRANS R2 Adjusted R2 Obs: 67

303

1.965345 0.549924 0.162916 0.731467 0.255851 1.115709 0.000976 0.299106 8.931548 3.296868 0.541352 3.00708 0.147371 0.46014 0.001068 0.53882 0.006914 0.017864 0.309457 0.341307 0.181457 0.334087 0.168229 0.263455 0.059737 0.112376

Yen DYEN DYENLTA DYENQUICK DYENDEBTEQ D1Y *

D1YLTA D1YQUICK D1YDEBTEQ MANU RETAIL WHOLE FIN TRANS

0.64714 0.568727

3.130378 3.148121 0.151963 2.35387 0.084077 0.966603 0.001356 1.79007 2.633038 1.232363 0.193856 1.346688 0.034403 0.158443 0.004209 0.885799 0.258629 1.409531 0.066273 0.15551 0.228567 0.882529 0.544339 1.889731 0.392538 1.624665

**

**

0.574387 0.479806

Note: Results of cross-sectional regression: b ¼ a+bLTA+cQUICK+dDEBTEQ+e. t-statistics in italics: *, **, *** connote 1%, 5%, 10% probabilities, respectively. RMB NEER, renminbi nominal, effective exchange rate.

5. CONCLUSION In assessing the exposure to and management of foreign exchange risk by Chinese-listed companies, the foregoing analysis suggests that despite the currency peg, internationally oriented Chinese ﬁrms have experienced signiﬁcant foreign exchange exposure. I ﬁnd that whereas a small number of the ﬁrms in my sample exhibit exchange rate exposure when measured

304

PATRICK J. SCHENA

against a trade-weighted index, approximately 34% of my sample displays a signiﬁcant exposure to changes in the value of one or more of the currencies of China’s major trading partners against which the RMB is not pegged. Indeed, this exposure is particularly acute against the yen. Furthermore, when measured against the trade-weighted index, there is no empirical evidence to suggest that Chinese ﬁrms are engaged in hedging activities. Conversely, when exposures are measured in yen terms, my results indicate, albeit weakly, that Chinese ﬁrms, particularly exporters, engage in some form of active foreign exchange risk management. The ﬁndings of this study focus attention on the complexity of the Chinese Government’s decision process concerning its currency regime. From a dimension not yet fully explored empirically, this study draws attention to the structural problems that confront Chinese ﬁrms resulting from the non-convertibility of the RMB. Speciﬁcally, the peg acts to make hedging unnecessary against the USD and HKD and is likely sustainable only as long as currency controls remain in effect. In addition, the peg burdens the establishment of a culture of currency risk management, while creating a false sense of security. Similarly, the underdevelopment of the markets for currency derivatives make hedging difﬁcult and costly, while also detracting from the evolution of a culture of effective risk management. The Chinese Government must move beyond the peg for macro-economic reasons. What this study suggests is that capital controls must be gradually removed and derivatives markets developed in order to establish the tools with which, as well as the culture in which, Chinese ﬁnancial managers can effectively manage risk. This conclusion is all the more pointed in light of the dubious ﬁnancial condition of both China’s banks and corporates. The results of this study are certainly not without several limitations, which I highlight here as opportunities for further research. First, data availability issues constrained the scope of my study and the selection criteria used to establish the study sample. Future versions of this paper will include a broader data set of Chinese listed companies selected on the basis of import, export, and foreign operations credentials. Second, I limited my study to the use of a single market proxy. The empirical literature (Bodnar & Wong, 2000; Pritamani et al., 2004) cautions on the structure of the market proxy and, in particular, its constitution and weighting scheme. Future work will incorporate robustness tests using alternative proxies for the Chinese market. Third, in testing theories of hedging policy, I focus on issues related exclusively to ﬁnancial distress. Tax and agency cost effects must also be examined. Last, an extension of this analysis might address the important issue of ‘‘pass-throughs’’ as they may impact exposures.

Foreign Exchange Exposure of Chinese Companies

305

NOTES 1. A noted exception is Choi and Prasad (1995), who ﬁnd signiﬁcant exchange rate exposure among 60% of the 409 multinationals studied in their ﬁrm-level analysis. 2. A complementary study might be that by Allayannis, Brown, and Klapper (2001), who analyze the use of derivatives by selected East Asian non-ﬁnancial corporations during and after the Asian Crisis to hedge foreign debt exposure. Included among their Hong Kong company observations are a small number of Chinese ‘‘Red Chips,’’ i.e. companies incorporated and listed in Hong Kong but majority owned by PRC interests. 3. Bodnar and Wong (2000) and Pritamani et al. (2004) test alternative speciﬁcations of the market proxy and ﬁnd signiﬁcant exposure estimates in some models. While this dimension of analysis is beyond the scope of this preliminary assessment, it offers guidance for robustness testing in extensions of this work. 4. Because of the USD peg, the RMB exhibits little to no change against either currency over the reference period. 5. Non-deliverable forwards are currency derivatives traded over the counter and settled not in an exchange of the underlying, but via a net payment in a convertible currency proportional to the difference between the contracted forward rate and a realized sopt rate (Ma et al., 2004). 6. While tax and agency cost effects might also explain hedging behavior, I do not test for these inﬂuences in this analysis. 7. Based on the practice in the asset-pricing literature to estimate the market model using monthly data, I do the same here. Bodnar and Wong (2000) investigate the sensitivity of exposure elasticities to longer horizons and ﬁnd evidence of sensitivity to other return horizons. 8. I also tested using the log of a ﬁrm’s net revenues to proxy for size and found no substantive change in my results. 9. http://chinaindex.dowjones.com/eng/introduction.htm 10. In contrast, approximately 25% of He and Ng’s sample exhibited signiﬁcant positive or negative exposure. 11. This ﬁnding is also consistent with McKinnon and Schnabl’s (2003) contention that the predominant invoice currencies among East Asian ﬁrms (for both exports and imports) have been the USD and the Japanese yen.

REFERENCES Allayannis, G., Brown, G. W., & Klapper, L. F. (2001). Exchange rate risk management: Evidence from East Asia (Policy research working paper). Washington, DC: World Bank. Allayannis, G., Ihrig, J., & Weston, J. (2001). Exchange-rate hedging: Financial versus operational strategies. American Economic Review, 91, 391–395. Bartov, E., & Bodnar, G. M. (1994). Firm valuation, earnings expectations, and the exchangerate exposure effect. Journal of Finance, 49, 1755–1786.

306

PATRICK J. SCHENA

Bodnar, G. M., & Franco Wong, M. H. (2000). Estimating exchange rate exposure: Some ‘‘weighty’’ issues (Working Paper No. 7497, pp. 1–44). Cambridge, MA: National Bureau of Economic Research. Chakravarty, S., Sarkar, A., & Wu, L. (1998). Information asymmetry, market segmentation, and the pricing of cross-listed shares: Theory and evidence from Chinese A and B shares. Journal of International Financial Markets, Institutions, and Money, 8, 325–356. Choi, J. J., & Kim, Y.-C. (2003). The Asian exposure of U.S. ﬁrms: Operational and risk management strategies. Pacific-Basin Finance Journal, 11, 121–138. Choi, J. J., & Prasad, A. M. (1995). Exchange risk sensitivity and its determinants: A ﬁrm and industry analysis of U.S. multinationals. Financial Management, 24, 77–88. Eichengren, B. (2004). Chinese currency controversies (Working paper, pp. 1–58). Berkeley: University of California. He, J., & Ng, L. K. (1998). The foreign exchange exposure of Japanese multinational corporations. Journal of Finance, 53, 733–753. Jorion, P. (1990). The exchange rate exposure of US multinationals. Journal of Business, 63, 331–345. Jorion, P. (1991). The pricing of exchange rate risk in the stock market. Journal of Financial and Quantitative Analysis, 26, 363–376. Ma, G., Ho, C., & McCauley, R. N. (2004). The markets for non-deliverable forwards in Asian currencies. BIS Quarterly Review, June, 81–94. McKinnon, R., & Schnabl, G. (2003). The East Asian dollar standard, fear of floating, and original sin (Working paper). Stanford, CA: Stanford University. Moreno, R., Pasadilla, G., & Remolona, E. (1998). Asia’s financial crisis: Lessons and policy responses. San Francisco: Federal Reserve Bank of San Francisco, Center for Paciﬁc Basin Monetary and Economic Studies. Nance, D. R., Smith, C. W., Jr., & Smithson, C. W. (1993). On the determinants of corporate hedging. Journal of Finance, 58, 267–284. Parsley, D., & Popper, H. (2006). Exchange rate pegs and foreign exchange exposure. Journal of International Money and Finance, Forthcoming. Pritamani, M. D., Shome, D. K., & Singal, V. (2004). Foreign exchange exposure of exporting and importing ﬁrms. Journal of Banking and Finance, 28(7), 1697–1710. Sawyer, N. (2002). Scaling the wall. Retrieved 15 April, 2004, from http://www.asiarisk.com.hk/ april02/corporatehedging.htm Sawyer, N. (2003, September). Setting the framework. AsiaRisk. Retrieved 15 April, 2004, from http://www.asiarisk.com.hk/public/showPage.html?page=17571 Smith, C. W., & Stulz, R. M. (1985). The determinants of ﬁrms’ hedging policies. Journal of Financial and Quantitative Analysis, 20, 391–405. Sun, Q., & Tong, W. H. S. (2000). The effect of market segmentation on stock prices: The China Syndrome. Journal of Banking and Finance, 24, 1875–1902. Wang, S. S., & Jiang, L. (2004). Location of trade, ownership restrictions, and market illiquidity: Examining Chinese A- and H-shares. Journal of Banking and Finance, 28, 1273– 1297. Wang, T. (2004). China: Sources of real exchange rate fluctuations (Working paper, pp. 1–22). Washington, DC: International Monetary Fund.

UK MEASURES OF FIRM-LIVED EQUITY DURATION Richard A. Lewin, Marc J. Sardy and Stephen E. Satchell ABSTRACT Investors often have much of their portfolios invested in equities that are exposed to interest rate risk. Hedging underlying exposures are not easy; whereas fixed income investors have duration to immunize bond portfolios from small fluctuations in interest rates. US equity duration estimates from dividend discount models result in long durations – often in excess of 50 years. Based on the UK data, we develop an alternative approach to generate equity duration as a by-product of asset pricing. Our analysis suggests that the equity premium puzzle may comprise an important element in reconciling this approach to equity duration, with traditional DDM alternatives.

Currently, ﬁrms are taking greater advantage of global capital markets and the inherent opportunities to ﬁnance internationally through various ﬁnancial instruments that lower capital costs. While this may decrease overall borrowing costs, by allowing ﬁrms to match country-speciﬁc income streams with local borrowing, it inevitably increases ﬁrm exposure to currency risk. Much of this risk can be offset by various forms of currency hedging, such Value Creation in Multinational Enterprise International Finance Review, Volume 7, 307–338 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07013-0

307

308

RICHARD A. LEWIN ET AL.

as illustrated by Click and Coval (2002) through the use of debt or derivatives with periodic rebalancing, or accomplished more traditionally through the use of investment tools such as duration (see Fabozzi, 2001; Kolb, 2002). Duration1 has long been a method in which investors or ﬁrms have balanced their debt instruments to suit the needs of their respective portfolios or strategies and as an early warning proxy to managers about the availability and changing costs of borrowing. Even though duration is widely used as a measure to quantify and control interest rate risk exposure in asset management, little attention is paid to the exposure of ﬁrm equity costs and their sensitivity to those same risks. Our approach will examine equity duration in the context of modiﬁed duration and then in the context of asset pricing. As our model uses dividends and bankruptcy risk as factors for the determination of equity duration, the model will lead to contrary results for countries with different bankruptcy or insolvency laws and different treatment and taxation of dividends; as an example the UK, Japan and the US would differ considerably on both these issues.2 In the ﬁxed income literature, Macaulay (1938) ﬁrst deﬁned what subsequently became known as Macaulay3 duration as the weighted average term to maturity of the cash ﬂow of a bond. These weights are simply the present values of each of the anticipated cash ﬂows, both coupons and principal repayment, as a percentage of the price of the bond. Macaulay chose each cash ﬂow’s contribution to the total price as its weight, which although somewhat arbitrary, remains extremely close to the measure in use today. Duration has been subsequently deﬁned by Fabozzi (2001) as ‘‘the approximate percentage change in the price of a bond (or bond portfolio) to a 100 basis point (or 1%) change in yields’’. Hence, the concept of duration expresses the amount by which price ﬂuctuates with respect to changes in underlying bond yields.4 This principally depends on the term to maturity and associated coupon rate, as well as any embedded options that may be present in more complex ﬁxed income instruments. More generally, these combined effects may be captured by a single summary measure of duration,5 which is the differential of the price yield function. Although duration is considered an invaluable tool for gauging the sensitivity of ﬁxed income portfolios to movements in underlying interest rates, fund managers essentially require the derivation of a similar model for the equity component of their investment portfolios. Such a tool could then have potential further application to the other asset classes held by fund managers, to provide for a more complete measure of interest rate risk. Traditionally the standard Dividend Discount Model6 (hereinafter referred

UK Measures of Firm-Lived Equity Duration

309

to as the DDM), provides extremely long duration estimates for equity securities. This methodology has thus been criticised by practitioners for its apparent unrealistic treatment of the observed pricing elasticity of equity securities with respect to changing discount rates.7 Moreover, any structural mismatch between assets and liabilities (particularly in a pensions management context) will necessarily increase the volatility in earnings and hence affect overall future investment returns.8 The tilt in asset allocation decisions towards more substantial emphasis on equity investment over the last few decades has created additional risks among investment managers and pension providers. These issues need to be quantiﬁed through a better understanding of equity duration9 since equities are considered well suited to long-term investment planning by providing considerable ﬁnancial growth as well as acting as a natural hedge against inﬂation.10 Indeed, equity investment now plays such a dominant role in most funded investments and savings vehicles that a better theoretical understanding by analysts for the behaviour of such instruments is of real economic signiﬁcance.

ALTERNATIVE METHODOLOGIES FOR ESTIMATING EQUITY DURATION Here initially we will analyse the traditional DDM approach to derive a suitable measure of equity duration. In the academic literature, aside from the traditional DDM approach, there have been other proposals for constructing measures of equity duration. We therefore complement our investigation with an examination of an alternative intuitive framework for calculating equity duration empirically. This concept was ﬁrst proposed by Leibowitz (1986) and subsequently considered in relation to the traditional DDM approach by Leibowitz, Sorensen, Arnott and Hanson (1989). The computation of equity duration in the seminal work of Leibowitz (1986) is based upon the observable historic correlation between stocks and bonds. This elegant and intuitive formulation relies upon the variance parameters found in conventional asset allocation procedures and provides for considerably shorter equity duration measures than those usually obtained from utilising the more orthodox DDM approach. The initial Leibowitz model provided equity duration estimates for the stock market of around two to six years, whereas the traditional DDM methodology often resulted in much longer equity duration estimates – in excess of 50 years for high growth stocks.

310

RICHARD A. LEWIN ET AL.

We will therefore begin our work by outlining this alternative methodology, where the correlation between stocks and bonds is used in conjunction with a measure of bond market duration to derive an estimate for stock market duration. In combination with the typical portfolio allocation decision between stocks and bonds, the beta of the stock component of any given portfolio can readily be used to obtain a measure of total portfolio duration, thus assisting in asset and liability matching. This is an appealing intuitive solution to the equity duration problem, pertaining to a CAPM style framework. It requires the estimation of ex ante values for the variance of stock market returns, the variance of bond market returns, and their historic correlation as follows (Leibowitz, 1986, Eq. (1))11: sE DE ¼ (1) rðE; BÞDB sB The co-movement of equity returns RE can then be expressed with respect to bond returns RB (Leibowitz, 1986, Eq. (A1))12: R~ E RF ¼ A1 þ BðR~ B RF Þ þ ~

(2a)

The regression coefﬁcient, relating DE and DB in Eq. (1), when combined ~ the with stock market returns, can thus be expressed as a linear function of d; 13 change in any given long-term benchmark yield (Leibowitz, 1986, Eq. (2)). RE ¼ A DE d~ þ ~ (2b) ~ In the Leibowitz’s (1986) paper, d could be speciﬁed as any long-term benchmark bond index, as considered relevant by the fund manager in the context of his speciﬁc risk mandate and liability schedule, and so the model relies on the effective duration of the bond market relative to yield shifts in any given benchmark proxy. The ﬁnal step necessary in combining these intuitive concepts allows for the establishment of a direct statistical relationship between the returns on the representative equity market index and changes in the representative benchmark yield. The resulting stock market durations are simply empirical estimates of actual stock price sensitivity to underlying changes in interest rates. These duration measures are by their very nature purely statistical relationships and would be subject to randomness. However, they retain econometric credibility in relating stock market returns directly to variations in long-term interest rates. In the follow-up paper, Leibowitz et al. (1989) build upon the original conceptual development in Leibowitz (1986) to derive a further measure of equity duration through price sensitivity to both the real and inﬂationary components of nominal interest rates. We derived our

UK Measures of Firm-Lived Equity Duration

311

approach by examining the models proposed by Leibowitz et al. (1989), which attempted to incorporate the impact of real interest rates and inﬂation within the DDM framework. Leibowitz et al. (1989) reformulated the DDM to obtain a total differential for stock price, or mean adjusted DDM duration measure (hereinafter referred to as DDDM), as follows14: dP @h @h ¼ DDDM 1 g þ dr DDDM 1 l þ dI DDDM dk (3) P @r @I Although our research is motivated by the concepts within these earlier papers, our own analysis differs from this previous work, by incorporating real changes in consumption and dividend growth directly into our model, to derive equity duration directly from stock market data as a by-product of asset pricing. We have re-examined the derivation of the earlier equity duration measures, such as the DDM, and instead have chosen to combine the variance parameters used in asset allocation decisions to emphasise the importance of the covariance between changes in real consumption and real dividend yields, for subsequent portfolio immunisation strategies. The use of such an approach enables us to overcome one of the standard criticisms levelled at the DDM, as cited by Hurley and Johnson (1995); that of failing to incorporate within the model realistic patterns of future dividend growth. The appropriateness of the traditional DDM approach has been further challenged by the anecdotal evidence presented by the persistent and considerable divergence in prospective broker forecasts for individual stocks. For if opinion within the brokerage profession is itself divided on any particular stock, when it comes to analysing and forecasting the fundamental earnings and dividend growth rates of ﬁrms collectively, what, indeed if any, useful conclusions can fund managers draw?15 This highlights a major practical caveat in any application of the DDM, since the more complicated task undertaken by fund managers is that of continuously reassessing a company’s longer-term performance, which would be confounded by long-term interest rate volatility itself. This is often masked by the mechanistic solutions provided for equity duration, as inculcated in the traditional DDM, since the value of the underlying corporate earnings will be implicitly affected by real changes in aggregate consumption. Leibowitz et al. (1989) modiﬁed the traditional DDM techniques at this critical stage by assuming that real interest and inﬂation rates are in fact the underlying variables that ‘‘relate changes in the interest rate and the equity risk premium to equity duration’’. Their approach re-examined the equity discount rate, dissecting it into a real interest rate, an inﬂation rate and an equity premium component.

312

RICHARD A. LEWIN ET AL.

The veriﬁcation of such a method of analysis is of critical importance, as any immunisation strategy employed against the current liability structure of portfolios (where they will inevitably include an equity component) fundamentally relies upon the conﬁdence pertaining to the duration values thereby assigned. Of course, these duration measures imply some relationship between duration and the established view of market participants in the investment community.16 By considering equities as part of the ﬁnancial assets available for investment, we recognise that the composition of a scheme is an individual investment decision, and thus a utility-based model appears to be an appropriate foundation. Therefore, we adopt a utility framework, wherein the equity risk-premium can be related to aggregate consumption. This in turn implies a link between the equity premium puzzle and reconciling the traditionally long duration estimates for equities. We consider that fund managers, as a central tenet of their investment strategy, are continually assessing and revising the risk premium for speciﬁc companies. We believe that this is not the same concept as purely mean adjusting the level of a constant risk premium term structure, as advocated in the Leibowitz et al.’s (1989) paper. The majority of the risk premium models in ﬁnance are singleperiod models of one form or another, and so it is natural for us to start from the original single-period DDM approach. The work of Damant and Satchell (1995) will therefore be used as the basis for incorporating the important macroeconomic variable of real consumption into the traditional DDM framework. When considering common stocks, we begin by assuming that the mandate of the incumbent management team is to maximise shareholders’ wealth in the context of the current business-operating environment. This notion is justiﬁed by management’s vulnerability to potential replacement by shareholders’ votes, cast at the annual general meeting, in accordance with pro rata interest in the ﬁrm’s equity capital. However, the dispersed nature of individual and even institutional shareholdings may dilute the ready exercise of these ultimate powers of administrative veto. Nonetheless, shareholders’ wealth, or the price of a common stock, can be evaluated using the ﬁrm’s expected proﬁtability or more speciﬁcally the discounted expected after-tax cash ﬂows that could be proportionately distributed to shareholders as dividend payments:

P¼

1 X t¼1

Dt ð1 þ kÞt

(4)

UK Measures of Firm-Lived Equity Duration

This is the generalised form of Williams’s (1938) seminal DDM: 1 X Dt P¼ ð1 þ r^t Þt t¼1

313

(5)

This multi-period model assumes that the future cash ﬂows are known with certainty and that the market determined interest rate is non-stochastic and constant over all time-periods. In practical terms, investors are faced with a conundrum as to which discount rate to apply across such future investment horizons. This is due to the continual re-shaping and twisting of whatever benchmark yield curve is used. Such an obvious dilemma suggests that perhaps we should be thinking of the problem in terms of an equity risk premium, implicit within the discount structure. In the following generalisation, r^t is taken as the yield on a risk free bond over the period and Pt can be considered as a per unit-time risk premium. 1 X Dt P¼ (6) t ^ ð1 þ r t þ Pt Þ t¼1 However, the more traditional approach leads us to the following generalised formula: 1 X ð1 þ gÞt P¼ D0 (7) ð1 þ kÞt t¼1 This can be rearranged thus: P¼

D0 ð1 þ gÞ ðk gÞ

(8)

And taking logs of Eq. (8), we arrive at: ln P ¼ ln D0 þ lnð1 þ gÞ lnðk gÞ

(9)

By differentiating this result, we obtain the standard DDM duration formulation (Leibowitz et al. 1989, Eq. (4)): DDDM ¼

d ln P 1 ¼ dk ð k gÞ

(10)

The preceding equation will be recognised as the DDM of Gordon (1962) that has been criticised for its lack of practical considerations, most notably the assumption that dividend growth is modelled as an inﬁnite geometric progression.

314

RICHARD A. LEWIN ET AL.

DIFFERENCES IN DURATION ESTIMATION ARISING FROM PRICE AS OPPOSED TO RETURN SENSITIVITY: THE RUBINSTEIN MODEL The theory behind our asset-pricing model investigating this apparent paradox is derived in this section.17 We have elected to build upon the work of Damant and Satchell (1995) and apply a variation of that work in addressing the classic equity duration problem. This should allow us to circumvent what we consider as the underlying practical limitations of the Leibowitz et al.’s (1989) methodology. We are therefore able to include an equity premium directly into our revised formula, with our discrete-time inter-temporal model echoing the earlier work of Rubinstein (1975): Pt ¼

1 X i¼1

E t ðDtþi Y tþi Þ ð1 þ rt;tþi Þi E t ðY tþi Þ

(11)

While the above equation is extremely general, further assumptions can be made such as those presented in Rubinstein’s (1975) Theorem 2, which imply that the state variable Yt is the marginal utility of consumption of the representative agent, i.e., Y t ¼ U 0 ðC t Þ: Our state variable thus represents changes in consumption (i.e., marginal utility). Expected dividends are based on the consumption model of expected returns on stocks. The dividend on a stock or yield is related to the consumption level in the economy, which itself inﬂuences the ﬁrm’s response to the state of economic cycle (boom and bust). Most people are well aware that recent events have reinforced the fact that dividends are inﬂuenced by factors in the marketplace, such as consumer expenditure, which have a considerable affect on corporate earnings and hence share prices reﬂect decisions ﬁrms have to make in respect of maintaining the payout level or cutting dividends, such that high growth/gearing ﬁrms are very sensitive to interest rates due to the high potential reinvestment returns on retained earnings. Firms which pay more in dividends, i.e., those ﬁrms with lower durations typically have a lower retention ratio as they are usually in industries with lower reinvestment opportunities, therefore they tend to borrow more thus resulting in higher duration as gearing rises. This gearing effect enhances returns on capital to investors.18 We argue that dividends are directly related to consumption (in that the deferral of consumption allows the investment of capital) and that the economic and consumption cycle are related to the ﬁrms well-being; being positively correlated with returns. Conversely, slow growth and stagnation in the economy will be correlated with low to negative

UK Measures of Firm-Lived Equity Duration

315

returns. In the literature, this has been broken into two components: real and inﬂationary. These are real interest and real inﬂation rates and the nominal component. Subject to the differential interest rates across borders interest rate parity, i.e., Fischer effects (1+i) (1+r) that times some E equal to the other side 1+P drive exchange rates, by real changes in inﬂation and interest and arguably consumption as a result of that. With respect to k, we are really looking at real interest and real inﬂation. That would be D/dk. So we can begin with an equation of the following form: Pt ¼

1 X E t ðDtþi Y tþi Þ i i¼1 ð1 þ rt Þ E t ðY tþi Þ

(12)

As implied in Eq. (6), rt can be split into a real yield component, deﬁned here as k, and an equity premium related to consumption. Pt ¼

1 X E t ðDtþi Y tþi Þ i i¼1 ð1 þ kÞ E t ðY tþi Þ

(13)

Thus, by taking logs of Eq. (13): ln Pt ¼ ln

1 X E t ðDtþi Y tþi Þ ð1 þ kÞi E t ðY tþi Þ i¼1

(14)

In addition, by differentiating Eq. (14), we can now deﬁne duration as follows: , 1 d ln Pt X E t ðDtþi Y tþi Þ ¼ i (15) Pt iþ1 dk E t ðY tþi Þ i¼1 ð1 þ kÞ We will now add an extra set of assumptions, in similar manner to the treatment pursued in Damant and Satchell (1995). Hence, the logs of the UK aggregate consumption and the UK aggregate dividends are deﬁned as being random walks with drift, such that they will jointly follow a bivariate random walk with drift. We further deﬁne our representative agent to have a power utility function consisting of the following assumption: Given the state of the economy, investors who place much more value on dividends would ﬁnd an impact on marginal utility as a result of small changes in consumption. Thus, value to growth investment fads ironically drive the dividend yields down. The yields fall by the real value of consumption deﬂating the nominal value of consumption utility to an investor by the incremental difference between nominal and the real. The systematic component is measured as the real component and thus can speciﬁcally be eliminated.

316

RICHARD A. LEWIN ET AL.

Assumption 1. UðC t Þ ¼

C 1b t ; 1b

U 0 ðC t Þ ¼ C b t ¼ Y t;

U 00 ðC t Þ ¼ bC tð1þbÞ

By substituting Assumption 1 into Eq. (15), we arrive at the following representation, where d represents duration: , 1 E t ðDtþi C b d ln Pt X tþi Þ ¼ i (16) d¼ Pt iþ1 dk E t ðC b i¼1 ð1 þ kÞ tþi Þ Now making the assumption that b ¼ 0, which is the risk neutral case, then Eq. (16) will become: , 1 X E t ðDtþi Þ i (17) d¼ Pt iþ1 i¼1 ð1 þ kÞ By assuming that E t ðDtþi Þ ¼ Dt ð1 þ gÞi ; where g is the expected growth rate in dividends, we can solve for Pt ¼ Dt ð1 þ gÞ=ðk gÞ; to ﬁnd again that d ¼ 1=ðk gÞ; as in Eq. (10). More generally, however, we need to specify certain assumptions about the joint probability density function (pdf) of Dt and Ct, which we will now assume follows a bivariate lognormal random walk. Assumption 2. lnðDt Þ ¼ ad þ lnðDt1 Þ þ dt

(18a)

lnðC t Þ ¼ ac þ lnðC t1 Þ þ ct

(18b)

where " 0

0

ðdt ; ct Þ N ð0; 0Þ ;

sdd scd

sdc scc

!#

Proposition 1. If Assumptions 1 and 2 hold, then the price of an asset Pt is given by: lnðPt Þ ¼ lnðDt Þ lnð1 þ k fÞ þ ln f

(19)

UK Measures of Firm-Lived Equity Duration

where

317

1 f ¼ exp ad þ sdd bscd 2

Thus, d¼

d lnðPt Þ 1 ¼ dk 1þkf

(20)

where k represents real yields and fo1 þ k: Proof. See Appendix A. The above calculations can be considered a form of the Gordon Growth Model, where f corresponds to the ð1 þ gÞ term, and represents the expected per period risk-adjusted relative dividend growth. Since b 0; we see that if scd 40; which we would expect, an increase in relative risk aversion leads to a decrease in f and hence a decrease in equity duration. However, we must be aware of the equity premium puzzle, as identiﬁed by Mehra and Prescott (1985), which suggests that for power utility, risk premia are far too low to explain empirical values. Alternatively, b needs to be quite large to generate appropriate risk premia of the magnitude estimated in mature ﬁnancial markets. Accordingly, we select to use an alternative linear risk-tolerance utility function, the HARA speciﬁcation, whereby we redeﬁne UðC t Þ as follows: Assumption 3.

¯ t Þ1b 1 b; UðC t Þ ¼ ðC t C

¯ t; b40; C t 4C

¯ t 40 C

¯ t represents base consumption or the minimum subsistence level Here C of consumption required to provide the necessities in life.19 This is analogous to what Marx considered the proletariat’s starvation bundle. We will therefore require a procedure such that Ct will always be greater than ¯ t subsistence level in all periods. We need to choose C¯ t in a systhe C tematic manner, and so we refer to the earlier work on external-habit models by Abel (1990) in formulating a difference approach. We thus deﬁne base consumption as an auto-regressive function of C t ; such that ¯ t ¼ yC t1 ; where values for y are taken from a grid of values where C 14y40: This approach is consistent with other empirical procedures in the consumption literature as adopted by Constantinides (1990) and Sundaresan (1989). Before we evaluate this model empirically, we should consider re-specifying the data generating process of Ct. Accordingly, we

318

RICHARD A. LEWIN ET AL.

will assume Ct is deﬁned as follows: Assumption 4.

X ¯ t exp a0c t þ ðC t C¯ t Þ ¼ C 0 C 0ctþj In addition,

X ¯ t b exp ba0 t b 0ctþj ðC t C¯ t Þb ¼ C 0 C c

(21)

where d

0ct Nð0; s0cc Þ and scd will become s0cd : Reformulating Proposition 1 with Assumptions 3 and 4, we expect a0c to be reduced relative to ac and s0cc to be larger than scc : Proposition 2. If Assumptions 3 and 4 hold, then the net effect of these changes on Eqs. (A.3) and (A.4) and f in particular, is that f becomes f0 . 1 where f0 ¼ exp ad þ sdd bs0cd (22) 2 whilst d 0 ¼

1 1 þ k f0

(23)

Thus if s0cd becomes larger for any given b, f0 becomes smaller and d 0 becomes smaller. The above argument effectively establishes a link between the equity premium puzzle and the large values found for equity duration. Therefore, to reinforce this point, we will adopt the following numerical example in our work. If b ¼ 2; ad ¼ 0:04 (ignoring inﬂation), sdd ¼ 0:001; scd ¼ 0:002 and k ¼ 0:05: 1 1:05 expð0:04 þ 0:0005 0:004Þ 1 1 ¼ 78 years ¼ 1:05 expð0:0365Þ 0:0128

Therefore; d ¼

Therefore supposing the average pension fund is invested in 80% equities and 20% bonds, and further assuming the duration of the diversiﬁed bond component would be at most 15 years, we can estimate the duration of the average fund’s portfolio.

UK Measures of Firm-Lived Equity Duration

319

Hence, total portfolio duration equals (0.2) 15 years +(0.8) 78 years, or 65.4 years, providing for a portfolio duration in excess of the retirement age in many countries. This would be exacerbated in the US market, where typically ad ðUSÞ ﬃ 12 ad ðUKÞ is regarded as a fair assumption.20 Emphasising the signiﬁcance of the equity premium puzzle to our previous numerical example, if b ¼ 20; which is generally thought to be too large, we might ﬁnd d 20 years, giving equity the same duration as long-term bonds. We therefore proceed to compute s0cd ; ad and sdd from our UK data series, to shed light on this anomaly. From our empirical analysis of the UK market, we may conclude reasonable values for our parameters to be taken as k ¼ 4, b ¼ 4 and where y ¼ 75%: Thus, according to our empirical calculations in Table 2, an average UK pension fund would have a total portfolio duration equalling (0.2) 15 years +(0.8) 28.5 years, or 25.8 years. This provides for a total portfolio duration that is consistent with the majority of pension funds’ contractual liabilities. Our ﬁndings therefore reinforce the traditional asset allocation decisions of pension fund providers, since the structure of the majority of pension schemes appears consistent with suitable matching strategies for their underlying assets and liabilities. This is in contrast to earlier conclusions on immunisation using the traditional DDM framework. In the empirical section that follows, we examine the relationship between our equity duration estimates and the underlying parameters in allowing for variations in y, b and k. We obtained dividend and consumption data for the UK market, as speciﬁed in the section ‘‘Empirical Estimation of Equity Duration in the UK Stock Markets’’. We are of course conscious of the important role derivatives play in the context of global fund management, and direct the interested reader to an extension of our work applicable to computing equity option duration, as contained in Appendix B. One area of empirical research that we have not addressed in our paper is the manner in which one might proceed to estimate reasonable values for b from observable market data. Instead, we refer the interested reader to Damant, Hwang, and Satchell (2000) for a discussion of the obstacles involved in such empirical calculations. Another controversial issue in the area of empirical research concerns the average expected lifetime of equity securities. The traditional DDM approach logically implies that equities are inﬁnitely lived ﬁnancial assets in regard to any prospective investment horizon, thus generating excessively long equity duration estimates. This stands in contrast to the bond markets, where ultimate available redemption periods are generally shorter.

320

RICHARD A. LEWIN ET AL.

For example, in ﬁxed income markets investors are seldom faced with liquid bond issues with maturities in excess of 30 years. More recently, however, some innovative international corporate bond issues have begun to appear with 50 year maturities; for example, the British Gas 7.25% 1994 issue, redeemable in 2044. We have thus considered whether it is reasonable to treat equities as undated instruments, when in fact there is clearly an observable corporate failure rate across all companies in the UK, listed and otherwise. This has been shown to ﬂuctuate with the economic cycle, between 1% and 2% per quarter, by Joyce and Lomax (1991). In terms of the pension fund industry, investments are predominantly in listed companies, but undoubtedly even the most risk-averse blue chip investors are exposed to potential bankruptcy risks. Former FTSE 100 companies have failed spectacularly, as exempliﬁed by Polly Peck, British & Commonwealth or Coloroll, and given the volatility in new technology constituents of the FTSE 100, bankruptcy remains a real issue with respect to pension fund portfolios. Even the most conservative of fund managers will undoubtedly retain some real residual exposure to potential insolvency and so we pose this legitimate concern for future research, but present our own initial contribution to the debate. We therefore include a modiﬁed form of our equity duration results in Table 3, where we adopt an estimate of average equity horizon as 100 years – assuming a 1% annual failure rate (a binomial distribution implies that this is not an unreasonable assumption). Proposition 3. If the life of the average ﬁrm is m years, where we consider 100 years as a reasonable assessment of average equity life expectancy, we can rework our theorems to calculate equity duration for ﬁnite-life companies, adjusting Eq. (16) accordingly: , m E t ðDtþi C b Þ d ln Pt X tþi d¼ ¼ i (24) Pt iþ1 dk E t ðC b i¼1 ð1 þ kÞ tþi Þ We similarly extend our equity duration analysis, by making use of our earlier Assumptions 1 and 2, to derive the general conditions for any m-year horizon. Thus, from our m-year survival horizon, , m X fi d ¼ Dt i (25) Pt iþ1 i¼1 ð1 þ kÞ

UK Measures of Firm-Lived Equity Duration

321

Therefore, correspondingly, we also ﬁnd that: i m X f Pt ¼ Dt 1þk i¼1

(26)

Using the result that m X

bi ¼

i¼1

b bmþ1 1b

thus 0

f B1þk

Pt ¼ Dt @

f 1þk

mþ1 1

f 1 1þk

C A

And so simplifying,

1 0 Þm fm f ð1þk m ð1þkÞ A A ¼ Dt @ Pt ¼ Dt @ 1þkf 1þkf 0

mþ1

f f ð1þk Þm

1

where f ¼ exp½ad þ 12sdd bscd ; as before, in Eq. (19). Now, however, lnðPt Þ ¼ lnðDt Þ lnð1 þ k fÞ þ lnðfÞ m lnð1 þ kÞ þ lnðð1 þ kÞm fm Þ; Thus, Duration ¼

d lnðPt Þ 1 m mð1 þ kÞm1 ¼ þ dk 1 þ k f ð1 þ k Þ ð1 þ k Þm f m

(27)

We note for convergence fo1 þ k; as m ! 1; but for ﬁnite m, no such condition is required. Under Assumption 3, f will again become f0 as in Eq. (22). Likewise, substituting this into Eq. (27) will now allow us to recalculate our equity duration estimates to be consistent with any given m-year average survival horizon for our listed equity securities. Thus, our general formulation could allow pension managers to deﬁne their own equity duration estimates based on any empirically identiﬁed value for m. An indication of the anticipated life span of a company, or portfolio of securities, could of course be determined empirically from an examination of the risk premium in corporate debt. The default probabilities, implicit in the pricing of corporate debt instruments, would thus provide a useful proxy to the expected time to

322

RICHARD A. LEWIN ET AL.

default. In any event, howsoever our reader may choose to derive an estimate of the expected lifetime of listed companies, we provide an efﬁcient method for generating subsequent equity duration estimates. We therefore illustrate an application of this approach within our results by including comparative recalculations of the corresponding equity duration parameters in Table 3, where we have taken m ¼ 100 years as a plausible illustrative assumption.

EMPIRICAL ESTIMATION OF EQUITY DURATION IN THE UK STOCK MARKETS Our empirical equity duration calculations, based upon an application to the UK market data, are reported next. Additionally, we include a further subset of results for each country arising from an extension of our work towards resolving an additional practical investment consideration: the question of a ﬁnite survival horizon for equity securities. The real possibility of bankruptcy should undoubtedly play an important part in regard to formulating long-term equity investment strategies, as applicable to the pensions industry. Lewin and Satchell (2001) explore some of the more pertinent ﬁscal and insolvency issues that are essential in underpinning any reconciliation between the differing equity yields observable across global markets. This is an important secondary issue with respect to this paper, since any subsequent empirical calculations for equity duration will be materially inﬂuenced by divergent yield characteristics and observed corporate failure rates among listed companies across the segmented international investor markets. Our empirical work is based on a time series of quarterly data derived from ICV Datastream, spanning a 37-year period from 1963 to 1999. Our market data have been generated from returns on the FT-All Share Index, which we have taken as representative of the broader UK equity market. Real yields have been computed from the UK Treasury bond data, both conventional and index-linked, to provide an empirical insight into the historical bounds for real yields. We also examined the index-linked yield gap between gilt issues over the period, which vindicates empirical bounds for k of around 3% to 8% (see Barr & Campbell, 1997). During 1998 and 1999, conventional gilts yielded barely more than 3% over comparable indexlinkers. This historically low real yield suggests that investors believe the recently independent Bank of England will, over the long run, succeed in maintaining inﬂation within the Government’s self-imposed target range of

UK Measures of Firm-Lived Equity Duration

323

2.5%. Our historical sample period, however, has witnessed successive political regimes that consistently failed to achieve anything close to this somewhat ambitious target – indeed real rates again almost reached 8% as recently as 1990. These low current yields appear consistent with an overriding expectation that the UK will enter the European single currency in the imminent future, which may yet allow conventional gilt yields to fall further in line with their European counterparts. Global interest rates for both conventional and index-linked bonds are also at historically low levels presently, increasing their vulnerability to any resurgence in commodity or wage price inﬂation, which could provoke a co-ordinated international rate rise over the longer term. Building upon the conventional approach in the empirical literature, our consumption measure has been derived from non-durable and service industry expenditure in the UK. This implicitly assumes that utility is separable across this form of consumption and other sources of utility. To allow for the non-separability in utility over time, we have incorporated a simple adjustment to allow for habit formation, or the positive effect of today’s consumption on tomorrow’s marginal utility of consumption. Our per capita adjustment factor makes use of the UK population series as compiled by the Ofﬁce of National Statistics (ONS). We have made use of an implicit price deﬂator series to obtain our inﬂation measures, and to further adjust our data to reﬂect our desire to work in real terms. By using Eqs. (19)–(21), we can derive estimates for s0cd ; ad and sdd from our UK market data. We then use our theoretical framework to empirically estimate equity duration parameters directly from UK market data, using Eqs. (23) and (27). The characteristics of our underlying empirical data set are summarised by the following Table 1 and accompanying Fig. 1. The preceding illustrations of real dividend and real consumption growth in Figs. 1 and 2 highlight the divergence in dividend growth during the 1970s and 1980s. Real dividends fell in response to the ravages of inﬂation in the 1970s, followed by a period of rapid real dividend growth in the 1980s, partly in response to the subsequent market reforms introduced under the Regan and Thatcher administrations. The graphs illustrate that re-based real consumption has been far less volatile than dividends, at least from a historical perspective.21 The next section contains our summary results in Tables 2 and 3, calculated using Eqs. (23) and (27), respectively. In these tables, we present empirical results for real yields of k ¼ 4%, for both equity duration models for each country, where we have taken m ¼ 100 years, to illustrate differences between estimates for ﬁnite and inﬁnite horizon equity securities. Our

324

RICHARD A. LEWIN ET AL.

Table 1.

Empirical Datasheet Summary Statistics for UK Data.

Quarterly Data (1963–1999) Mean Standard deviation Skewness Kurtosis Minimum Maximum

Mean Standard deviation Skewness Kurtosis Minimum Maximum

FT-All Share Yield

0.0466 0.0128 1.6774 9.8439 0.0239 0.1171

Inﬂation Rate

FT-All Share Real Return

0.0667 0.0469 1.4307 4.5766 0.0121 0.2205

0.0740 0.2260 1.5056 8.8012 0.9559 0.7495

Real Dividend Growth

Real Treasury Yield

Real Consumption Growth

0.0137 0.0614 0.1766 2.3589 0.1390 0.1348

0.0298 0.0408 0.7824 4.3458 0.1069 0.1165

0.0216 0.0192 0.2352 2.7037 0.0188 0.0677

Rebased Series (Q1 1963 =100 )

250

200

150

100

50

Real Consumption

Real Dividends

19

63 19 Q1 65 1 9 Q1 67 19 Q1 69 19 Q1 71 19 Q 1 73 19 Q1 75 1 9 Q1 77 19 Q1 79 1 9 Q1 81 19 Q1 83 1 9 Q1 85 19 Q1 87 1 9 Q1 89 19 Q 1 91 19 Q1 93 1 9 Q1 95 19 Q1 97 1 9 Q1 99 Q 1

0

Fig. 1.

Real Consumption Versus Real Dividends (1963–1999).

comparative static illustrations that appear in Figs. 2–5 provide an additional insight into the effects of changing the parameter values for b, k, y and m, respectively. Clearly s0cd has a pronounced effect on equity duration values, as we vary the level of base consumption via changes in y. Whatever

UK Measures of Firm-Lived Equity Duration

UK FT All Share (1963–1998) Equity Duration Estimates for 4% Real Yields.

Table 2. h Value 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

325

s0cd 0.01282 0.00922 0.00783 0.00708 0.00661 0.00629 0.00606 0.00588 0.00574 0.00563 0.00553 0.00546 0.00539 0.00533 0.00528 0.00524 0.00520 0.00517 0.00513 0.00511

b ¼ 1 b ¼ 2 b ¼ 3 b ¼ 4 b ¼ 5 b ¼ 6 b ¼ 7 b ¼ 8 b ¼ 9 b ¼ 10 35.99 44.85 49.57 52.55 54.60 56.11 57.26 58.17 58.90 59.51 60.02 60.46 60.83 61.16 61.45 61.70 61.93 62.13 62.32 62.48

26.69 34.12 38.26 40.92 42.80 44.18 45.25 46.10 46.79 47.37 47.85 48.26 48.62 48.93 49.21 49.45 49.67 49.87 50.04 50.20

21.25 27.57 31.18 33.54 35.22 36.47 37.43 38.21 38.84 39.36 39.81 40.19 40.52 40.81 41.06 41.29 41.49 41.67 41.84 41.99

16.75 21.99 25.04 27.07 28.52 29.61 30.46 31.14 31.69 32.16 32.55 32.89 33.18 33.44 33.67 33.87 34.05 34.21 34.36 34.49

13.85 18.32 20.95 22.72 23.99 24.95 25.70 26.30 26.79 27.21 27.56 27.86 28.12 28.35 28.55 28.73 28.89 29.04 29.17 29.29

11.83 15.71 18.03 19.59 20.72 21.57 22.24 22.78 23.22 23.59 23.91 24.18 24.41 24.62 24.80 24.96 25.11 25.24 25.36 25.47

10.34 13.77 15.84 17.24 18.25 19.02 19.62 20.11 20.51 20.84 21.13 21.37 21.58 21.77 21.94 22.08 22.22 22.34 22.44 22.54

9.20 12.27 14.14 15.40 16.32 17.01 17.56 18.00 18.37 18.67 18.93 19.16 19.35 19.52 19.68 19.81 19.93 20.04 20.14 20.23

8.29 11.08 12.77 13.92 14.76 15.40 15.90 16.31 16.64 16.92 17.16 17.37 17.55 17.71 17.85 17.97 18.08 18.18 18.27 18.36

7.56 10.10 11.66 12.72 13.49 14.08 14.54 14.91 15.22 15.48 15.70 15.89 16.06 16.21 16.33 16.45 16.55 16.65 16.73 16.81

value the reader may select as most appropriate can then be followed through Tables 2 and 3, to illustrate the effects of varying b in our example. Higher real yields have the effect of reducing equity duration estimates, as predicted from our theory. This is readily highlighted in Fig. 2, providing comparative statics for k. This is again consistent with ﬁndings on the effects of real interest-rate changes, as presented in the Leibowitz et al.’s (1989) paper. Again in the second comparative static illustration, Fig. 3, b plays an unambiguous role in altering our equity duration estimates. This further establishes our theoretical link between the historically high measures of equity duration attributable to the DDM approach and the underlying equity premium puzzle. The importance of m is emphasised in Fig. 4, portraying differences in equity duration estimates obtained from both models. This reveals the practical signiﬁcance of considering the longevity of listed equity securities themselves, by changing values for m, our chosen measure of the average life span for equity securities. Finally, Fig. 5 explores changes in y, and thereby the level of base consumption upon our equity duration estimates.

326

RICHARD A. LEWIN ET AL.

Table 3. y Value 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

s0cd 0.01282 0.00922 0.00783 0.00708 0.00661 0.00629 0.00606 0.00588 0.00574 0.00563 0.00553 0.00546 0.00539 0.00533 0.00528 0.00524 0.00520 0.00517 0.00513 0.00511

Recalculation of the UK Equity Duration Estimates for m ¼ 100 years and k ¼ 4%. b ¼ 1 b ¼ 2 b ¼ 3 b ¼ 4 b ¼ 5 b ¼ 6 b ¼ 7 b ¼ 8 b ¼ 9 b ¼ 10 29.12 32.42 33.78 34.54 35.02 35.35 35.60 35.78 35.93 36.05 36.15 36.23 36.30 36.36 36.42 36.46 36.50 36.54 36.57 36.60

24.17 28.27 30.07 31.09 31.75 32.20 32.54 32.80 33.01 33.17 33.31 33.43 33.53 33.61 33.69 33.75 33.81 33.86 33.91 33.95

20.30 24.73 26.79 27.99 28.78 29.33 29.74 30.05 30.30 30.51 30.68 30.82 30.94 31.05 31.14 31.22 31.29 31.36 31.42 31.47

16.48 20.88 23.09 24.42 25.30 25.93 26.40 26.77 27.07 27.31 27.51 27.67 27.82 27.94 28.05 28.15 28.24 28.31 28.38 28.44

13.78 17.88 20.07 21.43 22.35 23.02 23.53 23.92 24.24 24.50 24.72 24.90 25.06 25.20 25.32 25.42 25.52 25.60 25.68 25.75

11.81 15.54 17.63 18.96 19.88 20.56 21.07 21.47 21.80 22.07 22.30 22.49 22.65 22.80 22.92 23.03 23.13 23.22 23.30 23.37

10.33 13.71 15.66 16.93 17.82 18.48 18.98 19.38 19.71 19.98 20.20 20.40 20.56 20.71 20.84 20.95 21.05 21.14 21.22 21.30

9.20 12.25 14.05 15.25 16.09 16.73 17.22 17.60 17.92 18.19 18.41 18.60 18.76 18.90 19.03 19.14 19.24 19.33 19.41 19.49

8.29 11.07 12.73 13.85 14.65 15.25 15.71 16.09 16.39 16.65 16.86 17.05 17.20 17.34 17.47 17.57 17.67 17.76 17.84 17.91

7.56 10.10 11.64 12.68 13.43 13.99 14.44 14.79 15.08 15.32 15.53 15.71 15.86 15.99 16.11 16.22 16.31 16.39 16.47 16.54

50 Equity Duration Model for Firms with Infinite Lifespan

Equity Duration (in years)

40

Equity Duration Model for Firms with Lifetime m

30

20

10

0 2.00% 2.50% 3.00% 3.50% 4.00% 4.50% 5.00% 5.50% 6.00% 6.50% 7.00% Real Rate of Return (k)

Fig. 2.

Comparative Statistics for k.

UK Measures of Firm-Lived Equity Duration

327

Equity Duration Model for Firms with Infinite Lifespan

50 Equity Duration (in years)

Equity Duration Model for Firms with Lifetime m 40 30 20 10 0 1.5 2

2.5

3

3.5

Fig. 3.

4

4.5

5 5.5 6 6.5 Values for β

7

7.5

8

8.5

9

9.5

140

150

Comparative Statistics for b.

50 Equity Duration Model for Firms with Infinite Lifespan Equity Duration (in years)

40

Equity Duration Model for Firms with Lifetime m

30

20

10

0 30

40

50

60

Fig. 4.

70 80 90 100 110 Equity Horizon m (in years)

Comparative Statistics for m.

120

130

328

RICHARD A. LEWIN ET AL. 50

Equity Duration (in years)

40

30

20

10

Equity Duration Model for Firms with Infinite Lifespan Equity Duration Model for Firms with Lifetime m

0 95

85

75

Fig. 5.

65

55 45 Values for θ

35

25

15

5

Comparative Statistics for y.

CONCLUSIONS We conclude our article by discussing the further implications of our study and the scope for future research, as well as framing our analysis within the context of global portfolio management. The unsatisfactory resolution of the legitimacy of the inclusion of equity instruments into what are otherwise ostensibly dedicated pension portfolios clearly merits such further investigation. The need for higher returns, as pension liabilities become more onerous due to demographic effects, have steered investment committees towards placing higher contributions into equity holdings. But this situation appears precarious unless supported by a fundamental quantitative assessment of equity portfolio behaviour to macroeconomic events. The traditional DDM literature provides ambiguous evidence, given the practical experience of investment managers. Our work thus aims to synthesize an alternative approach, by reconciling the needs of the investment community with academic integrity proffered by the traditional DDM approach. We began our paper by acknowledging a criticism of the traditional DDM, which provides for extremely long equity duration estimates, often in excess of 50 years for high growth stocks. Empirical estimates derived from the intuitive approach of Leibowitz (1986), which relies on standard

UK Measures of Firm-Lived Equity Duration

329

regression techniques to estimate actual stock price sensitivity, result in duration estimates of between two and six years. Our work provides an alternative methodology to address this apparent paradox and establishes equity duration estimates within these polarised bounds. We proceeded to demonstrate the ﬂexibility in our approach, by considering the impact of corporate failure rates on equity duration calculations for the US and UK equity markets. Listed company failure rates can average as high as 1% to 2% per annum in the UK suggesting that there needs to be a reasonable approximation to the average equity survival horizon. We thus calibrated our model to reﬂect the real investment evidence of equities behaving as more ﬁnite horizon securities, choosing 100 years as an average lifetime benchmark. We argued that the resultant calculations are consistent with the liability structure of most pension funds, which have upper liability durations considerably below the 49-year maximum contribution period for a male worker in the UK. Both average contribution and redemption periods across all policyholders are almost certain to be of a considerably shorter duration characteristics, and our duration estimates suggest a far closer correspondence in asset matching. The degree to which pension funds can immunise a multi-asset portfolio relies more generally upon the deﬁnition of equity duration used and hence how we intend to model total portfolio duration to thereby control a fund’s overall interest rate risk. From our empirical work, we have concluded that a portfolio’s total duration, as a measure of its sensitivity to interest rates, may well be substantially different from that implied solely by calculating the duration of its ﬁxed income components. This has profound implications for fund managers attempting to implement secure immunisation strategies for mixed asset classes, as they may be potentially overlooking a fundamental risk factor, that of equity duration. Inevitably, a signiﬁcant portion of portfolio risk relies on the accuracy of the equity duration measures used, and so our work retains particular relevance to the investment community in improving such estimates. This is paramount to the trustee context of overall pension liability frameworks, around which most pension funds operate. The introduction of the minimum funding requirement under the Pensions Act 1995 has further increased the need to adopt matched investment strategies to prevent cash calls on plan sponsors during downturns in the equity markets. Our paper has relied on the US and the UK market dividend data, but we acknowledge its application in a global asset management context. The dividend yield on the UK market is dissimilar to that available in other international markets, demonstrably the US and Japan.

330

RICHARD A. LEWIN ET AL.

Historically, the central objection to adopting the traditional DDM framework as a useful tool for pension and asset management was that it tended to provide extremely long measures of equity duration, as distinct from practitioner observations. This suggested, contrary to asset management convention, that the ﬁxed income portion of a fund manager’s portfolio might well be used as a means of actually shortening the overall duration of the asset pool. This puzzling result stood in sharp contrast to fund management practice, where the equity component of a portfolio tended to be thought of at least in terms of representing much shorter durations than those implied by traditional DDMs. We thus set out to derive a new theoretical model, which whilst retaining the intuitive appeal of Leibowitz’s (1986) methodology, crucially maintains the authenticity of the more traditional approach. Our work allowed us to compute more intuitively acceptable values for the US and the UK equity duration, consistent with actual asset allocation decisions undertaken by the major pension funds. Moreover, our approach provides an inherent consistency with the actuarial assessment of equity fund valuation. This has become increasingly important through recent UK government initiatives to review the pension legislation with reference to the Myners committee inquiry recommendations and the report from the Institute of Actuaries. Pension fund actuaries assess the value of an equity investment portfolio by projecting future dividends, based upon long-term real dividend growth and real interest rate assumptions, as deﬁned in Grifﬁn (1998). Our model captures these features comprehensively, aligning the risk management tools of pension fund trustees directly with the mechanical actuarial assumptions used in monitoring their performance. We therefore anticipate considerable further investigation in this area, involving the computation of equity duration by sectors, to allow for portfolio construction and subsequent immunisation strategies. The propensity for additional funded pension provision should ensure that this area of research continues to expand in relation to the needs of a burgeoning pension fund industry. Furthermore, the proliﬁc growth in pension funds since the 1960s, coupled with the ageing demographic proﬁle in the developed world, have eroded the traditional advantages of unfunded pay-as-you-go schemes, exacerbating the need to monitor and improve pension fund performance.

NOTES 1. Our paper considers an extension of duration, a standard measure of the price volatility of a bond as deﬁned in the ﬁxed income literature, to equity securities.

UK Measures of Firm-Lived Equity Duration

331

Fund managers typically use duration in their investment schedules as a method of matching ﬁxed income instruments against known contractual liabilities. Hence immunisation, as ﬁrst recognised by Redington (1952), was deﬁned as ‘‘the investment of the assets in such a way that the existing business is immune to a general change in the rate of interest’’. Immunisation corresponds not only to the matching of assets with the present value of liabilities, as formalised by Fisher and Weil (1971), but also to the replication of the interest rates sensitivities themselves, as discussed in Bierwag, Kaufman and Toevs (1983). This creates an ostensibly immunised frontier position, or fund surplus against stochastic changes in interest rates, which consequently have no effect on ﬁnal valuation. Matching liabilities in this manner adds value directly by allowing corporate management to focus on their core business, while simultaneously allowing shareholders and future pension recipients to monitor the level of available funds more precisely. International accounting standards enforce a mark-to-market convention for pension fund reporting, thereby exacerbating the need for an improved methodology for quantifying equity duration risk. 2. We brieﬂy explore two plausible explanations to account for these differences in observed equity yields across the UK, Japan and the US markets; namely ﬁscal and bankruptcy issues. In the UK, higher dividend payouts are more common as a result of more favourable taxation treatment in the hands of recipients some of which are tax exempt institutions. In Japan, the traditionally high gearing and close relationship to the banking sector has meant that effective payout ratios for shareholders have been kept extremely low, in addition, culturally at least, many Japanese Boards would have considered it ‘an admission of failure’ to return cash to shareholders. In the US, dividends play a smaller role in the US ﬁrm treatment of earnings. Dividend irrelevance, as famously posited by Miller and Modigliani (1961) and Black (1976), or double taxation may be used as a rationale for lower or no dividend payouts. In particular, the inclusion of dividends as a factor in our model highlights the importance that any discrepancy between the yields available in the US, Japan and the UK equity markets would have on our equity duration calculations. Bankruptcy is also a risk explicitly included in our model, and corporate survivorship will be fundamentally affected by the legal bankruptcy framework in which ﬁrms operate. The UK insolvency practices tend to accentuate the position of creditors with the overreaching legal objective to facilitate the immediate dissolution of a company in order to maximize creditor recovery rates. Moreover, the regulatory and legal guidelines for insolvency in the UK tend to favour the investor, while the US bankruptcy rules through Chapter 11 protection tend to favour the ﬁrm/management – US Chapter 7 hearings typically only following on from unsuccessful Chapter 11 ﬁlings. In Japan, contrary to the more formal operation of either the US or the UK capital markets, bankruptcies of publicly traded ﬁrms are quite rare (see Aoki & Patrick, 1994). Culturally there are important alternatives to dealing with ﬁrms in ﬁnancial distresses, such as private debt restructurings often initiated by the major lenders themselves, as opposed to formal bankruptcy procedures (see Weiss, 1990). Japanese ﬁrms rely heavily on bank loans, partly due to regulatory restrictions on the issuance of corporate bonds. Moreover, loan syndication across the Japanese banking community makes it collectively vulnerable to the failure of any major industrial grouping or Keiretsu. Thus in Japan, bank dependence appears related to a reduction

332

RICHARD A. LEWIN ET AL.

in bankruptcy risk (see Hoshi, Kashyap, & Scharfstein, 1990). Since almost all large distressed corporate debtors in Japan restructure via direct bank intervention, rather than bankruptcy, it implies that the system may be less effective in redeploying capital away from declining industries. There have been recent changes to the bankruptcy laws in Japan to address such inefﬁciencies; however, it is too early to tell what effect, if any, this will have on the risk behaviour of global ﬁrms. In conclusion, the more draconian procedures in the UK tend to wind up ﬁrms (thereby eliminating shareholder equity) whereas the more ﬁrm/management friendly US Chapter 11 procedures typically allows more modest shareholder participation in the restructured entity, whereas the private debt restructuring (preferred in Japan as an alternative to formal bankruptcy) potentially leaves shareholders residual interests unchanged but almost always replaces incumbent management. 3. Independently, Hicks (1939) contemporaneously derived a similar duration measure to Frederick Macaulay. 4. Elton, Gruber, Goetzmann, and Brown (2002) identify the mapping between Macaulay’s ‘average maturity’ duration and Fabozzi’s ‘elasticity’ duration measures. 5. The interested reader is referred to Bierwag (1987) for a discussion of different duration measures. 6. For a review see Reilly and Brown (2002) or Bodie, Kane, and Marcus (2002). 7. This problem is particularly acute in the pension industry, where trustees are typically faced with very clear liability schedules arising from the nature of deﬁned beneﬁt packages on offer. It is the trustee’s irrefutable responsibility to offset such liabilities, and hence duration matching has become a recognised procedure to immunise pension assets against interest rate risk. Hence the pension industry would particularly beneﬁt from a more intuitive theoretical framework within which to match speciﬁc contractual liabilities against the overall duration of a multi-asset class portfolio. In the absence of a general formula for equity duration, the assets available for stringent portfolio dedication need to be restricted to those of the ﬁxed income variety. Pension mandates, however, are characterised by the recent tendency to increase equity investment, to underpin the long-term growth in assets as global interest rates have fallen, leaving the industry vulnerable to the adoption of an unsatisfactory hybrid arrangement of partially immunised assets and liability schedules. The absence of a suitable theoretical solution to the problem of quantifying equity investment interest rate exposure has become a more urgent issue as global competition for international pension mandates intensiﬁes. The issues raised by an ageing population, coupled with increased ﬂexibility in western labour markets, have exacerbated the requirement for greater equity holdings in the typical pension fund, to sustain beneﬁts for an expanding occupational and private pension sector. Indeed a signiﬁcant fraction of a country’s population now undertakes a variety of jobs over a lifetime, so that portability in pension design, allowing for disrupted contribution schedules, is now an essential consideration. The onus has fallen back on providers to maximise the return schedules on deﬁned contribution schemes, exacerbating the reliance on the equity component of investment returns. It is the rapidity and magnitude of these combined changes over a few decades, which are causing trustees to focus attention on better ways of measuring asset portfolio sensitivities to macroeconomic changes in the economy. Because anticipated demographic trends reinforce

UK Measures of Firm-Lived Equity Duration

333

the need for higher levels of funding to ensure sustainability in future pension provision, fund managers have adopted increasing equity exposure in ongoing schemes. 8. Even the historically unfunded UK State Earnings-Related Pension Scheme (SERPS) is tentatively moving towards a more funded basis, with the introduction of the new Individual Savings Account (ISA). Moreover, the Myners’ Report into pension reform in the UK which led to the abolition of the Minimum Funding Requirement placed greater emphasis on the accountability of trustees for selecting suitable investments in pension arrangements. 9. For completeness, our paper considers equity duration as pertinent to the pension fund management industry and not to internal asset and liability measurement referred to in the banking literature as ‘equity duration’. In two articles by Idol (1997a, 1997b), the term ‘equity duration’ is used erroneously in our view, with reference to the effective asset duration of a bank’s corporate assets and internal funding sources. Simonson (1993) also examines the interest rate exposure of a banking institution’s equity capital base, as derived from the duration of its aggregate loan and deposit portfolio, to calculate yet another albeit distinct measure again referred to as ‘equity duration’. Whilst such additional concepts exist in the broader literature, they remain distinct from the deﬁnitions in our asset pricing approach to modelling interest rate sensitivity features of listed equity securities. 10. For a discussion of the inﬂation hedging properties of equity investment see Bodie (1976). 11. Where DB is the duration of a broad based measure of the bond market, DE is the estimated duration for the equity market, sB is the standard deviation of the bond market index returns, sE is the standard deviation of the equity market returns, and r(E, B) is the correlation between the bond and equity market returns. 12. Where RF represents the risk-free rate and ~ represents non-bond market factors affecting equity returns, given that E ð~Þ ¼ 0 and E ~; R~ B ¼ 0: 13. The reader is referred to Appendix A, in Leibowitz (1986), for the complete derivation of his model. 14. Where the nominal discount rate for equities, k, is split into an inﬂation rate I, and a real return r, h is an equity premium, g represents growth rate sensitivity to real interest rates and l is an inﬂation ﬂow-through parameter. The reader is referred to Eq. (8) in Leibowitz et al. (1989) for the full derivation of this model. 15. Naturally this would depend on the correlation between broker forecasts themselves which may be inﬂuenced for example by the herd effect, groupthink or heuristic bias. 16. The interested reader is referred to the discussion of these issues as explored further in Lewin and Satchell (2001). 17. It should be noted that Johnson (1989) has also presented an alternative explanation for this anomaly, in which he emphasises that the differences arise from estimating duration from price, as opposed to return sensitivity. Whilst we recognise this important contribution, which examines the separate issues of price and reinvestment risk, such an approach does not form an integral part of our own analysis. 18. Especially in stable low growth/regulated industries such as utilities, partially explaining global LBO opportunities in these sectors, most recently that of UK airport operator BAA by Spanish Group Ferrovial.

334

RICHARD A. LEWIN ET AL.

19. Assumption 3 becomes Assumption 1, and likewise Assumption 4 will become ¯ ¼ 0: Assumption (18b) if C 20. The divergent yield characteristics across the US, Japan and the UK markets are considered in more detail in Endnote 2, which contains a brief discussion of dividend payments and related ﬁscal and bankruptcy issues affecting international equity yields. 21. For the UK, the dividend series illustrates the impact of the successive dividend control regimes applied during the 1970s. Dividend controls were ﬁrst introduced as UK policy instruments in the late 1960s and were continuously employed by successive governments from December 1972 until July 1979. The rationale behind imposing curbs on dividend payments assumes a degree of stock market inefﬁciency in the allocation of funds, and thereby seeks to increase retained earnings in order to improve the supply of organic company ﬁnancing for future investment opportunities. An important factor that contributed to the longevity of these mechanisms was their inherent consistency with the macroeconomic objectives of the day. These were essentially based on income policies that acted as restraints on the rising level of earned incomes, with dividends control tantamount to restraints on unearned income, and thus a natural ideological complement towards formulating consistent economic policy. Here we highlight the implications of these controls on our own dividend series, but refer the interested reader to a more complete account of these effects as described in Goudie and Hansen (1988).

REFERENCES Abel, A. B. (1990). Asset prices under habit formation and catching up with the Joneses. American Economic Review, 80(2), 38–42. Aoki, M., & Patrick, H. (1994). The Japanese main bank system: Its relevance for developing and transforming economics. Oxford: Oxford University Press. Barr, D. G., & Campbell, J. Y. (1997). Inﬂation, real interest rates, and the bond market: A study of UK nominal and index-linked government bond prices. Journal of Monetary Economics, 39, 361–383. Bierwag, G. O. (1987). Duration analysis. Cambridge, MA: Ballinger. Bierwag, G. O., Kaufman, G., & Toevs, A. (1983). Immunisation strategies for funding multiple liabilities. Journal of Financial and Quantitative Analysis, 18, 113–123. Black, F. (1976). The dividend puzzle. Journal of Portfolio Management, Winter 5–8. Black, F., & Scholes, M. (1973). The pricing of options and corporate liabilities. Journal of Political Economy, 81, 637–654. Bodie, Z. (1976). Common stocks as a hedge against inﬂation. Journal of Finance, 31, 459–470. Bodie, Z., Kane, A., & Marcus, A. (2002). Investments (5th ed.). New York: McGraw-Hill Irwin. Campbell, J. Y., Lo, A. W., & MacKinlay, A. C. (1997). The econometrics of financial markets. Princeton, NJ: Princeton University Press. Click, R. W., & Coval, J. D. (2002). The theory and practice of international financial management. Englewood Cliffs, NJ: Prentice Hall. Constantinides, G. M. (1990). Habit formation: A resolution of the equity premium puzzle. Journal of Political Economy, 98, 519–543.

UK Measures of Firm-Lived Equity Duration

335

Damant, D. C., Hwang, S., & Satchell, S. E. (2000). Using a model of integrated risk to assess U.K. asset allocation. Applied Mathematical Finance, 7, 127–152. Damant, D. C., & Satchell, S. E. (1995). Testing for short termism in the UK stock market: A comment. Economic Journal, 105, 1218–1223. Elton, E. J., Gruber, M. J., Brown, S. J., & Goetzmann, W. N. (2002). Modern portfolio theory and investment analysis (6th ed.). Somerset, NJ: Wiley. Fabozzi, F. J. (2001). The handbook of fixed income securities (6th ed.). New York: McGraw-Hill. Fisher, L., & Weil, R. (1971). Coping with risk of interest rate ﬂuctuations: Returns to bondholders from naı¨ ve and optimal strategies. Journal of Business, 44, 410–431. Garman, M. B. (1985). The duration of option portfolios. Journal of Financial Economics, 14, 309–315. Gordon, M. J. (1962). The investment, financing, and valuation of the corporation. New York: Homewood Irwin. Goudie, A. W., & Hansen, S. L. (1988). The effect of dividend controls on company behaviour. Applied Economics, 20, 143–164. Grifﬁn, M. W. (1998). A global perspective on pension fund asset allocation. Financial Analysts Journal, 54, 60–68. Hicks, J. R. (1939). Value and capital. Cambridge, UK: Oxford University Press. Hoshi, T., Kashyap, A., & Scharfstein, D. (1990). The role of banks in reducing the costs of ﬁnancial distress in Japan. Journal of Financial Economics, 27, 67–88. Huang, C.-F., & Litzenberger, R. H. (1988). Foundations for financial economics. Englewood Cliffs, NJ: Prentice Hall. Hurley, W. J., & Johnson, L. D. (1995). A note on the measurement of equity duration and convexity. Financial Analysts Journal, 51, 77–79. Idol, C. R. (1997b). ALM III: Practical issues with equity duration. Credit Union Executive, 37, 16–19. Idol, C. R. (1997a). Equity duration to assess interest-rate risk. Credit Union Executive, 37, 32–36. Johnson, L. D. (1989). Equity duration: Another look. Financial Analysts Journal, 45, 73–75. Joyce, M., & Lomax, J. (1991). Patterns of default in the non-ﬁnancial private sectors. Bank of England Quarterly Bulletin, November 534–537. Kolb, R. W. (2002). Futures, options, and swaps (4th ed.). London: Blackwell. Leibowitz, M. L. (1986). Total portfolio duration: A new perspective on asset allocation. Financial Analysts Journal, 42, 18–29, 77. Leibowitz, M. L., Sorensen, E. H., Arnott, R. D., & Hanson, H. N. (1989). A total differential approach to equity duration. Financial Analysts Journal, 45, 30–37. Lewin, R. A., & Satchell, S. E. (2001). The derivation of a new model equity duration. Working Paper no. 104. Department of Applied Economics, Cambridge University. Macaulay, F. R. (1938). Some theoretical problems suggested by the movement of interest rates, bond yields, and stock prices in the United States since 1856. New York: National Bureau of Economic Research. Mehra, R., & Prescott, E. C. (1985). The equity premium: A puzzle. Journal of Monetary Economics, 15, 145–161. Miller, M. H., & Modigliani, F. (1961). Dividend policy growth and the valuation of shares. Journal of Business, 34, 411–433. Redington, F. M. (1952). Review of the principle of life-ofﬁce valuations. Journal of the Institute of Actuaries, 78, 286–340.

336

RICHARD A. LEWIN ET AL.

Reilly, F., & Brown, K. (2002). Investment analysis and portfolio management (7th ed.). New York: Thomson/South-Western. Rubinstein, M. (1975). The valuation of uncertain income streams and the pricing of options. The Bell Journal of Economics, 7, 407–425. Simonson, D. G. (1993). How one leader manages A/L risk. United States Banker, 103, 56–57. Sundaresan, S. (1989). Intertemporally dependent preferences and the volatility of consumption and wealth. Review of Fiscal Studies, 2, 73–88. Weiss, L. A. (1990). Bankruptcy resolution. Journal of Financial Economics, 27, 285–314. Williams, J. B. (1938). Evaluation by the rule of present worth. In: The theory of investment value (pp. 55–75). Cambridge, MA: Harvard University Press.

APPENDIX A. DERIVATION OF THE EQUITY DURATION MODEL Under Assumptions 1 and 2, it follows that: Dt ¼ Dt1 expðad þ dt Þ and Dtþi ¼ Dt exp ad i þ

i P

! dtþj

(A.1)

j¼1

E t ðDtþi Þ ¼ Dt exp ad i þ 12 sdd i C tþi ¼ C t exp ac i þ

i P

! ctþj

j¼1

C b tþi

¼ C b t exp bac i b

E t ðC b tþi Þ

¼

C b t

so that

¼

Dt C b t

!

(A.2)

ctþj

j¼1

expðbac i þ 12b2 scc iÞ

" Dtþi C b tþi

i P

exp ðad bac Þi þ

i X j¼1

dtþj b

i X

# ctþj

(A.3)

j¼1

Using moment generating functions, the right hand side of Eq. (A.1), can be calculated so that:

i b 2 b E t Dtþi C tþi ¼ Dt C t exp ðad bac Þi þ sdd 2bscd þ b scc (A.4) 2

UK Measures of Firm-Lived Equity Duration

337

Eqs. (A.1) and (A.2) can be substituted into Eq. (A.4), thus, we ﬁnd that: i 1 X X fi f fDt Pt ; where Pt ¼ Dt i d ¼ Dt ¼ 1 þ k 1 þ kf ð1 þ kÞiþ1 i¼1 1 1 2 2 and where f ¼ exp ðad bac Þ þ sdd 2bscd þ b scc þ bac b scc 2 2 1 ¼ exp ad þ sdd bscd 2 ðA:5Þ Hence lnðPt Þ ¼ lnðDt Þ lnð1 þ k fÞ þ ln f, d lnðPt Þ 1 ¼ and d ¼ dk 1þkf

ðA:6Þ

We note that for convergence fo1 þ k:

APPENDIX B. EXTENSIONS TO COMPUTE OPTION DURATION An advantage of our procedure is that we can apply this directly to pricing options consistently with Black and Scholes (1973). Thus duration can be computed for derivatives as well, in similar treatment to Garman (1985), where this can be thought of as the usual ‘Greek’, rho (r) with respect to interest rates, where Ot is the option price and where d ¼ r=Ot : Assuming Black and Scholes holds, which we have not shown in our paper: Duration ¼

txð1 þ kÞt Fðd 2 Þ Pt Fðd 1 Þ xð1 þ kÞt Fðd 2 Þ

(B.1)

Where x is the exercise price of the call, Fð Þ is the standard normal cumulative distribution function t is the maturity, and d1 and d2 are deﬁned as follows (in similar treatment to Campbell, Lo, & MacKinlay, 1997, p. 352): log Pt =x þ k þ 12s2 t pﬃﬃﬃ d1 s t log Pt =x þ k 12s2 t pﬃﬃﬃ d2 s t

338

RICHARD A. LEWIN ET AL.

That Black and Scholes holds follows from the fact that we can assume that the option pays out Ptþt x at maturity if Ptþt 4x; and pays zero in all other periods. Therefore, Ot ¼

E t ½maxðPtþt X ; 0ÞY tþt ð1 þ kÞt E t ðY tþt Þ

(B.2)

The above, under our assumptions, is equivalent to Rubinstein’s model (as set out in Huang & Litzenberger, 1988, Chapter 5). It is straightforward to extend our option price to Assumptions 3 and 4. In fact, the formula turns out to be the same. To show the above, we would need to allow a general pricing formula to hold for all assets, in particular riskless assets. This would allow us to ‘endogenise’ interest rates and consequently ‘eliminate’ the terms in scd or s0cd from our pricing formula.

MULTINATIONALS AND EXCHANGE RATE PASS-THROUGH Alexandra Lai and Oana Secrieru ABSTRACT We examine the impact of multinational firms (MNEs) on exchange rate pass-through when an MNE engages in Cournot competition with domestic and foreign rivals. The MNE can locate its production for the foreign market domestically — intra-firm trade (IT) — or in the foreign country — international production (IP). In addition to incomplete exchange rate pass-through, we show that an MNE increases the sensitivity of domestic market prices and reduces the sensitivity of foreign market prices to exchange rate movements. Finally, IT prices are more sensitive to exchange rate movements than their IP counterparts and react in the opposite direction.

1. INTRODUCTION Multinational enterprises (MNEs) play an important role in international trade. Recently, MNEs’ pricing behaviour has attracted a lot of attention from both policy makers and researchers. To the extent that MNEs are different from other ﬁrms, one would expect that their trade and pricing patterns are also different. A few empirical studies have attempted to compare the responsiveness of MNEs’ trade prices to arm’s-length trade (AT) Value Creation in Multinational Enterprise International Finance Review, Volume 7, 339–364 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07014-2

339

340

ALEXANDRA LAI AND OANA SECRIERU

prices, but no theoretical analyses of this exist.1 In this paper, we develop a model of an MNE and attempt to shed some light on the issue of the sensitivity of MNEs’ trade prices to exchange rate movements. In particular, we examine how the presence of an MNE affects the exchange rate passthrough relative to AT, and how the MNE’s location of production matters for the exchange rate pass-through. Exchange rate pass-through into consumer prices has fallen considerably in Canada and other industrialized countries over the last two decades. One potential explanation for this decline is the increasing importance of intraﬁrm trade (IT).2 The existing evidence shows that MNEs’ exports have increasingly shifted away from AT towards IT (Rangan, 2001). The empirical evidence, otherwise limited, on the sensitivity of IT prices to exchange rate pass-through is somewhat mixed. For example, Clausing (2001) ﬁnds that IT prices are more sensitive to exchange rate movements than AT prices. Rangan and Lawrence (1999) also show that relative to AT, imports (quantities) to the U.S. by MNEs exhibit both stronger and faster responses to exchange rate changes, which the authors attribute to informational advantages arising from multinational operations. However, Pain (2002) ﬁnds the opposite for the UK: IT prices are less sensitive to exchange rate movements than AT prices. MNEs can organize their production in a number of different ways. An MNE can set up production at one location and transfer ﬁnished goods among its different branches. This is the case of IT. An MNE can also undertake production locally for sale in the local market. We refer to this scenario as international production (IP). According to UNCTAD 1998– 2004 reports, IP has increased substantially in the last decade. Finally, an MNE can locate different stages of production in different locations and this will entail IT in intermediate products. We do not examine this case as we do not model intermediate production in this paper.3 Instead, we focus on an MNE which produces a ﬁnal good for two locations, domestic and foreign, using local resources at a constant input price at the point of production. The MNE competes with local producers in both markets, but is able to treat the domestic and foreign markets as completely segmented. We consider two different production scenarios: (1) IT under which production is undertaken at the location of the MNE’s parent company (this case results in trade of the ﬁnal good between the two countries) and (2) IP in which production takes place locally in the location of sales. The distinguishing features of our model of MNE are (1) economies of scope or complementarity in the production of outputs, (2) competition in both markets, and (3) differential taxation of MNE proﬁts in the two countries.

Multinationals and Exchange Rate Pass-Through

341

The notion of economies of scope is related to intangibles such as R&D, advertising, marketing, distribution and management services that an MNE is able to share across plants avoiding duplication of such expenditures (Markusen, 1984). Desai et al. (2005) provide empirical evidence of a complementarity relationship between foreign and domestic investment for U.S. MNEs. Their study implies that MNEs combine home production with foreign production to generate a ﬁnal product at a lower cost than would be possible with production in just one country. In our model, the MNE competes in quantity (Cournot competition) with local producers in both markets. This is a departure from the MNE literature in which it is assumed that the MNE is a monopolist in both the domestic and the foreign markets. Following a stream of the industrial organization pricing-to-market literature, we choose to work under the assumption of quantity (Cournot) competition rather than that of price (Bertrand) competition (Bodnar et al., 2002, Dornbusch, 1987, Feenstra et al., 1996, Lapham, 2004). As is common in the MNE literature, the MNE in our model is subject to differential taxation of proﬁts in the two countries in which it operates. In this context, cost assignment rules (also referred to as transfer pricing rules) becomes particularly important. The common cost splitting rule we employ breaks up the cost into two: the stand-alone cost (SAC), the MNE incurs in the case where it only produces for the domestic market and the incremental cost (IC), the MNE incurs when producing for the foreign market as well. This rule is realistic and has been used in other studies of MNEs, such as Calzolari and Scarpa (2001) and Eden (1998). However, contrary to other MNE papers that focus on the regulation of MNEs by means of taxation, in our paper, we assume that proﬁt taxes are exogenous.4 We incorporate taxes in our analysis to investigate the effects of these on the sensitivity of prices with respect to the exchange rate. There exists substantial evidence of the relationship between countries’ tax rates and prices of intra-ﬁrm transactions. One such study is Clausing (2003), which shows that as the tax rates of the destination and origin countries are lower, U.S. intra-ﬁrm export prices are lower and U.S. intra-ﬁrm import prices are higher. It is therefore important to include taxes in any analyses of MNEs’ pricing behaviour. Our paper combines two strands of literature: the MNE literature and the literature on the exchange rate pass-through. To our knowledge, our analysis is novel and no other analytical studies of the effect of MNEs’ pricing behaviour for exchange rate pass-through exist. The only related paper that we came across is Hegji (2003). Hegji develops a model in which an MNE produces locally and exports some of the output to a foreign subsidiary

342

ALEXANDRA LAI AND OANA SECRIERU

which incurs additional costs. This framework allows Hegji to derive simple expressions for the exchange rate pass-through in terms of elasticities of demand and marginal costs. Our paper builds on Hegji’s by introducing a rationale for the existence of MNEs (economies of scope), competition in both markets, and considers both IT and IP as alternative means of delivering goods and services. Consistent with other studies on exchange rate pass-through under imperfect competition, our analysis shows that the exchange rate pass-through into domestic and foreign prices is incomplete. Moreover, the presence of an MNE increases the sensitivity of domestic market prices and reduces the sensitivity of foreign market prices to exchange rate movements, relative to a situation of AT. Finally, IT domestic and foreign prices are more sensitive to exchange rate movements than their IP counterparts and react in the opposite direction. Our results indicate that it is important to distinguish between the domestic market (the location of the MNE’s parent) and the foreign market (the location of the subsidiary) when looking at the sensitivity and the direction of change of the prices. This could potentially explain why some empirical studies ﬁnd the IT prices more sensitive to exchange rate movements, while others ﬁnd them less sensitive. While our approach does not allow us to directly link the observed decline in the exchange rate passthrough to the increasing importance of IT, it does, however, allow us to shed some light on how the presence of an MNE and its location of production can affect the exchange rate pass-through compared with AT. The paper is organized as follows: In Section 2 we describe the model, in Section 2.1 we derive the equilibrium outputs and exchange rate passthrough in both markets. In Section 2.2, we compare outputs and exchange rate pass-through across cases. Section 3 concludes.

2. THE MODEL In this model, we consider an MNE which sells its products in two locations, domestic (where its parent company is located) and foreign. It supplies outputs yd to its domestic market and yf to its foreign market. In each of these markets, it faces Cournot competition from local ﬁrms which produce homogeneous products. We assume that there are n identical local ﬁrms in the domestic market which each supply the quantity yd and m identical local ﬁrms in the foreign market which each supply the quantity yf. We allow the MNE to differ from its local competitors in its production

Multinationals and Exchange Rate Pass-Through

343

technology, and we assume that the domestic and foreign markets are segmented, that is, demands are independent. We consider two different cases for the location of production by the MNE, IT and IP. In the IT case, the MNE produces yd+yf in its domestic (parent company) location and transfers yf to its foreign afﬁliate for sale in the foreign market. In the IP case, the MNE produces yd domestically and yf through its foreign afﬁliate in its foreign location. The MNE’s production technology is summarized by a minimum cost function C(e, yd, yf, b), where e A {e, 1}, and the exchange rate e, is the price of the domestic currency divided by the price of foreign currency. For simplicity, we assume the domestic currency as the numeraire and express the MNE’s proﬁt and cost functions in terms of the domestic currency. b is an efﬁciency parameter which we assume is observable to the MNE and its competitors in both the domestic and foreign markets. The MNE’s cost function satisﬁes the following properties: C yi 40; C b 40; C byi 0; C yi yi 0 and C yi yj ¼ co0; a constant, for i6¼j. For e ¼ e, we also assume: Ce>0, C eyd ¼ 0 and C eyf 40: The assumption co0 reﬂects the fact that there are economies of scope for the MNE. That is, yd and yf are complements in production. Complementarity can arise when the MNE uses common inputs such as R&D, brand name and reputation. Finally, where results are clearer with speciﬁed functional forms, we assume the cost function takes the following functional form: Cð; yd ; yf ; bÞ ¼ bð yd þ yf Þ þ cyd yf ,

(1)

which satisﬁes all the required restrictions. e ¼ 1 corresponds to the case of IT (since this indicates that the cost of producing yf is incurred in domestic currency) and e ¼ e corresponds to the case of IP. The AT case corresponds to c ¼ 0 and e ¼ 1. This is the case of a purely domestic exporter that produces in its domestic location and exports its output to the foreign market. The cost assignment rule we introduce breaks up the common cost into two components. The ﬁrst component is the SAC, and it is deﬁned as the cost the MNE incurs if it produces for the domestic market alone, that is, yf ¼ 0: SACð; yd ; 0; bÞ ¼ byd .

(2)

The second component is the IC, and it is the additional cost the MNE incurs if it produces for the foreign market as well. The IC is thus: ICð; yd ; yf ; bÞ ¼ Cð; yd ; yf ; bÞ SACð; yd ; 0; bÞ ¼ byf þ cyd yf .

(3)

344

ALEXANDRA LAI AND OANA SECRIERU

The IC has two terms. The ﬁrst term beyf, is positive, while the second term cydyf, is negative since c ¼ 0, reﬂecting the economies of scope (or production complementarity) assumption. In the case of AT, c ¼ 0 and e ¼ 1, so the foreign cost for a domestic exporter is just byf>0. The MNE, however, has economies of scope and is able to produce at a lower foreign cost than the domestic exporter. For simplicity, we assume that the MNE’s competitors—domestic and foreign—have constant marginal costs of production equal to 1. We also assume that the MNE is more efﬁcient than its competitors, that is, br1. This is a reasonable assumption given the existing empirical evidence that exporters and, a fortiori, MNEs are more efﬁcient than purely domestic ﬁrms.5 Market conditions in the domestic and foreign markets are given by inverse demand functions. We assume linear demands and allow the two markets to differ along a scale (intercept) and an elasticity (slope) component. The inverse demand function in the domestic market is given by Pd ¼ abYd, where Yd ¼ yd +nyd is total output, n is the number of ﬁrms operating in d, each producing output yd. Similarly, the inverse demand function in f is Pf ¼ hkYf, where Yf ¼ yf +myf is total output, m is the number of ﬁrms operating in f, each producing output yf. We assume that a>1 and h>1, which is necessary for the existence of local ﬁrms with marginal cost of production equal to 1. The MNE is subject to proﬁt taxes, td and tf, in the domestic and foreign market, respectively. The MNE is risk neutral and its objective is to choose quantities yd and yf, to maximize global proﬁts: Pð; yd ; yf ; bÞ ¼ ð1 td ÞðPd yd SACÞ þ ð1 tf ÞðePf yf ICÞ ¼ ½ð1 td ÞPd yd þ ð1 tf ÞePf yf ½ð1 td ÞSAC þ ð1 tf ÞIC.

ð4Þ

The terms in the ﬁrst square brackets represent the after-tax revenue and the terms in the second square brackets represent the after-tax cost of the MNE. Its competitors in the domestic market simultaneously choose yd to maximize proﬁts: ðPd 1Þyd ,

(5)

while its competitors in the foreign market simultaneously choose yf to maximize proﬁts: ðPf 1Þyf .

(6)

Multinationals and Exchange Rate Pass-Through

345

2.1. Equilibrium Output and Exchange Rate Pass-through The MNE simultaneously chooses yd and yf to maximize global proﬁts, taking as given the output levels yd and yf, chosen by the domestic and foreign competitors. Assuming interior solutions, we have two ﬁrst-order conditions associated with the MNE’s problem, Eqs. (7) and (8), and two ﬁrst-order conditions associated with its competitors in the domestic and foreign markets, Eqs. (9) and (10): ð1 td ÞðP 0d yd þ Pd Þ ¼ ð1 td ÞSAC yd þ ð1 tf ÞIC yd ,

(7)

eP 0f yf þ ePf ¼ IC yf ,

(8)

P 0d yd þ Pd ¼ 1,

(9)

P 0f yf þ Pf ¼ 1.

(10)

For future reference, it is useful to have a closer look at the MNE’s ﬁrstorder conditions Eqs. (7) and (8). The left-hand side in Eq. (7) is the MNE’s after-tax marginal revenue with respect to yd, while the right-hand side is the MNE’s after-tax marginal cost with respect to yd. Both domestic and foreign taxes affect the MNE’s marginal cost of domestic output since the cost assignment rule that breaks up total cost into the SAC and IC does not take into account production complementarities in the sense that foreign output decreases the marginal cost of domestic output. It is this economies-of-scope effect that results in foreign taxes entering into the MNE’s ﬁrst-order condition for domestic output. On the other hand, both the marginal revenue and marginal cost of foreign output are reduced by (1tf) and this drops out of the ﬁrst-order condition for foreign output, which is why taxes do not inﬂuence the MNE’s equilibrium choice of foreign output directly (it does so only indirectly through changes in domestic output). The solution to the Cournot games in the domestic and foreign market is obtained by solving Eqs. (7)–(10) simultaneously. The functional forms also allow us to derive the expressions for competitors’ outputs, total market outputs and market prices as functions of the MNE’s (or arms-length exporter’s) outputs, which we present in Lemma 1. Lemma 1. Equilibrium quantities, prices and demand elasticities expressed in terms of the MNE’s quantities, yd and yf, are given by yd ¼

h 1 kyf a 1 byd ; yf ¼ , bðn þ 1Þ kðm þ 1Þ

(11)

346

ALEXANDRA LAI AND OANA SECRIERU

Yd ¼

yf nða 1Þ y mðh 1Þ þ d ; Yf ¼ þ , bðn þ 1Þ n þ 1 kðm þ 1Þ m þ 1

(12)

m þ h kyf n þ a byd ; Pf ¼ . nþ1 mþ1

(13)

Pd ¼

Proof. The proof is relegated to Appendix A. Denote by E Pi ;e ¼ @ðln Pi Þ=@ðln eÞ the exchange rate pass-through into market i prices, for i ¼ {d, f }. Proposition 1 gives the general expressions for the pass-through elasticities. Proposition 1. The exchange rate pass-through into foreign and domestic prices are E Pi ;e ¼ E yi ;e ZPi ;yi ; i 2 fd; f g

(14)

where E yi ;e ¼ @ðln yi Þ=@ðln eÞ is the elasticity of the MNE’s output in country i with respect to the exchange rate, ZPi ;yi ¼ @ðln Pi Þ=@ðln yi Þ is the elasticity of the market demand i, i A{d, f } with respect to the MNE’s output in market i.6 Proof. The proof is relegated to Appendix B. Proposition 1 shows the exchange rate pass-through mechanism by breaking it down into two factors. First, a change in the exchange rate affects the MNE’s outputs. Second, the change in outputs affects the elasticity of market demand with respect to the MNE’s outputs. In the pricing-to-market literature, the exchange rate pass-through is typically expressed in terms of mark-ups over marginal costs. We can also express the exchange pass-through elasticities as a function of ﬁrms’ cost elasticities with respect to the exchange rate and show that exchange rate pass-through is proportional to the elasticity of marginal cost with respect to the exchange rate. E Pd ;e ¼

ð1 tf Þ=ð1 td ÞC yd @ ln C yd , a þ nC yd þ ðtf td Þ=ð1 td ÞSAC yd þ ð1 tf Þ=ð1 td ÞC yd @ ln e (15) E Pf ;e ¼

C yf =e h þ mC yf þ C yf =e

@ lnðC yf =eÞ @ ln e

.

With some manipulation, we can derive the expressions in Lemma 2.

(16)

Multinationals and Exchange Rate Pass-Through

347

Lemma 2. The price elasticities with respect to the MNE’s outputs are: ZPd ;yd ¼

byd 40, a þ n byd

(17)

ZPf ;yf ¼

kyf 40. h þ m kyf

(18)

Proof. The proof is relegated to Appendix C. Due to economies of scope in the MNE’s cost function (c6¼0), the price elasticities of the MNE’s domestic and foreign output are related. Lemma 3 gives the relationship between the price elasticities of the MNE’s domestic and foreign output. Lemma 3. The MNE’s output elasticities in the domestic and foreign markets are related: 1 t f n þ 1 yf j E , 1 td bðn þ 2Þ yd yf ; e " # mþ1 yd b j c ð1 þ E yd ;e Þ þ I , ¼ ekðm þ 2Þ yf yf E jyd ;e ¼ c

E jyf ;e

(19)

(20)

where jA{AT, IT, IP} and Ie is an indicator function equal to one for e ¼ 1 and zero for e ¼ e. Proof. The proof is relegated to Appendix D. Lemmas 2 and 3 allow us to predict that the MNE changes the quantities it supplies to both the domestic and foreign market in the same direction in response to an exchange rate shock. Consequently, prices in both markets move in the same direction when an MNE is present. Under AT, however, domestic market prices are invariant to exchange rate movements. These results are summarized in Proposition 2. Proposition 2. The following is true: (1) E jyd ;e and E jyf ;e have the same sign, for c o 0, j A{IT, IP}. j j (2) E Pd ;e and E Pf ;e have the same sign, for c o 0, j A{IT, IP}. AT (3) E AT Pd ;e ¼ 0 and E Pf ;e o0:

Lemma 3 implies that for co0, E jyd ;e and E jyf ;e ; j A{IT, IP} have the same sign. Since elasticities ZPi ;yi are positive by deﬁnition, Lemmas 2 and 3

348

ALEXANDRA LAI AND OANA SECRIERU

together imply that the exchange rate pass-through into domestic and foreign prices also have the same sign. We can also see from Lemma 3 that under AT, which corresponds to c ¼ 0 and e ¼ 1 the exporter’s output, yd is AT invariant to the exchange rate, that is, E AT yd ;e ¼ 0; while E yf ;e o0: This implies that the exchange rate pass-through into domestic and foreign prices is AT E AT Pd ;e ¼ 0 and E Pf ;e o0; respectively. Solving Eqs. (7)–(10) for yjd and yjf ; we obtain Proposition 3. Proposition 3. Equilibrium quantities in the domestic and foreign markets respectively are yjd ¼

1 ð1 td Þekðm þ 2Þ½a þ n bðn þ 1Þ D ð1 tf Þcðn þ 1Þ½eðh þ mÞ bðm þ 1Þ ,

ð21Þ

1 ð1 td Þbðn þ 2Þ½eðh þ mÞ bðm þ 1Þ D ð1 td Þcðm þ 1Þ½a þ n bðn þ 1Þ ,

ð22Þ

and yjf ¼

for j A{AT, IT, IP}, where D ð1 td Þekbðn þ 2Þðm þ 2Þ ð1 tf Þc2 ðn þ 1Þðm þ 1Þ40:7 We can easily show that ydj oða þ nÞ=2b and yfj oðh þ mÞ=2k; which implies that ZPi ;yi o1; for iA{d, f }. For future reference, it is useful to obtain the comparative statics properties of the equilibrium output levels, ydj and yfj : They are summarized in Lemma 4. Lemma 4. The MNE’s output yij ; iA{d, f }, jA{AT, IT, IP} is increasing in a, h, n, m and td, and decreasing in b, k, b, c and tf: @yij @y j @y j @y j 0; i 0; i 0; i 0, @a @h @b @k

(23)

@yij @y j @y j @y j 0; i 0; i 0; i 0, @n @m @td @tf

(24)

Multinationals and Exchange Rate Pass-Through

349

@y ij @y j o0; i 0, @b @c

(25)

@y ij @yIP 0; j 2 fAT; ITg; i o0. @e @e

(26)

To understand the intuition behind these comparative statics results, we rewrite the MNE’s ﬁrst-order conditions in terms of its own outputs for the domestic and foreign markets (i.e., we substitute in the reaction functions of its domestic and foreign competitors) as the following: ð1 td Þ

n þ a bðn þ 2Þydj ¼ ð1 td Þb þ ð1 tf Þcy fj , nþ1 e

h þ m kðm þ 2Þyfj ¼ b þ cydj , mþ1

(27)

(28)

for j A{IT, IP}. The marginal revenue is on the left-hand side while the marginal cost is on the right-hand side. Notice that while marginal revenue functions are independent across markets, marginal cost functions are not for the MNE (co0). The comparative statics results in Lemma 4 are not surprising, with the exception of those with respect to the tax rates, td and tf. As expected, the MNE’s domestic and foreign output increases with market size, decreases with the elasticity of demand and increases with competition in the two markets. An increase in the MNE’s efﬁciency, that is, lower b, increases output levels in the two markets. The output levels are also increasing in the degree of complementarity c, that is, the lower c (the more negative) the higher the output level y ij : The more surprising result is that domestic and foreign taxes have opposite effects on the MNE’s output levels in both markets. An increase in the domestic tax rate td reduces both the marginal revenue on the left-hand side of Eq. (27) and the marginal cost on the right-hand side of Eq. (27). The MNE’s ﬁrst-order condition for its foreign market output is unaffected by a change in td. The net effect of the reduction in marginal revenue and marginal cost of domestic output production is to increase domestic output. This occurs because the effect on marginal cost, given by b is greater than the effect on marginal revenue ðn þ a bðn þ 2Þydj Þ=ðn þ 1Þ ¼ ð1 ZPd ;yd ÞPd : To see why this kis so, consider a re-statement of Eq. (27): ð1 td Þ j ð1 ZPd ;yd ÞPd b ð1 tf Þc ¼ 0: Since the second term on the left-hand

350

ALEXANDRA LAI AND OANA SECRIERU

side is positive (due to co 0), the expression ð1 ZPd ;yd ÞPd b must be negative. The increase in domestic output reduces the marginal cost of producing foreign output due to complementarities in production; hence, foreign output also increases. On the other hand, an increase in the foreign tax rate tf, increases the marginal cost of domestic output by cy fj : The MNE responds to this by decreasing domestic output. This also leads to a decrease in the foreign output. An increase in the foreign proﬁt tax tf, thus results in a decrease of output in both markets. 2.2. Comparison Across Cases Since the purpose of this paper is to examine how the presence of an MNE and how the location of the MNE’s production affect exchange rate passthrough, we need to compare equilibria across the cases of AT when an MNE is absent, and IT and IP when an MNE is present. Proposition 4. Domestic and foreign output by an MNE is higher compared with AT, assuming that e is not too large: IP AT yIT d ; yd 4yd ,

(29)

AT yIT f 4yf ,

(30)

AT yIP f 4yf ; e 1 and e41 but not too large:

(31)

If e o 1, both the MNE’s domestic and foreign outputs increase with a IT shift from IT to IP; that is, yIP i yi ; i ¼ d, f. If e>1, the reverse is true, IP IT that is, yi yi : Proof. This is can be easily shown by directly comparing y ij ; for i A{d, f}, j A{AT, IT, IP}. The intuition is as follows. When eo1, the average marginal cost of the MNE falls with IP and therefore, output increases. This increase is true for both the domestic and foreign markets due to the economies of scope the MNE enjoys. The opposite argument holds if e>1. Proposition 5. In response to an increase in exchange rate e, the MNE’s outputs increase under IT while they decrease under IP: IP E IT yi ;e o0oE yi ;e ; i 2 fd; f g.

(32)

Multinationals and Exchange Rate Pass-Through

351

Proof. The proof is relegated to Appendix F. Since ZPi ;yi 40; i 2 fd; f g; market prices fall in response to an increase in exchange rate E IT Pi ;e o0 (i 2 fd; f g) for an MNE with IT. On the other hand, under an MNE with IP, market prices increase in response to an increase in exchange rate E IP Pi ;e 40; for i 2 fd; f g: The next proposition compares exchange rate pass-through under the different cases of an arms-length exporter (AT), and MNE with IT and an MNE with IP. Proposition 6. The exchange rate pass-through into domestic and foreign prices is incomplete under AT or an MNE, that is, E jPi ;e o1; for i A{d, f} and j A{AT, IT, IP}. Moreover, (1) Introducing an MNE increases the exchange rate pass-through into domestic prices and reduces the exchange rate pass-through into foreign prices, regardless of the MNE’s location of production: j AT j 0 ¼ E AT Pd ;e oE Pd ;e ; E Pf ;e 4E Pf ;e

(33)

for co0 but close to zero, where j A{IT, IP} corresponds to the case with an MNE. (2) The exchange rate pass-through is higher under IT than under IP: IT E 4E IP , Pi ;e Pi ;e

(34)

for i A{d, f}, co0 but close to zero and e>b/(m+2). Proof. The proof is relegated to Appendix G. Table 1 summarizes the results in Propositions 4–6. 6 shows that exchange rate pass-through is incomplete, Proposition AT E Pf ;e o1; due to imperfect competition, a common result from the industrial organization exchange rate pass-through literature. Proposition 6 implies that introducing an MNE (1) increases the exchange rate pass-through into domestic prices and decreases the exchange rate passthrough into foreign prices and (2) pass-through is always higher under IT than under IP. The intuition is as follows. An AT exporter’s domestic

352

ALEXANDRA LAI AND OANA SECRIERU

Table 1.

Comparison Across Cases.

eo1

e>1

IT IP yAT d oyd oyd AT IT yf oyf oyIP f

IP IT yAT d oyd oyd AT IP yf oyf oyIT f

IP ZIT Pd ;yd ZPd ;yd

IP ZIT Pd ;yd ZPd ;yd

ZIT Pf ;yf

IP ZIT Pf ;yf ZPf ;yf

ZIP Pf ;yf

IP E IT yd ;e o0oE yd ;e IP E IT yf ;e o0oE yf ;e

IT IP 0 ¼ E AT Pd ;e oE Pd ;e oE Pd ;e o1 IT AT 0 ¼ E IP Pf ;e oE Pf ;e oE Pf ;e o1

output is invariant to exchange rate changes and, therefore, the exchange rate pass-through into domestic prices is zero under AT.8 This is obvious from the ﬁrst-order condition Eq. (27) for c ¼ 0. An MNE’s domestic output is, however, affected by exchange rate changes because of linkages between the domestic and foreign markets due to economies of scope, that is, co0. The exchange rate pass-through into domestic prices is, thus, positive in the presence of an MNE and, consequently, greater than under AT. On the other hand, the exchange rate pass-through into foreign prices is lower in the presence of an MNE than under AT. The reason is that the MNE is more ‘‘diversiﬁed’’ compared with the AT exporter. An MNE adjusts both foreign and domestic production in response to an exchange rate change due to linkages between the two markets resulting from economies of scope. We can easily see this from the ﬁrst-order conditions Eqs. (27) and (28). A change in the exchange rate e, affects the marginal revenue of producing for the foreign market and the MNE is going to adjust foreign output accordingly. This, in turn, affects the marginal cost of domestic production, that is, the left-hand side of Eq. (27), and the MNE adjusts domestic output as well. This is no longer the case for an arm’s-length exporter who adjusts only its foreign output in response to a change in the exchange rate. Since the MNE has two degrees of freedom, it does not have to adjust foreign output as

Multinationals and Exchange Rate Pass-Through

353

much as the AT exporter. The output adjustment is then passed through into prices via changes in the elasticity of market demands with respect to output. This, in turn, renders foreign prices less sensitive to exchange rate movements in the presence of an MNE compared with AT. For part (2) of Proposition 6, the intuition is as follows. Under IP, a change in the exchange rate makes production in one market more expensive, and makes it less expensive in the other market. These two effects are offsetting. Under IT, however, a change in the exchange rate makes domestic production more or less expensive without an offsetting effect in the foreign market. As a consequence, the MNE must adjust domestic output by a greater extent under IT than under IP. The MNE is, thus, more ‘‘diversiﬁed’’ under IP than under IT and output is less sensitive to exchange rate movements under IP than under IT. The output adjustment translates into price adjustment via changes in the elasticity of market demands with respect to output. As a result, IT prices are more sensitive to exchange rate movement than their IP counterparts. Analytically, we can see this from the ﬁrst-order conditions of Eqs. (27) and (28). Under IP, which corresponds to co0 and e ¼ e, a change in the exchange rate e, affects both marginal revenue and marginal cost of foreign production in Eq. (28), since we use the domestic currency as numeraire. Under IT, which corresponds to co0 and e ¼ 1, a change in the exchange rate affects only the marginal revenue of foreign production in Eq. (28). The MNE must adjust foreign production more under IT than under IP because there is no offsetting marginal cost effect. To complete our analysis of how the presence of an MNE affects exchange rate pass-through, the following three propositions examine how exchange rate pass-through is affected by taxes, competition and exchange rate movements. The results are obtained by taking partial derivatives of pass-through elasticities and direct inspection of these derivatives. Proposition 7. Domestic taxes increase while foreign taxes reduce the exchange rate pass-through in both domestic and foreign markets. Neither domestic nor foreign taxes affect the exchange rate pass-through in both domestic and foreign markets under AT: @E jPi ;e 40, (35) @td @E jPi ;e @tf

o0,

(36)

354

ALEXANDRA LAI AND OANA SECRIERU

@E AT Pi ;e ¼ 0, @tk

(37)

i; k 2 fd; f g; j 2 fIT; IPg: Proposition 7 shows that domestic and foreign taxes have asymmetric effects on exchange rate pass-through. This result emphasizes once more the importance of distinguishing between the domestic and foreign markets when analysing exchange rate pass-through. Proposition 8. When the MNE’s domestic market becomes more petitive, exchange rate pass-through falls for all cases but one, the through to foreign prices under IP: @E jPd ;e o0; j 2 fIT; IPg, @n j IP @E AT @ E @ E Pf ;e Pf ;e Pf ;e ¼ 0; 40; o0. @n @n @n

compass-

(38)

(39)

On the other hand, when the MNE’s foreign market becomes more competitive, exchange rate pass-through increases for domestic prices and falls for foreign prices: @E jPd ;e 40; j 2 fIT; IPg, (40) @m @E jPf ;e o0; j 2 fAT; IT; IPg. (41) @m Our results regarding the effects of competition on exchange rate passthrough when an MNE is present contrasts with results of the pricingto-market literature on exchange rate pass-through which feature AT only and competition with other exporters in the domestic market. In that literature, an increase in competition in the domestic market always increases the sensitivity of domestic prices to exchange rate movements. In our context, the price sensitivity of a given market increases only in some cases, and, in those cases, always in response to an increase in the competition the MNE faces in the other market.

Multinationals and Exchange Rate Pass-Through

355

Proposition 9. An appreciation/depreciation of the domestic currency leads to a decrease/increase in pass-through into both domestic and foreign prices: @E jPi ;e 0; i 2 fd; f g; j 2 fAT ; IT; IPg. (42) @e

3. CONCLUSIONS In this paper, we develop a model that allows us to look at the effects of MNEs’ pricing behaviour on the sensitivity of prices to exchange rate movements. Our simple model allows us to draw some powerful results. We ﬁrst ﬁnd that the exchange rate pass-through into domestic and foreign prices is incomplete. We also show that IT domestic prices are more sensitive to exchange rate movements than AT prices, whereas IT foreign prices are less sensitive to exchange rate movements than AT prices. Moreover, IT domestic and foreign prices are more sensitive to exchange rate movements than their IP counterparts. Our results are consistent with some of the empirical evidence on exchange rate pass-through. First, the empirical evidence is somewhat mixed with respect to the sensitivity of IT prices to exchange rate movements compared with that of AT prices. Our results imply that it is important to distinguish between the domestic market, that is, the location of the MNE’s parent, and the foreign market, that is, the location of the subsidiary, when looking at the sensitivity of the IT prices versus AT and IP prices. This could potentially explain why some empirical studies ﬁnd the IT prices more sensitive to exchange rate movements, while others ﬁnd them less sensitive. Second, the empirical evidence also shows that the exchange rate passthrough into U.S. import prices is lower than into Canadian import prices. One explanation in our model could be that foreign MNEs choose to deliver goods into the U.S. mainly by IP and into Canada mainly by IT. This is, of course, something that would have to be tested empirically. Our analysis also shows that exchange rate pass-through does not always increase with competition which is somewhat in contradiction with the pricing-to-market literature. The pricing-to-market literature, however, considers only arm’s-length exporters which produce at constant marginal cost. This is no longer the case in our model of the MNE which could explain why we are obtaining the mixed results.

356

ALEXANDRA LAI AND OANA SECRIERU

In this paper, we abstract from intermediate production and assume the MNE produces a homogeneous ﬁnal good for two locations. In order to better compare our results with those in the standard industrial organization models of exchange rate pass-through, we need to consider intermediate production in our future work. This is important as an increasing proportion of MNEs’ trade is accounted by intermediate goods. Relaxing the homogeneity assumption and introducing product differentiation would also allow us to look explicitly at transfer prices. In our model, the MNE only competes with purely domestic and foreign ﬁrms and we do not consider competition from exporters. Extending the analysis to include competition from exporters is left for future research.

NOTES 1. In arm’s length trade, an exporter sells a product to a local ﬁrm that is responsible for the retailing of that product. The exporter has no access to the local market. 2. Other possible explanations for the decline of the exchange rate pass-through suggested in the literature are (1) the transition to a low-inﬂation environment in industrialized countries; (2) substitution by consumers to lower-priced items, substitution by retailers to lower-cost suppliers, and greater productivity improvements by retailers during periods of depreciation of the national currency; (3) the shift in the composition of imports in industrialized countries towards sectors with lower pass-through. A Canadian study by Lapham (2004), shows that both lower inﬂation and restructuring in the retail sector explain the observed lower pass-through to consumer prices in Canada. 3. An increasing proportion of intra-ﬁrm trade by MNEs is accounted for by intermediate inputs and we hope to examine intermediate production in future research. 4. The regulation of MNEs literature is extensive. For more on this, see, for example, Calzolari and Scarpa (2001), Calzolari (2004), Bond and Gresik (1996), Dasgupta and Sengupta (1995) and Gresik and Nelson (1994). 5. See, for example, Bernard and Jensen (1995), Bernard and Jensen (1997a, 1997b, 1997c), Richardson and Rindal (1995), Bernard and Wagner (1997), Bernard, Jensen, and Wagner (1997), and Aw and Hwang (1995). 6. The elasticity of the market demand i, i A {d, f} with respect to the MNE’s output in market i is positively related to the market demand elasticity: 1 y byd ZPd ;yd ¼ Z , d Z ¼ n þ 1 Y d d byd þ nða 1Þ d ZPf ;yf ¼

yf kyf 1 Z , Zf ¼ kyf þ mðh 1Þ f m þ 1 Yf

where Zi ¼ @ðln Pi Þ=@ðln Y i Þ; i ¼ d, f.

Multinationals and Exchange Rate Pass-Through

357

7. See Appendix E for a proof and expressions for yi, Yi, Yi, Pi, and Pi, for iA{d, f}. 8. This is an artifact of the assumption that there are no exporters based in the foreign country. Modifying our model to allow for competition by other exporters will lead to a non-zero exchange rate pass-through under AT, perhaps a more realistic scenario.

ACKNOWLEDGMENTS The authors gratefully acknowledge comments from Richard Dion, Sharon Kozicki, Robert Lafrance, Larry Schembri and seminar participants at the Bank of Canada. All errors and omissions are our own.

REFERENCES Aw, B., & Hwang, A. (1995). Productivity and the export market: A ﬁrm-level analysis. Journal of Development Economics, 47(2), 313–332. Bernard, A., & Jensen, J. (1995). Exporters, jobs, and wages in U.S. manufacturing, 1976–1987. Washington DC: Brookings Papers on Economic Activity, Microeconomics. Bernard, A., & Jensen, J. (1997a). Exporters, skill-upgrading, and the wage gap. Journal of International Economics, 42, 3–31. Bernard, A., & Jensen, J. (1997b). Inside the U.S. export boom. (Working Paper No. 6438). National Bureau of Economic Research. Cambridge, MA. Bernard, A., & Jensen, J. (1997c). Why some firms export: Experience, entry costs, spillovers, and subsidies. (Photocopy). New Haven, CT: Yale University. Bernard, A., Jensen, J., & Wagner, J. (1997). The good go abroad. In: S. Laarsonen (Ed.), The evolution of firms and industries. Helsinki, Finland: Statistics Helsinki. Bernard, A., & Wagner, J. (1997). Exports and success in German manufacturing. Weltwirtschafliches Archiv, 133(1), 134–157. Bodnar, G. M., Dumas, B., & Marston, R. C. (2002). Pass-through and exposure. Journal of Finance, 57(1), 199–231. Bond, E. W., & Gresik, T. A. (1996). Regulation of multinational ﬁrms with two active governments: A common agency approach. Journal of Public Economics, 59, 33–53. Calzolari, G. (2004). Incentive regulation of multinational enterprises. International Economic Review, 45(1), 257–285. Calzolari, G., & Scarpa, C. (2001). Regulation at home, competition abroad: A theoretical framework. (Photocopy). FEEM Working Paper No. 39. 2001, Milano, Italy. Clausing, K. A. (2001). The behavior of intrafirm trade prices in U.S. international prices data. (Working Paper No. 333). Bureau of Labor Statistics. Washington, DC. Clausing, K. A. (2003). Tax-motivated transfer pricing and U.S. intraﬁrm trade prices. Journal of Public Economics, 87, 2207–2223. Dasgupta, S., & Sengupta, K. (1995). Optimal regulation of MNEs and government revenues. Journal of Public Economics, 58, 215–234.

358

ALEXANDRA LAI AND OANA SECRIERU

Desai, M. A., Foley, C. F., Hines, J. R. J. (2005). Foreign direct investment and the domestic capital stock. American Economic Review, 95(2), 33–38. Papers and Proceedings. Dornbusch, R. (1987). Exchange rate and prices. American Economic Review, 77(1), 93–106. Eden, L. (1998). Taxing multinational: Transfer pricing and corporate income taxation in North America. Toronto, Canada: University of Toronto Press. Feenstra, R. C., Gagnon, J. E., & Knetter, M. M. (1996). Market share and exchange rate pass-through in world automobile trade. Journal of International Economics, 40, 187–207. Gresik, T. A., & Nelson, D. R. (1994). Incentive compatible regulation of a foreign-owned subsidiary. Journal of International Economics, 36, 309–331. Hegji, C. E. (2003). A note on transfer prices and exchange rate pass-through. Journal of Economics and Finance, 27(3), 396–403. Lapham, B. (2004). Canadian restructuring and exchange rate pass-through. (Photocopy). Ontario, Canada: Queen’s University. Markusen, J. R. (1984). Multinational, multi-plant economies, and the gains from trade. Journal of International Economics, 16, 205–226. Pain, N. (2002). Networks, multinational firms and U.K. export performance. (Photocopy). National Institute for Economic and Social Research. London, UK. Rangan, S. (2001). Explaining tranquility in the midst of turbulence: U.S. multinationals’ intrafirm trade, 1966–1997. (Working Paper No. 336). Bureau of Labor Statistics. Washington, DC. Rangan, S., & Lawrence, R. Z. (1999). Search and deliberation in international exchange: Learning from multinational trade about lags, distant effects, and home bias. (Working Paper No. 7012). National Bureau of Economic Research. Cambridge, MA. Richardson, J., & Rindal, K. (1995). Why exports really matter!. Washington, DC: The Institute for International Economics and the Manufacturing Institute. UNCTAD. (1998–2004). World investment report. New York: United Nations Program on Transnational Corporations.

APPENDIX A. PROOF OF LEMMA 1 Using Pd ¼ a bY d ; Pf ¼ h kY f ; Y d ¼ yd þ nyd ; and Y f ¼ yf þ myf in the ﬁrst-order conditions (9) and (10), we obtain: byd þ bðn þ 1Þyd ¼ a 1,

(A.1)

kyf þ kðm þ 1Þyf ¼ h 1.

(A.2)

Eqs. (A.1) and (A.2) can be solved for yd and yf in terms of yd and yf, respectively. The solutions give Eq. (11). Substituting Eq. (11) into the domestic and foreign aggregate output gives Eq. (12).

Multinationals and Exchange Rate Pass-Through

359

APPENDIX B. PROOF OF PROPOSITION 1 The proof is straightforward: @Pi e @Pi yi @y e ¼ E Pi ;e ¼ i ¼ E yi ;e ZPi ;yi . @e Pi @yi Pi @e yi

(B.1)

APPENDIX C. PROOF OF LEMMA 2 The price elasticities with respect to the MNE’s outputs are by deﬁnition: ZPi ;yi ¼

@Pi yi . @yi Pi

(C.1)

Since Pd ¼ b and Pf ¼ k we can use Eq. (12) to obtain the result in Lemma 2.

APPENDIX D. PROOF OF LEMMA 3 The ﬁrst-order condition with respect to the MNE’s domestic output, Eq. (7) can be re-written as: 1 tf 1 tf a þ n bðn þ 2Þ ¼ SAC yd þ IC yd ¼ b þ cy . nþ1 1 td 1 td f

(D.1)

Differentiating both sides of Eq. (D.1) with respect to e and multiplying by e/yd, gives: 1 tf @yf e yf bðn þ 2Þ @yd e ¼ c , (D.2) nþ1 @e yd 1 td @e yf yd which can also be written as: E yd ;e ¼ c

1 t f n þ 1 yf E y ;e , 1 td bðn þ 2Þ yd f

(D.3)

which proves Eq. (19). The proof for Eq. (20) is similar. The ﬁrst-order condition with respect to the MNE’s foreign output, Eq. (8) can be rewritten as: e

h þ m kðm þ 2Þyf ¼ b þ cyd . mþ1

(D.4)

360

ALEXANDRA LAI AND OANA SECRIERU

Differentiating both sides of Eq. (D.4) with respect to yf, multiplying by both sides by e/yf, and re-arranging terms gives Eq. (20).

APPENDIX E. EQUILIBRIUM QUANTITIES AND PRICES Solving Eqs. (7)–(10) yd, yd, yf, and yf using the expressions in Lemma 1 we obtain: (1) The equilibrium quantities and prices in the domestic market yjd ¼

1 fð1 td Þekðm þ 2Þ½a þ n bðn þ 1Þ D ð1 tf Þcðn þ 1Þ½eðh þ mÞ bðm þ 1Þg,

yjd ¼

Y jd ¼

1 ð1 td Þekðm þ 2Þða 2 þ bÞ D þ ð1 tf Þc½eðh þ mÞ bðm þ 1Þ a1 ð1 tf Þc2 ðm þ 1Þ , b

1 ð1 td Þekðm þ 2Þ½aðn þ 1Þ n b D ð1 tf Þc½eðh þ mÞ bðm þ 1Þ a1 2 , ð1 tf Þc nðm þ 1Þ b

Pjd ¼

ðE:1Þ

ðE:2Þ

ðE:3Þ

1 fð1 td Þekbðm þ 2Þ½a þ n þ b D þ ð1 tf Þcb½eðh þ mÞ bðm þ 1Þ ð1 tf Þc2 ðm þ 1Þða þ nÞg.

ðE:4Þ

Multinationals and Exchange Rate Pass-Through

361

(2) The equilibrium quantities and prices in the foreign market yjf ¼

1 fð1 td Þbðn þ 2Þ½eðh þ mÞ bðm þ 1Þ D ð1 td Þcðm þ 1Þ½a þ n bðn þ 1Þg,

yjf ¼

Y jf ¼

1 ð1 td Þbðn þ 2Þ½eðh 2Þ þ b D þ ð1 td Þc½a þ n bðn þ 1Þ h1 2 , ð1 tf Þc ðm þ 1Þ k

1 ð1 td Þebðn þ 2Þ½eðm þ 1Þh m bg D ð1 td Þc½a þ n bðn þ 1Þ h1 2 ð1 tf Þc mðn þ 1Þ , k

Pjf ¼

ðE:5Þ

ðE:6Þ

ðE:7Þ

1 fð1 td Þkbðn þ 2Þ½eðh þ mÞ þ b D þ ð1 td Þck½a þ n bðn þ 1Þ ð1 tf Þc2 ðn þ 1Þðh þ mÞg,

ðE:8Þ

for jA{IT, IP}, where D ð1 td Þekbðn þ 2Þðm þ 2Þ ð1 tf Þc2 ðn þ 1Þðm þ 1Þ40:

APPENDIX F. PROOF OF PROPOSITION 5 IP We can easily show that E IT yd ;e oE yd ;e :

1 1 ð1 tf Þc2 ðn þ 1Þðm þ 1Þ þ ð1 tf Þcbðn þ 1Þðm þ 1Þ D DyIT d 1 ð1 tf Þcbðn þ 1Þðm þ 1ÞoE IP ¼ E IP yd ;e þ yd ;e . DyIT d

E IT yd ;e ¼

ðF:1Þ

362

ALEXANDRA LAI AND OANA SECRIERU

We can easily derive the elasticities of the foreign outputs with respect to the exchange rate under IT and IP respectively and show that they are: E IT yf ;e ¼

eDIT 40, DyIT f

(F.2)

E IP yf ;e ¼

eDIP 40, DyIP f

(F.3)

where DIT ¼ Dð1 td Þðn þ 2Þbðh þ mÞ DyIT f ð1 td Þðn þ 2Þkb40,

(F.4)

DIP ¼ Dð1 td Þðn þ 2Þb½ðh þ mÞ ðm þ 1Þb DyIP f ð1 td Þðn þ 2Þðm þ 2Þkbo0,

ðF:5Þ

IP for c close to zero. This implies that E IT yf ;e oE yf ;e for c close to zero. Using the results in Proposition 4, we can also show that DIP 4DIT ; IT which implies that E yf ;e oE IP yf ;e o1:

APPENDIX G. PROOF OF PROPOSITION 6 (1) The exchange rate pass-through into domestic and foreign prices are: E IT Pd ;e ¼

1 fð1 tf Þc2 ðm þ 1Þ½a þ n ðn þ 1ÞPIT d De þ ð1 tf Þcbbðm þ 1Þg,

ðG:1Þ

1 ð1 tf Þc2 ðm þ 1Þ½a þ n ðn þ 1ÞPIP (G.2) d . De IP Since, E IT Pd ;e o0 and E Pd ;e 40; we compute the difference between the IT absolute value of E Pd ;e and E IP Pd ;e in order to compare the sensitivity of the domestic prices to exchange rate movements under IT and IP: E IP Pd ;e ¼

1 IT ð1 tf Þcðm þ 1Þ½bb þ 2cða þ nÞ E Pd ;e E IP Pd ;e ¼ De IP cðn þ 1ÞðPIT d Pd Þ.

ðG:3Þ

Multinationals and Exchange Rate Pass-Through

363

For c close to zero, the left-hand side of Eq. (G.3) is positive and, IP AT therefore, E IT Pd ;e 4E Pd ;e 40 ¼ E Pd ;e : In order to show that the exchange rate pass-through into domestic prices is incomplete under both IT and IP, it is enough to show that IT E Pd ;e o1: For c close to zero: 1 IT fð1 tf Þc2 ðm þ 1Þ½a þ n ðn þ 1ÞPIT E Pd ;e 1 ¼ d De ð1 tf Þbbðm þ 1Þ ð1 td Þe2 kbðn þ 2Þðm þ 2Þ þ ð1 tf Þec2 ðn þ 1Þðm þ 1Þgo0.

ðG:4Þ

(2) The exchange rate pass-through into foreign prices under AT, IT, and IP respectively, is: E AT Pf ;e ¼

E IT Pf ;e ¼

E IP Pf ;e ¼

1 ð1 td Þbkbðn þ 2Þo0, De

(G.5)

1 fð1 tf Þc2 ðn þ 1Þ½h þ m ðm þ 1ÞPIT f De ð1 td Þkfc½a þ n bðn þ 1Þ þ bbðn þ 2Þggo0

ðG:6Þ

1 fð1 tf Þc2 ðn þ 1Þ½h þ m ðm þ 1ÞPIP f De ð1 td Þckf½a þ n bðn þ 1Þgg40

ðG:7Þ

for c close to zero. We can show that foreign prices are more sensitive to exchange ratemovements under AT than under IT by calculating the IT difference E AT E Pf ;e Pf ;e and show that it is positive: 1 AT IT cfð1 tf Þcðn þ 1Þ½h þ m ðm þ 1ÞPIT E Pf ;e E Pf ;e ¼ f De ð1 td Þk½a þ n bðn þ 1Þ 1 cðn þ 1Þðm þ 1Þg40, ðG:8Þ DDAT

364

ALEXANDRA LAI AND OANA SECRIERU

for c close to zero, where DAT ¼ ð1 td Þekbðn þ 2Þðm þ 2Þ: Also, 1 IT IP fð1 tf Þc2 ðn þ 1Þðm þ 1ÞðPIT E Pf ;e E IP Pf ;e ¼ f Pf Þ De 2ð1 tf Þc2 ðn þ 1Þðh þ mÞ þ 2ð1 td Þkc½a þ n bðn þ 1Þ þ ð1 td Þbkbðn þ 2Þg40,

ðG:9Þ

for c close to zero. IT AT So far we showed that E IP o E E o Pf ;e Pf ;e Pf ;e for c sufﬁciently close to zero. Since 1 AT ð1 td Þkbðn þ 2Þ½b eðm þ 2Þo0, E Pf ;e 1 ¼ De

(G.10)

for e4b=ðm þ 2Þ; it follows that the exchange rate pass-through into foreign prices is incomplete under AT, IT and IP.

CORPORATE GOVERNANCE IN RUSSIA: A CASE STUDY OF TIMELINESS OF FINANCIAL REPORTING IN THE TELECOM INDUSTRY Robert W. McGee ABSTRACT The present study focuses on the timeliness of financial reporting, which is an element of transparency. Specifically, it looks at the telecommunications industry in Russia and computes the number of days it takes companies to receive an audit opinion, then compares the time lag to the number of days it takes non-Russian companies in the telecommunications industry to receive an audit opinion. The study concludes that Russian companies take longer to report financial results than do non-Russian companies. Larger Russian companies take less time to report their financial condition than do small Russian firms, but the difference is not significant. The same was true for the non-Russian companies included in the sample. Companies using Russian Accounting Standards took significantly less time to report financial results than did companies using either International Financial Reporting Standards (IFRS) or US generally accepted accounting principles (GAAP). Companies using IFRS Value Creation in Multinational Enterprise International Finance Review, Volume 7, 365–390 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07015-4

365

366

ROBERT W. MCGEE

took significantly longer to report financial results than did companies using US GAAP. The dominant auditor in the Russian telecommunications industry did not complete audits in significantly less time than did nondominant auditors. Although Russian companies take far less time to issue financial statements now than they did a few years ago, it is premature to definitively conclude that the improvement is significant due to the limited data set.

INTRODUCTION It is generally perceived that the quality of corporate governance in the transition economies of Central and Eastern Europe is not on a par with that of the developed market economies. There is a lot of empirical and anecdotal evidence to support that belief (McGee & Preobragenskaya 2004, 2005, 2006; Preobragenskaya & McGee, 2004). A report by the European Bank for Reconstruction and Development (EBRD, 2005) looked at the extensiveness and effectiveness of corporate governance laws in transition economies. The report measured extensiveness using the well-known Organisation for Economic Cooperation and Development (OECD) benchmarks and measured effectiveness using the EBRD’s Legal Indicator Survey. The EBRD Report ranked 27 transition countries into ﬁve categories based on the extent of their compliance with international corporate governance standards. Table 1 shows the breakdown. As can be seen, none of the 27 transition economies in the EBRD study had the highest possible score. Nearly 30 percent had either low or very low levels of compliance. More than two thirds fell into the medium- to highcompliance category. Thus, there is much room for improvement. Chart 1 shows this relationship graphically. The World Bank has conducted a series of studies that rate various aspects of corporate governance in a number of transition and developing economies. One aspect these studies examined was the timeliness and fairness of ﬁnancial reporting. Table 2 shows the relative timeliness and fairness of ﬁnancial reporting for those transition economies where World Bank studies have been made. Unfortunately, the World Bank studies did not include some of the largest transition countries, such as Russia and Ukraine. Thus, the present study partially ﬁlls this void in the literature regarding the timeliness of ﬁnancial

Corporate Governance in Russia

Table 1.

Extent of Compliance with International Corporate Governance Standards.

Country

Very High

Albania Armenia Azerbaijan Belarus Bosnia and Herzegovina Bulgaria Croatia Czech Republic Estonia Georgia Hungary Kazakhstan Kyrgyz Republic Latvia Lithuania Macedonia Moldova Poland Romania Russia Serbia and Montenegro Slovak Republic Slovenia Tajikistan Turkmenistan Ukraine Uzbekistan Total Percentage of total

367

High

Medium

Low

Very Low

X X X X X X X X X X X X X X X X X X X X X X X X X X X 0 0.0

9 33.3

10 37.0

4 14.8

4 14.8

Source: EBRD (2005), McGee (2006a).

reporting in transition economies. Chart 2 shows the relative percentages of the countries that fall into each category. The present study focuses on the timeliness of ﬁnancial reporting, which is an element of transparency. Speciﬁcally, it looks at the telecommunications industry in Russia and computes the number of days it takes companies to receive an audit opinion, then compares the time lag to the number of days it takes non-Russian companies in the telecommunications industry to receive an audit opinion. The study concludes that Russian companies take

ROBERT W. MCGEE

368

Very low 15%

Very high 0%

High 33% Low 15%

Medium 37%

Chart 1.

Compliance with Corporate Governance Standards.

longer to report ﬁnancial results than do non-Russian companies. Larger Russian companies take less time to report their ﬁnancial condition than do small Russian ﬁrms, but the difference is not signiﬁcant. The same was true for the non-Russian companies included in the sample. Companies using Russian Accounting Standards (RAS) took signiﬁcantly less time to report ﬁnancial results than did companies using either International Financial Reporting Standards (IFRS) or US generally accepted accounting principles (GAAP). Companies using IFRS took signiﬁcantly longer to report ﬁnancial results than did companies using US GAAP. The dominant auditor in the Russian telecommunications industry did not complete audits in signiﬁcantly less time than did nondominant auditors. Although Russian companies take far less time to issue ﬁnancial statements now than they did a few years ago, it is premature to deﬁnitively conclude that the improvement is signiﬁcant due to the limited data set.

TRANSPARENCY AND TIMELINESS Transparency is one of those terms that has many facets. It is used in different ways. It can refer to the openness of governmental functions. It can

Corporate Governance in Russia

369

Table 2. Fair and Timely Dissemination of Information in Selected Transition Economies. Country

Observed

Armenia (World Bank, 2005a) Azerbaijan (World Bank, 2005b) Bulgaria (World Bank, 2002a) Croatia (World Bank, 2001) Czech Republic (World Bank, 2002b) Georgia (World Bank, 2002c) Hungary (World Bank, 2003a) Latvia (World Bank, 2002d) Lithuania (World Bank, 2002e) Macedonia (World Bank, 2005c) Moldova (World Bank, 2004a) Poland (World Bank, 2005d) Romania (World Bank, 2004b) Slovak Republic (World Bank, 2003b) Slovenia (World Bank, 2004c) Frequency

Largely Partially Materially Not Observed Observed Not Observed Observed X X X X X X

X X X X X X X X X 2

3

5

4

Materially not observed

Not observed

1

Source: World Bank, McGee (2006a).

35 30 25 %

20 15 10 5 0 Observed

Chart 2.

Largely observed

Partially observed

Fair and Timely Dissemination of Information.

refer to a country’s economy. Or it can refer to various aspects of corporate governance and ﬁnancial reporting. The OECD (1998) lists transparency as one element of good corporate governance. Kulzick (2004) and others

370

ROBERT W. MCGEE

(Blanchet, 2002; Prickett, 2002) view transparency from a user perspective. According to their view, transparency includes the following eight concepts:

Accuracy Consistency Appropriateness Completeness Clarity Timeliness Convenience Governance and enforcement

This paper focuses on just one aspect of transparency – timeliness. The International Accounting Standards Board considers timeliness to be an essential aspect of ﬁnancial reporting. In Accounting Principles Board (APB) Statement No. 4, the APB (1970) in the USA listed timeliness as one of the qualitative objectives of ﬁnancial reporting disclosure. APB Statement No. 4 was later superseded, but the Financial Accounting Standards Board (1980) continued to recognize the importance of timeliness in its Concepts Statement No. 2. The US Securities and Exchange Commission also recognizes the importance of timeliness and requires that listed companies ﬁle their annual 10-K reports by a certain deadline. The issue of timeliness has several facets. There is an inverse relationship between the quality of ﬁnancial information and the timeliness with which it is reported (Kenley & Staubus, 1974). Accounting information becomes less relevant with the passage of time (Atiase, Bamber, & Tse, 1989; Hendriksen & van Breeda, 1992; Lawrence & Glover, 1998). Studies show mixed conclusions regarding the relationship between quickness of reporting and the nature of the information being reported. Some studies show that good news is reported before bad news, whereas other studies show that bad news is reported before good news. There is some evidence to suggest that it takes more time to report bad news than good news (Bates, 1968; Beaver, 1968), both because companies hesitate to report bad news and because companies take more time to massage the numbers or resort to creative accounting techniques when they have to report bad news (Givoli & Palmon, 1982; Chai & Tung, 2002; Trueman, 1990). Stated differently, there seems to be a tendency to rush good news, such as better than expected earnings, to press and delay the reporting of bad news, such as less than expected earnings (Chambers & Penman, 1984; Kross & Schroeder, 1984). Dwyer and Wilson (1989) found this relationship to hold true for municipalities. Haw, Qi, and Wu (2000) found it to be the

Corporate Governance in Russia

371

case with Chinese companies. Leventis and Weetman (2004) found it to be the case for Greek ﬁrms. However, Annaert, DeCeuster, Polﬂiet, and Van Campenhout (2002) found that this was not the case for Belgian companies, and Han and Wang (1998) found that this was not the case for petroleum-reﬁning companies, which delayed reporting extraordinarily high proﬁts during the Gulf crisis of the 1990s, perhaps because political repercussions outweighed what would otherwise have been a good market reaction. Rees and Giner (2001) found that companies in France, Germany, and the UK tended to report bad news sooner than good news. A study by Basu (1997) found that companies tend to report bad news quicker than good news, presumably because of conservatism. Gigler and Hemmer (2001) discuss this point in their study, which ﬁnds that ﬁrms with more conservative accounting systems are less likely to make timely voluntary disclosures than are ﬁrms with less conservative accounting systems. Building on the Basu study (1997), Pope and Walker (1999) found that there were cross-jurisdictional effects when extraordinary items were either included or excluded, using US and UK ﬁrms for comparison. Han and Wild (1997) examined the potential relationship between the timeliness of earnings reporting and the share price reactions of competing ﬁrms. But Jindrichovska and Mcleay (2005) found that there was no evidence of conservatism in the Czech accounting system when it came to reporting bad news earlier than good news, presumably because the Czech tax system offers little incentive to do so. Ball, Kothari, and Robin (2000) found that companies in jurisdictions that have a strong shareholder orientation tend to disclose earnings information sooner than companies in countries operating under a legal code system. There is also a relationship between the speed with which ﬁnancial results are announced and the effect the announcement has on stock prices. If information is released sooner, the effect on stock prices is more pronounced. The longer the time lapse between year-end and the release of ﬁnancial information, the less effect there is on stock price, all other things being equal (Ball & Brown, 1968; Brown & Kennelly, 1972). This phenomenon can be explained by the fact that ﬁnancial information seems to seep into the stock price over time, so the more time that elapses between yearend and the release of the ﬁnancial reports, the more such information is already included in the stock price. Some countries report ﬁnancial results faster than other countries. DeCeuster and Trappers (1993) found that Belgian companies take longer to report their ﬁnancial results than do Anglo-Saxon countries. Annaert et al.

372

ROBERT W. MCGEE

(2002) found this to be the case for interim information as well. Companies can report ﬁnancial results faster on the Internet, and the information can be more widely disbursed, but posting two-year-old annual reports does nothing to improve timeliness (Ashbaugh, Johnstone, & Warﬁeld, 1999). Atiase et al. (1989) found that large companies report earnings faster than small companies and that the reporting of earnings has a more signiﬁcant market reaction for small ﬁrms than for large ﬁrms. In a study of Australian ﬁrms, Davies and Whittred (1980) found that small ﬁrms and large ﬁrms made signiﬁcantly more timely reports than medium-size ﬁrms and that proﬁtability was not a signiﬁcant variable. Whittred (1980) found that the release of ﬁnancial information for Australian companies is delayed the ﬁrst time an audit ﬁrm issues a qualiﬁed report and that the extent of the delay is longer in cases where the qualiﬁcation is more serious. Keller (1986) replicated that study for US companies and found the same thing to be true. Whittred and Zimmer (1984) found that it took Australian ﬁrms in ﬁnancial distress a signiﬁcantly longer time to publish their ﬁnancial information. A study of more than 5,000 annual reports of French companies found that it took longer to release audit reports where there had been a qualiﬁed opinion, and that the more serious the qualiﬁcation, the greater the delay in releasing the report (Soltani, 2002). Krishnan (2005) found that the audit ﬁrm’s degree of expertise has an effect on the timeliness of the publication of bad earnings news. Audit ﬁrms that specialize in the industry in which the company operates are timelier in reporting bad ﬁnancial news than are audit ﬁrms that have less industry expertise. One measure of the transparency and quality of ﬁnancial reporting is timeliness. The lapse of time between a company’s year-end and the date when ﬁnancial information is released to the public is related to the quality of the information reported. Issuing excellent, accurate, and comprehensive ﬁnancial information two or three years after year-end is not as desirable as issuing less comprehensive and complete ﬁnancial information a few months after year-end. Financial information becomes stale after a few months, and certainly after two or three years. The staler it is, the less relevant it is to potential investors and creditors. There are a number of reasons for the time lag between year-end and the issuance of the audit report and the publication of ﬁnancial information. Ashton, Graul, and Newton (1989) identiﬁed auditor size, industry classiﬁcation, the presence or absence of extraordinary items, and the sign of net income as some of the factors that inﬂuence timeliness. To that, one might

Corporate Governance in Russia

373

add the culture and the political and economic system of the country in which the particular ﬁrm is located. One purpose of the present study is to determine whether Russian companies are any less timely in the speed of ﬁnancial reporting than companies in Western Europe, the developed part of Asia, and the USA. In the not too distant past, some Russian enterprises were criticized for waiting too long to issue their ﬁnancial reports. Some Russian companies did not issue their annual reports until a year or more after the end of the ﬁnancial year. In some cases, Russian ﬁrms did not even have annual audits (McGee & Preobragenskaya, 2005). Since the quality of corporate governance in transition economies is signiﬁcantly lower than in developed market economies, and since timeliness of ﬁnancial reporting is an element of corporate governance, one might conclude that the timeliness of ﬁnancial reporting in transition economies is not on a par with that of the developed market economies. But such a conclusion might be hasty. Although companies in transition economies are not as accustomed to issuing timely ﬁnancial reports, or any kind of ﬁnancial reports for that matter, one might assume that it would take them longer to issue ﬁnancial statements, since they are at the beginning of the learning curve. Furthermore, disclosing information of any kind is not something that former communists feel comfortable doing, so there is a built-in cultural and political resistance to releasing ﬁnancial information. But there is another side to this coin. Practically all the largest companies in transition economies have their books audited by one of the Big-4 accounting ﬁrms or one of the other large international accounting ﬁrms. Most of them also issue ﬁnancial statements based on either IFRS or US GAAP. This being the case, perhaps it is premature to conclude that the quality of ﬁnancial reporting for large companies in transition economies is lower in qualitative terms than the ﬁnancial reporting of companies in the more developed market economies. The main purpose of the present study is to test the assumption that one aspect of quality of ﬁnancial reporting – timeliness – is lower in transition economies. Measuring timeliness is relatively easy. The present study measures timeliness by computing the number of days that elapse between the company’s year-end and the date of the auditor’s report. Data for some large Russian companies in the telecommunications industry are calculated and compared with those of selected non-Russian companies in the same industry. McGee (2005) conducted a single-year study of the Russian energy sector using a similar methodology. Another study discussed the opacity of Russian ﬁnancial reporting in general and examined the extent of disclosure and

374

ROBERT W. MCGEE

timeliness of ﬁnancial reporting for Russia’s 10 most transparent companies (McGee, 2006c).

COMPARING RUSSIAN AND NON-RUSSIAN COMPANY DATA IN THE TELECOM INDUSTRY Fortune does an annual ranking of the top 50 Russian companies in terms of sales (Demos, 2006). Table 3 lists the members of the top 50 Russian companies that are in the telecommunications industry and provides some basic ﬁnancial information. Standard & Poor’s does periodic rankings of Russian companies in terms of ﬁnancial transparency (Kochetygova et al., 2005). The most recent study ranked 54 Russian companies on a scale from 1 to 100, where 100 is the best score. The standings and scores for the Russian telecommunication companies that were included in the Standard & Poor’s study are also listed in Table 3. The conversion rate for monetary units was $1 ¼ 27.75 rubles.

Table 3. Size/ Rank (2005) 14 18 23 28 29 33 34 39

42 43 50

Russian Telecommunication Companies.

Company

MTS (www.company.mts.ru) Vimpelcom (www.vimpelcom.ru) Rostelcom (www.rt.ru) Uralsvyazinform (www.uralsviazinform.com) Centertelecom (www.centertelecom.ru) Sibirtelecom (www.sibertelecom.ru) Volgatelecom (www.vt.ru) Southern Telecom (UTK – Uralskaya Telefonia Kompania) (www.utk.ru, www.stcompany.ru) Golden Telecom (www.goldentelecom.com) North-West Telecom (www.nwtelecom.ru) MGTS (Moscow City Telephone) (www.mgts.ru)

Revenue ($ Proﬁts ($ million, million, 2004) 2004)

Market value ($ million)

Transparency Transparency Rank (54) Score (100 ¼ best)

3,887.0 2,146.6

987.9 350.4

13,770.0 8,230.2

1 5

84 75

1,344.8 969.4

154.9 81.2

1,745.3 1,367.0

2 12

82 64

938.6

(33.1)

787.1

13

63

814.5

73.1

815.5

14

63

755.6 624.2

89.3 (24.0)

970.8 224.6

11 10

65 67

584.0

64.8

na

5

75

555.2

25.4

556.2

7

71

480.8

75.0

958.2

30

48

Corporate Governance in Russia

375

Table 4 shows the number of days it took Russian telecommunication companies to issue their audit report. Data were collected from company Web sites. The dates used were the dates of the auditor’s report, which are not necessarily the same as the date the ﬁnancial information was released to the public. However, it was not possible to determine when the annual ﬁnancial reports were published, so the date of the audit opinion was selected as a surrogate. Using this date also made it possible to compare Russian and non-Russian company data. Information was not given in some cases. The range was 65 to 339 days, with an average of 138.3 days. In other words, assuming the company’s year-end was December 31, some companies published their audit opinion as early as March 6, while others took until December 5. The average company from this group published its audit report on May 19. Non-Russian companies were taken from the Fortune Global 500 list for Anon (2005), which was published in the July 25, 2005, issue of Fortune. The Fortune article classiﬁed the selected companies as being in the telecommunications industry. Data for the most recently reported year were selected, as well as data from some earlier years. Some companies reported more years than others. Many companies had already issued their 2005 annual reports by the date this paper was written, but a few had not. Table 5 shows the sales and ranking of the Fortune Global 500 companies that are in the telecommunications industry. The smallest company on the list is more than three times larger than the largest company on the Russian list. Table 6 shows the number of days’ delay between year-end and the issuance of the audit opinion for the Fortune Global 500 companies that are in the telecommunications industry. The range was 23–281 days, with an average of 63.2 days. In other words, assuming the company’s year-end was December 31, some companies published their audit opinion as early as January 23, while others took until September 8. The average company from this group published its audit report on March 5, or about 75 days before the average Russian company. A similar study of timeliness in the Russian energy sector found that Russian companies delayed the release of audited ﬁnancial statements for 148.7 days after year-end, compared with 78.6 days for non-Russian companies in the petroleum-reﬁning industry (McGee, 2006b, 2006c). Data were gathered from the companies’ Web sites, sometimes supplemented by accessing the US Securities and Exchange Commission Web site. Some companies reported more years than others. Some companies did not

376

Table 4.

ROBERT W. MCGEE

Timeliness Data for Russian Telecommunication Companies.

Company

Year

Days’ Delay

MTS

2004

81

2003

86

2002

140

2001

339

Vimpelcom

Rostelecom

2004 2003 2002 2001 2000 1999 1998 1997 1996 2004

90 75 73 73 79 95 85 Unknown Unknown 174

Auditor Deloitte & Touche Deloitte & Touche Deloitte & Touche Deloitte & Touche Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Unknown

2003

181

Ernst & Young

2002

157

Ernst & Young

2001

239

Ernst & Young

2000

138

PWC

Uralsvyazinform

2004 2003 2002 2001 2000 1999 1998 1997

168 170 230 228 176 172 158 181

Centertelecom

2004 2003 2002 2001 2004 2003 2002 2001 2000 2004 2003 2002 2001 2000

Ernst & Young Unknown Ernst & Young Unknown PWC PWC PWC Coopers & Lybrand Ernst & Young Ernst & Young Ernst & Young Arthur Andersen Unknown Ernst & Young Ernst & Young Ernst & Young Arthur Andersen Ernst & Young Ernst & Young Ernst & Young Unknown Astrea

Sibirtelecom

Volgatelecom

Unknown Unknown Unknown 88 91 110 220 Unknown Unknown 90 114 106 119 Unknown

Standards Used US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP IFRS with US GAAP reconciliation IFRS with US GAAP reconciliation IFRS with US GAAP reconciliation IFRS with US GAAP reconciliation IFRS with US GAAP reconciliation IFRS IFRS IFRS IFRS IFRS IFRS IFRS IFRS IFRS Unk IFRS IFRS RAS RAS IFRS IFRS Unknown RAS RAS RAS RAS RAS

Corporate Governance in Russia

377

Table 4. (Continued ) Company

Year

Southern Telecommunication Company (UTK) 2004 2003 2002 2001 2000 1999 Golden Telecom 2004 2003 2002 North-West Telecom 2004 2003 2002 2001 2000 1999 1998 1997 MGTS (Moscow City Telephone) 2004 2003 2002 2001 2000 1999 Average

Days’ Delay

Auditor

177 Unknown Unknown 228 110 98 70 70 65 166 213 212 258 105 106 99 86 89

Ernst & Young Unknown Unknown Unknown Unknown Unknown Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Ernst & Young Unknown A. Y. Grebenyuk A. Y. Grebenyuk V. A. Balashov V. A. Balashov Deloitte & Touche Deloitte & Touche Unknown Unknown PWC PWC

142 Unknown Unknown 118 117

Standards Used IFRS IFRS IFRS US GAAP US GAAP US GAAP US GAAP US GAAP US GAAP IFRS IFRS IFRS US GAAP RAS RAS RAS RAS RAS RAS Unknown Unknown RAS RAS

138.3

post their annual reports on their Web site. In a few cases, the annual report or other ﬁnancial information was posted, but the date of the auditor’s opinion was not given. As of the date of this writing, some companies had not yet posted their data for 2005. The ﬁrm that signed the audit opinion is also listed in Table 6. This information will be used to determine whether there is a correlation between the audit ﬁrm that signs the opinion and the amount of time it takes to issue the opinion.

TESTING HYPOTHESES Prior studies have reached certain ﬁndings for various aspects of timeliness. The present study examines data from Russian and non-Russian companies in the telecommunications industry to test whether the ﬁndings of some other studies can be veriﬁed using the Russian data. For example, DeCeuster and Trappers (1993) found that Belgian companies take longer

Fortune Global 500 Companies (2005): Telecommunications Industry (million USD, 2004).

Rank within Industry Global 500 Rank (2005)

Source: Fortune (2005).

18 37 38 53 63 102 111 114 140 162 192 194 199 224 244 247 262 290 401 414 416 418 451 463

Revenues

Proﬁts

Nippon Telegraph & Telephone (www.ntt.co.jp) 100,545 6,608 Deutsche Telekom (www.telekom.de) 71,989 5,764 Verizon Communications (merged with MCI) (www.verizon.com) 71,563 7,831 Vodafone (www.vodafone.com) 62,971 (13,910) France Telecom (www.francetelecom.com) 58,652 3,463 SBC Communications (merged with AT&T) (www.sbc.com) 41,098 5,887 Telecom Italia (www.telecomitalia.com) 39,228 971 Telefonica (www.telefonica.com) 38,188 3,579 BT (www.bt.com) 34,673 3,360 AT&T (www.att.com) 30,537 (6.469) Sprint (www.sprint.com) 27,428 (1,012) KDDI (www.kddi.com/english/) 27,170 1,866 Vivendi Universal (http://ﬁnance.vivendiuniversal.com) 26,651 938 China Mobile Communications (www.chinamobile.com) 23,958 4,078 Bell South (merged with AT&T) (www.bellsouth.com) 22,729 4,758 MCI (merged with Verizon) (www.mci.com) 22,615 (4,002) China Telecommunications (www.chinatelecom.com.cn) 21,562 2,422 Comcast (www.comcast.com) 20,307 970 Telstra (http://telstra.com) 15,193 2,939 KT (www.kt.com.kr) 14,901 1,119 BCE (www.bce.ca) 14,842 1,224 Royal KPN (www.kpn.com) 14,828 1,879 Qwest Communications (www.qwest.com) 13,809 (1,794) Nextel Communications (merged with Sprint) (www.nextel.com) 13,368 3,000

ROBERT W. MCGEE

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Company

378

Table 5.

Fortune Global 500 Companies (2005): Telecommunications Industry, Days’ Delay in Issuing Audit Opinion.

Company Nippon Telegraph & Telephone Deutsche Telekom

Verizon Communications GTE (merged with Verizon) Bell Atlantic (merged with Verizon) Vodafone France Telecom SBC Communications Telecom Italia Telefonica BT AT&T Sprint (merged with Nextel) KDDI Vivendi Universal

2005

2004

2003

2002

2001

2000

1999

1998

90 PWC 44 PWC, EY

90 PWC 53 PWC, EY 53 EY

88 PWC 71 PWC, EY 29 EY

114 PWC 76 PWC, EY 29 EY

165 PWC 77 PWC, EY 31 EY

90 PWC 85 PWC

90 – 87 PWC

87 PWC

27 AA 45 PWC 69 DT

28 AA 40 PWC 63 DT

54 EY

54 DT 41 EY, DT 56 EY 81 EY 60 EY 48 PWC 47 EY 66 KPMG 85 PWC

57 DT 43 EY, DT 40 EY 106 EY 57 DT 51 PWC 65 PWC 31 EY 85 Other 99 EY

57 DT 63 EY 38 EY 108 EY 57 DT 51 PWC 23 PWC 36 EY 86 Other 92 EY

59 DT 80 EY 39 EY 105 EY 73 AA 52 PWC 39 EY 35 EY 87 Other 87 AA

1996

1995

27 AA 40 CL 85 DT

28 AA 36 CL 88 DT

84 DT

50 CL 26 CL 34 EY

35 EY

45 EY

32 EY

59 DT 81 EY

141 PWC 59 AA 54 PWC 40 EY 32 EY 89 Other 92 AA

55 PWC 69 PWC 32 EY 90 Other 70 AA

25 PWC 33 EY 87 Other

379

55 DT 41 EY, DT 56 EY 76 EY 63 DT 49 PWC 67 PWC 69 KPMG 85 PWC 69 EY

1997

Corporate Governance in Russia

Table 6.

Company China Mobile Communications Bell South MCI (successor to Worldcom) China Telecommunications Comcast Telstra

KT BCE Royal KPN Qwest Communications Nextel Communications (merged with Sprint)

2005

2004

2003

2002

2001

2000

58 PWC

63 PWC 74 KPMG 90 KPMG 52 DT 43 EY

40 PWC 118 KPMG 77 KPMG 71 DT 59 EY

41 PWC 76 DT 83 KPMG

39 PWC 66 AA

57 PWC 89 AA

59 Aud. Gen.

60 Aud. Gen.

61 DT 59 PWC 49 KPMG 73

70 DT 68 PWC 62 KPMG 70

57 DT 59 PWC 281 KPMG 51

DT

DT

DT

81 KPMG 52 DT 42 EY

31 DT 59 PWC 47 KPMG

380

Table 6. (Continued ) 1999

1998

1997

1996

1995

61 EY

57 Aud. Gen.

57 PWC

60 Aud. Gen.

75 Aud. Gen.

73 Aud. Gen.

58 DT 74 PWC 29 AA 78

59 DT 84 PWC 24 AA 46

54 DT 80 PWC 31 AA 56

33 KPMG 53

53

72

79

DT

DT

DT

DT

DT

DT

DT

ROBERT W. MCGEE

Note: Code: AA ¼ Arthur Andersen; Aud. Gen. ¼ Auditor General; CL ¼ Coopers & Lybrand; DT ¼ Deloitte & Touche; EY ¼ Ernst & Young; PWC ¼ PricewaterhouseCoopers.

Corporate Governance in Russia

381

to report their ﬁnancial results than do Anglo-Saxon countries. Annaert et al. (2002) found this to be the case for interim information as well. The present study tests to see whether Russian companies in the telecommunications industry take longer to report annual results than do non-Russian companies in the same industry. The hypothesis to test this relationship might be stated as follows: H1. Russian companies in the telecommunications industry take signiﬁcantly longer to report annual audited ﬁnancial information than do nonRussian companies in the same industry. Firms that audited Russian companies took an average of 138.3 days to issue their audit opinion, compared with just 63.2 days for the audit of nonRussian companies on the Fortune Global 500 list. The Wilcoxon test was done to see whether this difference was signiﬁcant. An advantage of using a nonparametric test like the Wilcoxon test is that one need not make the assumption that the distribution is random. The Wilcoxon test found the difference to be highly signiﬁcant (pr1.05 1019). Chart 3 shows the difference between the Russian and non-Russian companies as a percentage of a year. Atiase et al. (1989) found that large companies report earnings faster than small companies. It was decided to test this ﬁnding using Russian data. This hypothesis might be stated as follows: H2. Large Russian companies take less time to publish audited ﬁnancial information than do small Russian companies. What constitutes a large company is subjective. Furthermore, the largest Russian company in the present study is much smaller than the smallest non-Russian company included in this study, which could lead one to

Russian Companies

Non-Russian Companies 63.2

138.3

365

Chart 3.

365

Comparison of Russian & Non-Russian Companies.

382

ROBERT W. MCGEE

conclude that all the Russian companies are small companies. Any attempt to divide a range of companies into large and small must be at least somewhat arbitrary. But a decision had to be made in order to conduct the necessary test. In the present case, a large Russian company was deﬁned as a company that had annual sales of more than $1 billion. Three companies were large according to this deﬁnition. The other eight companies were classiﬁed as small for the purposes of this test. The average delay for the large Russian companies was 131.6 days, compared with a delay of 141.4 days for the smaller Russian companies. The days’ delay for the two groups was then compared using the Wilcoxon test to determine whether the difference was signiﬁcant. The Wilcoxon test found that the difference was not highly signiﬁcant (pr0.1939). Since data were also available for the Fortune Global 500 telecommunication companies, it was decided to perform the Wilcoxon test for this group of companies. However, before this test could be done it was ﬁrst necessary to determine which companies were large and which were small. Again, a certain amount of arbitrariness was involved, since there is no clear-cut distinction between which companies are large, which are medium sized, and which are small. If large is deﬁned as the ﬁve largest telecommunication companies in the Fortune Global 500 and small is deﬁned as the smallest ﬁve companies on that list, then it took an average of 62.7 days for the large company auditors to issue their audit reports, compared with 64.4 days for small company auditors. Companies falling into the middle range took 63.1 days. So it appears that the larger a company is, the less time it takes to get an audit opinion, which conﬁrms the results of the Atiase et al. (1989) study. However, a Wilcoxon test comparing the days’ delay for the large and small companies found that the difference was not statistically signiﬁcant (pr0.7032). One might expect that it would take more time to audit ﬁnancial statements prepared using US GAAP, since US GAAP is more complicated than IFRS and covers more topics. One might also expect that it would take Russian auditors less time to audit statements prepared using RAS than either US GAAP or IFRS since RAS are not as complicated as either of the two internationally recognized sets of accounting standards and Russian auditors are probably more familiar with RAS than with either of the international standards. One might reach such a conclusion a priori because it seems logical. However, one can also test this hypothesis. Also, in cases where a company issues both RAS-based statements and statements that are based on one of the sets of international standards (US GAAP or IFRS), one can compare the issuance dates.

Corporate Governance in Russia

383

In the course of the present study, the author noticed that Uralsvyazinform’s annual report for 2000 had two audit opinions, one for the IFRS prepared statements and one for the RAS statements. The audit opinion for the RAS statements was dated 110 days after year-end. The IFRS audit opinion was dated 176 days after year-end. Thus, in this one case at least, there was an additional delay of 66 days. In other words, it took the company auditors an additional 60 percent in terms of days to issue their IFRS opinion. The present study sought to see whether this relationship can be generalizable by comparing the days’ delay for statements prepared using RAS with those prepared using one of the two international standards (US GAAP or IFRS). The delay using US GAAP was also compared with the delay for companies that used IFRS. The hypotheses might be stated as follows: H3. Independent accounting ﬁrms that audit the books of companies that issue ﬁnancial statements based on US GAAP or IFRS take longer to issue their audit opinions than do accounting ﬁrms that audit companies that use RAS. H4. Independent accounting ﬁrms that audit the books of companies that use IFRS take less time to issue their audit opinions than do accounting ﬁrms that audit companies that use US GAAP. The ﬁndings support the view that it takes less time to audit RAS statements than statements prepared using one of the two sets of international accounting standards. However, an examination of the numbers also found that it takes less time to audit ﬁnancial statements prepared according to US GAAP than statements prepared using IFRS, which is just the opposite of what was expected. The breakdown by days is as follows: RAS US GAAP IFRS

106.6 days 117.5 days 181.5 days

The Wilcoxon test found that the difference between the RAS and IFRS statements was highly signiﬁcant (pr1.587 1005). The difference between the US GAAP and IFRS statements was also highly signiﬁcant (pr0.0008302). But the difference between the RAS and US GAAP

384

ROBERT W. MCGEE

statements was not signiﬁcant (pr0.5688). It would be interesting to see if this highly signiﬁcant difference between US GAAP and IFRS statements is also present in nontransition economies. It would also be interesting to see whether these relationships are similar in different industries, or if telecommunications is a special case. Krishnan (2005) found that the audit ﬁrm’s degree of expertise has an effect on the timeliness of the publication of bad earnings news. Audit ﬁrms that specialize in the industry in which the company operates are timelier in reporting bad ﬁnancial news than are audit ﬁrms that have less industry expertise. The present study wanted to test a variation of this ﬁnding. But rather than testing the relationship between publishing bad news and the timeliness of publishing ﬁnancial information, it was determined that the time it takes for the dominant auditor in the telecommunications industry to issue its audit report would be compared with the time it takes the other audit ﬁrms to issue their audit reports. The hypothesis to test this relationship might be stated as follows: H5. Financial statements that are audited by the dominant auditor in the industry are published faster than are statements audited by a nondominant auditor. The ﬁrst step was to determine which audit ﬁrm dominates the telecommunications industry. Actually, two comparisons were made, one for the telecommunications industry in Russia and one for the telecommunication companies that were listed in the Fortune Global 500. The results of the comparisons are shown in Tables 7 and 8. As can be seen, the dominant audit ﬁrm in the Russian telecommunications industry is Ernst & Young, and there is no dominant auditor in the case of the large international telecommunication companies. It took Ernst & Young an average of 137.3 days to issue an audit opinion, compared with 134.3 days for other audit ﬁrms, which is just the opposite of what was expected. Thus, the present study does not conﬁrm the Krishnan (2005) study. The Wilcoxon test found the three-day difference to be not signiﬁcant (pr0.6624). As Russian companies move up the learning curve, they become more efﬁcient. That being the case, it is reasonable to assume that it does not take Russian companies as long to issue their ﬁnancial statements today as it did a few years ago.

Corporate Governance in Russia

Table 7.

385

Dominant Auditor in the Russian Telecommunications Industry.

Audit Firm

Most Recent Year

Deloitte & Touche Ernst & Young KPMG PricewaterhouseCoopers Others

Total Years

No.

%

No.

%

2 9 0 0 0

18.2 81.8

6 29 0 6 7

12.5 60.4 12.5 14.6

Table 8. Dominant Auditor in the Fortune Global 500 Telecommunications Industry. Audit Firm

Deloitte & Touche Ernst & Young KPMG PricewaterhouseCoopers (and Coopers Lybrand) Others

Most Recent Year

Total Years

No.

%

No.

%

4 8 3 6 0

19.0 38.1 14.3 28.6

38 47 13 47 21

22.9 28.3 7.8 28.3 12.7

The best way to test this hypothesis would be to compute the number of days’ delay for the most recent period and compare those ﬁgures with the ﬁgures for some distant period, such as 10 or 15 years ago. The problem with this approach is that most Russian companies do not report information that is more than a few years old on their Web site, which makes it impossible to do a quick comparison between, say, 2004 data and 1990 data (which are probably not available anyway, since many ﬁrms were not yet privatized by 1990). So a compromise was reached. It was thought that it would be worthwhile to compare data for the most recent year with those of the oldest year for which data were available. The data might not be very useful in the cases where the Russian company reported data of only two or three years, but the comparison would not be totally useless either. Any conclusions drawn would have to be tentative. A larger sample size and a longer and more uniform span of years would be required to make the comparison strong, but even a comparison with limited data might reveal something, even if only tentative.

386

ROBERT W. MCGEE

The hypothesis for this test might be as follows: H6. Russian companies issue their ﬁnancial statements faster now than they did a few years ago. The data show that it took audit ﬁrms an average of 119.6 days to publish their audit opinion for the most current year and 144.8 days for the oldest date reported on their Web site, a difference of 25.2 days, or 17.4 percent faster if we use 144.8 for the denominator. Thus, it would appear that Russian companies are becoming faster in issuing their ﬁnancial statements, as learning curve theory would suggest. However, the Wilcoxon test found that this difference was not signiﬁcant (pr0.5708). There may be several explanations for this insigniﬁcant change. For one, the sample size was small, just 10, since 1 of the 11 Russian companies in the telecommunications industry had data for just one year and was thus excluded from the calculation. Furthermore, for the companies that had data for more than one year, the average was only 4.8 years, which is not a very long range of time. A better comparison would have been to compare 50 or 100 or more companies over a 10- or 15-year period. However, the data needed to make this comparison simply were not available.

SUMMARY AND CONCLUDING COMMENTS If one were to summarize the results of the present study, the summary might go something like this: Russian companies in the telecommunications industry take signiﬁcantly longer to report audited ﬁnancial information than do non-Russian companies in the same industry – 138.3 days on average, compared with 63.2 days. Although larger Russian companies took less time to report audited ﬁnancial information than did small companies, the time difference was not highly signiﬁcant – 131.6 days compared with 141.4 days. The larger non-Russian companies in the sample also took less time to report than did small companies – 62.7 days compared with 64.4 days – but this difference was not signiﬁcant. It took signiﬁcantly less time for companies to issue audited ﬁnancial information if they prepared their ﬁnancial statements using RAS rather than one of the two sets of internationally recognized standards of ﬁnancial reporting. Financial statements using IFRS took signiﬁcantly longer to audit than US GAAP statements.

Corporate Governance in Russia

387

The dominant auditor in an industry may be presumed to have a higher level of expertise but does not ﬁnish audit work any sooner than audit ﬁrms that do not hold a dominant position in the industry. Although Russian companies take less time to issue audited ﬁnancial statements now than they did a few years ago, it is premature, due to lack of data, to determine whether the increased speediness is signiﬁcant. Corporate governance in general in Russia is not yet at the level of governance in companies located in the developed market economies. Timeliness, one aspect of corporate governance, is also not yet at market economy levels. Russian companies take longer to issue audited ﬁnancial statements than do their Western and Japanese competitors, even though they are audited by the same four independent auditors. Thus, it appears that local factors play a more dominant role than the spread of technology (invasion by the Big-4 accounting ﬁrms and the technology and expertise they bring with them). This situation will likely change over time, as international accounting and auditing standards become a part of the corporate culture in Russia. There is evidence that this gap is closing rapidly. The Russian companies included in the present study take 25 days less to issue their audited ﬁnancial statements than they did just ﬁve years ago, which is a signiﬁcant improvement in spite of the Wilcoxon test results. But they still take much longer than their Western counterparts.

REFERENCES Accounting Principles Board (1970). Basic concepts and accounting principles underlying financial statements of business enterprises: Statement No. 4. New York: American Institute of Certiﬁed Public Accountants. Annaert, J., DeCeuster, M. J. K., Polﬂiet, R., & Van Campenhout, G. (2002). To be or not be y ‘‘too late’’: The case of the Belgian semi-annual earnings announcements. Journal of Business Finance & Accounting, 29(3 & 4), 477–495. Anon. (2005). Fortune Global 500. Fortune, 152(2), 119–125 (July 25). Ashbaugh, H., Johnstone, K. M., & Warﬁeld, T. D. (1999). Corporate reporting on the Internet. Accounting Horizons, 13(3), 241–257. Ashton, R. H., Graul, P. R., & Newton, J. D. (1989). Audit delay and the timeliness of corporate reporting. Contemporary Accounting Research, 5(2), 657–673. Atiase, R. K., Bamber, L. S., & Tse, S. (1989). Timeliness of ﬁnancial reporting, the ﬁrm size effect, and stock price reactions to annual earnings announcements. Contemporary Accounting Research, 5(2), 526–552. Ball, R., & Brown, P. (1968). An empirical evaluation of accounting income numbers. Journal of Accounting Research, 6, 159–178.

388

ROBERT W. MCGEE

Ball, R., Kothari, S. P., & Robin, A. (2000). The effect of international institutional factors on properties of accounting earnings. Journal of Accounting and Economics, 29(1), 1–51. Basu, S. (1997). The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics, 24, 3–37. Bates, R. J. (1968). Discussion of the information content of annual earnings announcements. Journal of Accounting Research, 6(Suppl.), 93–95. Beaver, W. H. (1968). The information content of annual earnings announcements. Journal of Accounting Research, 6(Suppl.), 67–92. Blanchet, J. (2002). Global standards offer opportunity. Financial Executive, March/April, 28–30. Brown, P., & Kennelly, J. W. (1972). The information content of quarterly earnings: An extension and some further evidence. Journal of Business, 45, 403–415. Chai, M. L., & Tung, S. (2002). The effect of earnings-announcement timing on earnings management. Journal of Business Finance & Accounting, 29(9 & 10), 1337–1354. Chambers, A. E., & Penman, S. H. (1984). Timeliness of reporting and the stock price reaction to earnings announcements. Journal of Accounting Research, 22(1), 21–47. Davies, B., & Whittred, G. P. (1980). The association between selected corporate attributes and timeliness in corporate reporting: Further analysis. Abacus, 16(1), 48–60. DeCeuster, M., & Trappers, D. (1993). Determinants of the timeliness of Belgian financial statements (Working paper). Antwerp, Belgium: University of Antwerp. (Cited in Annaert et al., 2002) Demos, T. (2006). The Russia 50: The country’s largest public companies. Fortune, 153(2), 70–71. Dwyer, P. D., & Wilson, E. R. (1989). An empirical investigation of factors affecting the timeliness of reporting by municipalities. Journal of Accounting and Public Policy, 8(1), 29–55. European Bank for Reconstruction and Development. (2005). Transition report 2005. Business in transition. London. Financial Accounting Standards Board (1980). Statement of financial accounting concepts No. 2: Qualitative characteristics of accounting information. Stamford, CT: Financial Accounting Standards Board. Gigler, F. B., & Hemmer, T. (2001). Conservatism, optimal disclosure policy, and the timeliness of ﬁnancial reports. Accounting Review, 76(4), 471–493. Givoli, D., & Palmon, D. (1982). Timeliness of annual earnings announcements: Some empirical evidence. Accounting Review, 57(3), 486–508. Han, J. C. Y., & Wang, S. (1998). Political costs and earnings management of oil companies during the 1990 Persian Gulf crises. Accounting Review, 73, 103–117. Han, J. C. Y., & Wild, J. J. (1997). Timeliness of reporting and earnings information transfers. Journal of Business Finance & Accounting, 24(3 & 4), 527–540. Haw, I., Qi, D., & Wu, W. (2000). Timeliness of annual report releases and market reaction to earnings announcements in an emerging capital market: The case of China. Journal of International Financial Management and Accounting, 11(2), 108–131. Hendriksen, E. S., & van Breda, M. F. (1992). Accounting theory (5th ed.). Burr Ridge, IL: Irwin. Jindrichovska, I., & Mcleay, S. (2005). Accounting for good news and accounting for bad news: Some empirical evidence from the Czech Republic. European Accounting Review, 14(3), 635–655.

Corporate Governance in Russia

389

Keller, S. B. (1986). Reporting timeliness in the presence of subject to audit qualiﬁcations. Journal of Business Finance & Accounting, 13(1), 117–124. Kenley, W. J., & Staubus, G. J. (1974). Objectives and concepts of ﬁnancial statements. Accounting Review, 49(4), 888–889. Kochetygova, J., Popivshchy, N., Shvyrkov, O., Kazakov, D., Kuzmina, O., & Rozanova, A. (2005, September 21). Russian transparency and disclosure survey 2005: Continuing progress in transparency, but mainly among weaker disclosers. Moscow: Standard & Poor’s Governance Services. Krishnan, G. V. (2005). The association between Big 6 auditor industry expertise and the asymmetric timeliness of earnings. Journal of Accounting, Auditing & Finance, 20(3), 209–228. Kross, W., & Schroeder, D. A. (1984). An empirical investigation of the effect of quarterly earnings announcement timing on stock returns. Journal of Accounting Research, 22(1), 153–176. Kulzick, R. S. (2004). Sarbanes-Oxley: Effects on ﬁnancial transparency. S.A.M. Advanced Management Journal, 69(1), 43–49. Lawrence, J. E., & Glover, H. D. (1998). The effect of audit ﬁrm mergers on audit delay. Journal of Managerial Issues, 10(2), 151–164. Leventis, S., & Weetman, P. (2004). Timeliness of ﬁnancial reporting: Applicability of disclosure theories in an emerging capital market. Accounting and Business Research, 34(1), 43–56. McGee, R. W. (2005). Financial reporting in the Russian energy sector. Russian/CIS Energy & Mining Law Journal, 3(6), 23–28. McGee, R. W. (2006a). Timeliness of financial reporting in transition economies (Andreas School of Business working paper). Miami Shores, FL: Barry University. McGee, R. W. (2006b). Transparency and disclosure in Russia, in corporate governance in Russia and Poland in comparative perspective (Tomasz Mickiewicz, Ed.). New York: Palgrave Macmillan. McGee, R. W. (2006c). Timeliness of ﬁnancial reporting in the energy sector. Russian/CIS Energy & Mining Law Journal, 4(2), 6–10. McGee, R. W., & Preobragenskaya, G. G. (2004, July 9–11). Corporate governance in transition economies: The theory and practice of corporate governance in Eastern Europe. In Proceedings of the Global Conference on Business Economics (pp. 690–711). Amsterdam: Association for Business and Economics Research. (Available at www.ssrn.com) McGee, R. W., & Preobragenskaya, G. G. (2005). Accounting and financial system reform in a transition economy: A case study of Russia. New York: Springer. McGee, R. W., & Preobragenskaya, G. G. (2006). Accounting and financial system reform in Eastern Europe and Asia. New York: Springer. Organisation for Economic Cooperation and Development. (1998). Global corporate governance principles. Paris. Pope, P. F., & Walker, M. (1999). International differences in the timeliness, conservatism, and classiﬁcation of earnings. Journal of Accounting Research, 37(Suppl.), 53–87. Preobragenskaya, G. G., & McGee, R. W. (2004). Corporate governance in a transition economy: A case study of Russia. Corporate Ownership and Control, 1(4), 61–71 (Reprinted in Russian/CIS Energy & Mining Law Journal, 2004, Issue 4, 21–28). Prickett, R. (2002). Sweet clarity. Financial Management, September, 18–20. Rees, W. P., & Giner, B. (2001). On the asymmetric recognition of good and bad news in France, Germany and the UK. Journal of Business Finance & Accounting, 28(9 & 10), 1285–1332.

390

ROBERT W. MCGEE

Soltani, B. (2002). Timeliness of corporate and audit reports: Some empirical evidence in the French context. International Journal of Accounting, 37, 215–246. Trueman, B. (1990). Theories of earnings-announcement timing. Journal of Accounting and Economics, 13, 285–301. Whittred, G. P. (1980). Audit qualiﬁcation and the timeliness of corporate annual reports. Accounting Review, 55(4), 563–577. Whittred, G. P., & Zimmer, I. (1984). Timeliness of ﬁnancial reporting and ﬁnancial distress. Accounting Review, 59(2), 287–295. World Bank. (2001). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Croatia. World Bank. (2002a). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Bulgaria. World Bank. (2002b). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Czech Republic. World Bank. (2002c). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Georgia. World Bank. (2002d). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Latvia. World Bank. (2002e). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Lithuania. World Bank. (2003a). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Hungary. World Bank. (2003b). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Slovak Republic. World Bank. (2004a). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Moldova. World Bank. (2004b). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Romania. World Bank. (2004c). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Slovenia. World Bank. (2005a). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Armenia. World Bank. (2005b). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Azerbaijan. World Bank. (2005c). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Macedonia. World Bank. (2005d). Report on the observance of standards and codes (ROSC). Corporate Governance Country Assessment, Poland.

PART VI: THE FINANCIAL SERVICES SECTOR

This page intentionally left blank

MERGERS AND CONSOLIDATION OF FINANCIAL SERVICE FIRMS: GLOBAL TRENDS AND STRATEGIES FOR VALUE CREATION Edward C. Boyer and Jongmoo Jay Choi ABSTRACT The financial services industry is experiencing rapid consolidation globally. Consolidation has proceeded not only in the same market but also across different market segments and across national boundaries. In this paper, we (a) outline the general trend of the mergers and acquisitions (M&As) and consolidation of the financial service industry in the U.S. and in the global economy; (b) identify and analyze the reasons that contribute to the consolidation of the financial service industry; (c) examine some cases of successful and unsuccessful financial service M&As; and (d) arrive at some strategic implications.

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 393–417 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07016-6

393

394

EDWARD C. BOYER AND JONGMOO JAY CHOI

1. INTRODUCTION As the global economic environment changes in the ﬁnancial services industry, ﬁnancial institutions are positioning themselves to compete. While a few ﬁrms are going lean and mean, stripping themselves of some functions or divisions in order to focus on a particular market niche (e.g., State Street Bank & Trust Company that is focusing on servicing ﬁnancial assets as an investment manager), the most salient feature of competitive posturing has been a trend toward the consolidation and rapid development of bigger, fullservice, global ﬁnancial institutions. The ﬁnancial services industry has experienced rapid consolidation globally. The form of consolidations in the U.S. has been primarily in-market (that is, banks merging with banks, and insurance companies merging with insurance companies, and so forth), but there is also signiﬁcant diversiﬁcation across ﬁnancial services as commercial banks merge with securities ﬁrms and insurance companies, globally as well as in the U.S. Interesting patterns of consolidation have followed these developments as ﬁrms pursued new opportunities and defended themselves against threats relative to their perceived strengths and weaknesses. Some ﬁrms have strategically positioned themselves to offer a complete array of services, while others have stuck-to-their-knitting. Whether mergers and acquisitions (M&As) are inmarket or diversiﬁed across different market segments may have important implications for post-combination performance as well as the nature of global competition in the ﬁnancial sector. To some extent, the consolidation of the ﬁnancial service industry reﬂects the general M&A activity in the global economy. Table 1 shows that the pace of M&As in the U.S. has increased signiﬁcantly both in volume and monetary terms over time in the 1990s, although that of domestic M&As in the U.S. has moderated since 1998 because of the uncertainty created by the Asian ﬁnancial crisis and ensuing market volatility. The list of largest bank groups in Tables 2A and 2B indicates that the consolidation of the Japanese banking ﬁrms has progressed signiﬁcantly so as to reclaim some of the top spots, which were lost due to the market collapse in the early 1990s when the bubble burst. In this paper, we (a) outline the general trend of M&As and consolidation of the ﬁnancial service industry in the U.S. and in the global economy; (b) identify and analyze the reasons that contribute to the consolidation of the ﬁnancial service industry; (c) classify and examine the cases of successful and unsuccessful ﬁnancial service M&As; and (d) arrive at some strategic implications.

Year

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

U.S. Domestic Acquisitions, Foreign Acquisitions of U.S. Companies and U.S. Acquisitions Overseas.

U.S. Acquisitions of U.S. Companies

Foreign Acquistions of U.S. Companies

U.S. Acquisitions Overseas

Numbers of Deals

Value ($bil)

Numbers of Deals

Value ($bil)

Numbers of Deals

Value ($bil)

3,642 3,781 4,213 5,155 6,595 7,562 8,896 10,459 9,319 8,505

141.4 122.8 176.5 278.9 389.9 573.5 778.7 1,354.8 1,424.9 1,747.5

539 406 394 513 634 684 837 982 1,151 1,196

30.8 16.1 21.0 48.2 55.2 79.7 90.9 234.0 266.5 340.0

482 548 635 762 1,035 1,160 1,401 1,688 1,617 1,502

15.3 15.6 18.2 23.4 46.8 62.0 79.8 119.7 153.8 135.2

Mergers and Consolidation of Financial Service Firms

Table 1.

Source: M&A Almanac, February 2001.

395

396

EDWARD C. BOYER AND JONGMOO JAY CHOI

Table 2A.

Largest Banking Firms by Asset Size in Billions of U.S. Dollars. 2004

UBS Citigroup Mizuho Financial Group HSBC Holdings Cre´dit Agricole BNP Paribus JPMorgan Chase Deutsche Bank Royal Bank of Scotland Bank of America

1995

1,533 1,484 1,296 1,277 1,243 1,234 1,157 1,144 1,119 1,110

Deutsche Bank Sanwa Bank Sumitomo Bank Dai-Ichi Kangyo Bank Fuji Bank Sakura Bank Mitsubishi Bank Norinchukin Bank Cre´dit Agricole ICBCa

503 501 500 499 487 478 475 430 386 374

1985 Citicorp Dai-Ichi Kangyo Bank Fuji Bank Sumitomo Bank Mitsubishi Bank Banque Nationale de Paris Sanwa Bank Cre´dit Agricole Bank America Cre´dit Lyonnais

167 158 142 135 133 123 123 123 115 111

Source: The Banker. a Industrial and Commercial Bank of China.

Table 2B.

Dozen Biggest Banking Acquisitions.

Announced

Acquiring Company

April 6, 1998 April 13, 1998 June 8, 1998 September 13, 2000 April 13, 1998 November 18, 1997 March 14, 1999 August 29, 1997 October 18, 1995 April 30, 1999 August 28, 1995 August 30, 1996

Travelers Group Nations Bank Norwest Chase Manhattan Banc One First Union Fleet Financial Nations Bank Wells Fargo Firstar Chemical Bank Nations Bank

Target Company

Transaction Valuea ($bil)

Citicorp Bank America Wells Fargo Capital J. P. Morgan & Co. First Chicago NBD CoreStates Financial BankBoston Bamett Bank First Interstate Bancorp Mercantile Bancorp Chase Manhattan Boatman’s Bancshares

72.50 61.60 34.30 34.30 29.60 17.10 15.90 14.80 10.90 10.60 10.40 9.60

Source: Tompson Financial Services; WSJ research, September 14, 2000. a Ranked by value at announcement; excluding net debt of target company.

In Section 2 we review the patterns in M&A activity globally. In Section 3 the global environmental factors contributing to the competitive positioning of ﬁrms and the consolidation of ﬁnancial service industry is examined. In Section 4 we discuss motives for M&As. Section 5 includes an analysis of strategies and market valuation. Section 6 examines several cases to illustrate successful cases as well as difﬁculties in realizing expected synergistic beneﬁts. Section 7 concludes with some strategic implications.

Mergers and Consolidation of Financial Service Firms

397

2. CONSOLIDATION AND STRUCTURE Bank consolidation in the U.S. from December 31, 1992 to December 31, 2000 shows a decline of 27.5% from 11,462 institutions to 8,315 and has remained roughly the same in 2005 at 8,332 institutions (Table 3). Table 3 also indicates a tendency toward the megamerger. During the period 1992– 2000, the proportion of the number of smallest bank groups (asset size of $100 million or less) has decreased from 72% to 58%, and to 44% as of 2005. At the same time, the number of banks with asset size of $10 billion or more has increased from 51 to 82 in 2000 and to 118 banks in 2005. More important, these megabanks controled 74% of all banking assets at the end of 2005, compared with 41% at the end of 1992. In addition to changes in bank size, there has been an important change in the nature of their business. Table 4 shows the changes in service offering by type of ﬁnancial institution in 1999 as opposed to 1950, showing a pattern of transition from a segmented to an integrated ﬁnancial sector. The only product that all ﬁnancial institutions were allowed to offer in 1950 (except for ﬁnance companies) was a savings product. Other services were permitted only for certain classes of ﬁnancial ﬁrms. In 1999, all major categories of ﬁnancial institutions (depository institutions, insurance companies, securities ﬁrms, and ﬁnance companies) could offer a complete line of ﬁnancial services (payment, savings, ﬁduciary services, business and consumer lending, debt and equity underwriting, and insurance and risk management products). Even specialized asset management institutions such as pension funds and mutual funds could provide many of these ﬁnancial services in 1999. The pattern of consolidation by institution category is illustrated for the U.S. and for Europe in Tables 5 and 6. Table 5 on the U.S. indicates the preponderance of in-market M&As: 52% of all ﬁnancial M&As in the U.S. are in-market mergers of commercial banks, followed by in-market mergers of insurance companies (19%) and securities ﬁrms (16%). Only about 13% have been mergers across different institutions, reﬂecting regulatory restrictions in the U.S. that existed during the period 1985–1997. The second panel of Table 5 regarding international (U.S. – non-U.S.) mergers, however, shows a somewhat more integrated picture. Compared to domestic mergers, the in-market international combinations are more important for insurance companies (37% of the total) and securities ﬁrms (21%) rather than banks, which comprise only 14% of the total value. This is in stark contrast to domestic M&As, which are dominated by bankto-bank mergers. It is also noteworthy that 28% of all international M&As

398

Table 3.

Changes in Banking Firms by Asset Size from 1992 to 2000 to 2005. Asset Size $100 million–$1 billion

$1 billion–$10 billion

$10 billion +

Total

December 31, 1992 Number of banks Percent of U.S. banks Total assets Percent of total bank assets

8,292 72.34% 346,028,309 9.87%

2,790 24.34% 680,187,216 19.40%

329 2.87% 1,034,161,182 29.50%

51 0.44% 1,445,286,661 41.23%

11,462 1 3,505,663,368 100%

December 31, 2000a Number of banks Percent of U.S. banks Total assets Percent of total bank assets

4,842 58.23% 231,194,141 3.71%

3,078 37.02% 773,009,474 12.39%

313 3.76% 884,112,777 14.17%

82 0.99% 4,350,396,442 69.73%

8,315 1 6,238,712,834 100%

December 31, 2005b Number of banks Percent of U.S. banks Total assets Percent of total bank assets

3,863 43.74% 200,700,000 1.85%

4,339 49.13% 1,247,600,000 11.47%

512 5.80% 1,393,000,000 12.81%

118 1.34% 8,035,800,000 73.88%

8,832 1 10,877,100,000 100%

a

Source: FDIC – Statistics on Depository Institutions Report. Source: FDIC – Statistics on Depository Institutions Report.

b

EDWARD C. BOYER AND JONGMOO JAY CHOI

$0–$100 million

1950 Depository institution Insurance companies Finance companies Securities ﬁrms Pension funds Mutual funds 1999 Depository institution Insurance companies Finance companies Securities ﬁrms Pension funds Mutual funds

Financial Products and Services by Different Financial Firms, 1999 versus 1950. Payment Services

Savings Products

Fiduciary Services

Business Lending

Consumer Lending

X

X X

X

X

X

X X X X X

X X X

X

X X X X X X

X X X X X X

X X X X X

Underwriting Equity

Underwriting Debt

Insurance and Risk Management Products

X X

X X X X

X

X

#

#

#

#

#

#

X

X

X X X X X X

Mergers and Consolidation of Financial Service Firms

Table 4.

Source: Anthony Saunders (2000). Financial Institution Management (p. 3). McGraw-Hill. Minor involvement. # Selective involvement through afﬁliates.

399

400

Table 5.

EDWARD C. BOYER AND JONGMOO JAY CHOI

Value of Target Institutions in M&A Activity in Financial Services, U.S. and U.S. – non-U.S.

Target Institution

Commercial Banks

Securities Firms

Insurance Companies

241 51.77% 15 3.22% 0.2 0.04%

6 1.29% 74 15.90% 27 5.80%

0.3 0.06% 14 3.01% 88 18.90%

U.S. – Non-U.S. Acquiring Institution Banks 9.5 13.59% Securities 4.4 6.29% Insurance 0.2 0.29%

3 4.29% 14.7 21.03% 7.7 11.02%

0.6 0.86% 3.9 5.58% 25.9 37.05%

U.S. Acquiring Institution Banks Securities Insurance

Note: The percentages are based on total M&A activity during the period 1985–1997. Source: Berger et al. (1999, p. 142).

in the U.S. are across market segments involving ﬁnancial institutions of different types, compared with only 13% for the domestic M&As. Table 6 shows a pattern of consolidation in Europe. An interesting result here is the relative importance of insurance and securities ﬁrms as acquirers. Although bank-to-bank mergers are still the largest percentage (36%), an additional 12% represent acquisitions of securities and insurance ﬁrms by banks. Acquisitions by securities ﬁrms and insurance companies are also quite substantial – 27% and 26%, respectively. International mergers, either in Europe or with non-European ﬁrms show even greater diversity of the M&A activity. For intra-Europe mergers, acquisitions by insurance companies – either in-market insurance–insurance mergers or across-market insurance company mergers with banks – dominate at 58%. For Europe to non-Europe combinations, acquisitions by securities ﬁrms are largest at 44%, larger than the same by banks or insurance companies. An interesting feature of the recent combinations is that they have involved large ﬁrms and ﬁrms of different types, which would have not been possible due to legal reasons a few yearsago. Also, during this period, the eight-ﬁrm concentration ratio increased from 22.3% to 35.5% (Berger, Demstz, & Strahan, 1999), as many consolidations were ‘‘megamergers.’’

Mergers and Consolidation of Financial Service Firms

Table 6.

401

Value of Target Institutions in M&A Activity in Europe, intraEurope, and Europe – non-Europe.

Target Institution

Commercial Banks

Securities Firms

Insurance Companies

Europe Acquiring Institution Banks 89 36.03% Securities 9 3.64% Insurance 20 8.10%

23 9.31% 19 7.69% 24 9.72%

11 4.45% 6 2.43% 46 18.62%

Intra-Europe Acquiring Institution Banks 15 17.90% Securities 8.7 10.38% Insurance 0.4 0.48%

4.3 5.13% 5.8 6.92% 1.1 1.31%

11.2 13.37% 0.3 0.36% 37 44.15%

15.6 15.55% 15.9 15.85% 12.9 12.86%

1 1.00% 3.1 3.09% 32.7 32.60%

Europe – non-Europe Acquiring Institution Banks 14.5 14.46% Securities 4.3 4.29% Insurance 0.3 0.30%

Note: The percentages are based on total M&A activity during the period 1985 – 1997. Source: Berger et al. (1999, p. 142).

Recently, the industry has realized ‘‘supermegamergers’’ – M&As between ﬁrms with assets of over $100 billion each. Four large banking mergers occurred in 1998 alone: Bank America–Nations Bank, Citicorp–Travelers, Banc One–First Chicago, and Norwest–Wells Fargo (see Table 2B). The trend toward M&As and consolidation has been accompanied by bank failures and creations as well as globalization. The bank failures were most pronounced during the late 1980s in the U.S., occurring at a rate exceeding 200 banks a year, although this trend has declined in the late 1990s, and 2005 witnessed no bank failures, a ﬁrst for any calendar year. At the same time new banks have ranged from 228 in 1988 to a low of 48 in 1994 and back to 207 in 1997, averaging better than 135 de novo banks per year (Berger et al., 1999). Changes in bank consolidation among the larger U.S. banks have been particularly phenomenal over time (Table 7).

402

EDWARD C. BOYER AND JONGMOO JAY CHOI

Table 7. Banking Consolidations. J.P. Morgan

1990-95

96

97

98

99

2000

Manufacturers Hanover Trust Chemical Bank

J. P. Morgan Chase

Chemical Bank Chase Manhattan Bank Chase Manhattan Bank

Bank of America Continental Bank

Bank of America

Security Bank

Bank of America

Nations Bank Nations Bank Bamett

First Union First Union Signet

First Union

Corestates

Bank Boston Bank Boston Bay Bank Fleet Boston Fleet Fleet Shawmut

Source: Wall Street Journal, March 14, 2001.

3. ENVIRONMENTAL FACTORS FOR CONSOLIDATION OF THE FINANCIAL SERVICES INDUSTRY There are three fundamental reasons that precipitated the consolidation of the ﬁnancial service industry in the U.S.: deregulation, technological progress, and globalization. And these events have helped to shape strategic choices as ﬁrms assess their core competencies relative to a changing environment.

Mergers and Consolidation of Financial Service Firms

403

3.1. Technological Progress Technical progress is the driver of change in industrial structure and competition. This has been especially true in the ﬁnancial services industry, even though, and perhaps partly because, this industry has arguably been the most regulated industry. By the end of the 1980s, and owing in part to the loan failures of the decade, banks were under heavy regulation, restricting their ability to compete in businesses that were traditionally theirs. While, as is argued below, technological progress has opened opportunities for banks, it has also provided threats. Along with the development of information technology, nonbank ﬁrms began to offer investment products, credit cards, and business and consumer loans that had previously been the turf of banks. Even nonﬁnance ﬁrms, such as large retailers, got into the mix by offering their own credit cards. Events in the 1980s raised the cost of capital for banks and made the process of securitization a substitute for many of the traditional loans made by banks. Firms that were engaged in underwriting and placement could do the work of the bank cheaper. In addition, money market mutual funds and equity mutual funds became popular substitutes for deposits at banks (Cumming, 1987). As interest income began to disappear, the need to replace it became important and banks had to ﬁnd new sources of income. New services for which banks could charge a fee became a critical source of revenue. And if there are to be economies of scale and scope to capture, then size could be critical. In the recent past, ﬁnancial services institutions have experienced scale economies from technological innovations, which are of both process and product. Process innovations include ATMs, online banking, e-mail billing, home banking, electronic funds transfer (CHIPS and Fedwire), and pointof-sale debit and credit card sales. On the product side there are derivative securities, off-balance sheet activities, and risk management tools. The question is how these innovations improve proﬁtability. The ﬁnancial intermediary’s technology can impact the ﬁrm’s interest margin and other net income in a variety of ways. Providing an increased array of services can enhance interest income. Banks provide cash management services that collect, transfer, and disburse cash and enable ﬁrms to have a real-time access to their working capital status and enjoy interest savings by controlling excess cash balances. Information technology allows ﬁrms to match customers while telemarketing and the Internet can provide

404

EDWARD C. BOYER AND JONGMOO JAY CHOI

insurance and bank products directly to the customer, with potential savings for customers and institutions. ATMs, point-of-sale debit cards, home banking, and related services provide efﬁciencies for retail customers. Similar efﬁciency is possible through networked computer systems, such as CHIPS and Fedwire, linking domestic and international inter-bank lending markets. Other income derives from off-balance-sheet activities and is increasingly computer driven. Customers can originate their own letters of credit electronically, and derivatives as well as primitive securities are priced with the aid of computers and then traded via the computer screen. Securityrelated back-ofﬁce activities become more efﬁcient since the collection and storage of customer proﬁles and ﬁnancial products are automated rather than handled manually or on paper. Technological change provided the tools for harnessing economies of scale and scope, but deregulation unleashed banks to do their bidding. 3.2. Deregulation The competitive environment in each country follows largely from the regulatory restrictions within those countries. For example, there are a signiﬁcantly greater number of banks in the U.S. than in other major industrialized countries on a per capita basis. However, the U.S. does not exhibit substantially more branches on a per capita basis. This fact clearly follows from the branching restrictions that existed in the U.S. prior to 1994. In Europe, banks are engaged in a wider variety of ﬁnancial services than in the U.S. The Riegle-Neal Interstate Banking and Branching Efﬁciency Act of 1994 prompted immediate consolidation within the U.S. From 1995 to 1998 a total of 1,015 holding company consolidations and 1,763 unassisted bank mergers and acquisitions occurred. The act does continue to limit consolidation to 30% of deposits within a given state and 10% nationally. Also, this act has not removed all barriers to expansion. Banks may not open branches across state lines except by acquiring existing banks within that state. Glass-Steagall restrictions have been drastically reduced due to the liberal treatment by the Fed of Section 20 subsidiaries even prior to its elimination in 1999. The improved regulatory environment permitted a combination of commercial and investment banking (e.g., Piper Jaffray and U.S. Bancorp) as well as an emergence of ﬁnancial conglomerates involving insurance, commercial, and investment banking services (e.g., Citicorp and Travelers/ Salomon Smith Barney). At the same time, increasingly, nonbank ﬁrms are

Mergers and Consolidation of Financial Service Firms

405

providing ﬁnancial services, leading to a universal banking combining banking and commerce functions (e.g., General Motors Acceptance Corporation). The trend toward integration and liberalization of ﬁnancial markets in the U.S. was formalized by the abolition of the Glass-Steagall act in 1999. The shape of regulation and the emergence of deregulation have also contributed to the globalization of markets. 3.3. Globalization of Markets Globalization is often cited as a reason for consolidation in ﬁnancial services. As markets for products and services become global, ﬁrms entering these markets require ﬁnancial services as they expand across borders. This globalization provides opportunities for domestic ﬁnancial service ﬁrms abroad. But differences in regulation and information costs have not made cross-border activity easy for banks. Focarelli and Pozzolo (2001) ﬁnd that ‘‘ycross-border M&As are rarer in the banking industry than in other sectors, possibly owing to the importance of information asymmetries in banking relationships and to regulatory restrictions.’’ Also, Buch and DeLong (2004) ﬁnd that information costs and regulation impact cross-border activity, and state that, ‘‘y policy makers can create environments that encourage cross-border activity, but information cost barriers must be overcome even in (legally) integrated markets.’’ Most foreign activity by U.S. banks is of the offshore variety. Citibank is an exception, offering credit cards, banking by phone, and ATMs, which local banks do not provide. In 1985 there were over 160 U.S. banks with overseas ofﬁces but by 1998 that number declined by 50%. In 1998 there were 243 foreign banks in the U.S., but this was not without notable losses by some banks, forcing them to retreat (in 1991 Barclay’s lost $397 million). Traditionally, international banking and ﬁnance has been motivated by the necessity to escape the regulatory burden as well as the business need to follow the customers. In the 1960s and 1970s, U.S. banks established subsidiaries in the Eurocurrency market where the absence of reserve requirements and deposit insurance fees permitted improved proﬁt margins. Also, the elimination of interest rate ceilings allowed for more aggressive competition for deposits. Non-U.S. banks entered U.S. markets primarily to serve non-U.S. corporations for their international business ﬁnancing needs. Differing regulatory restrictions in each country means that it permits particular national banks to maintain a competitive advantage globally. Canadian commercial banks were subject to less regulation than U.S. banks as they could expand nationally and were not restricted from investment

406

EDWARD C. BOYER AND JONGMOO JAY CHOI

banking operations. Moreover, with the exception of U.S. banks, Canadian banks were protected from foreign entry, until 1999. European banks are also freer to engage in investment banking operations such as underwriting corporate securities. Japanese banks have less freedom than European banks in offering investment banking services, although they are permitted to invest in corporate equity using depositor funds. Starting in the early 1990s, however, there have been some regulatory changes intended to promote cross-border activity. In North America, the North American Free Trade Agreement of 1994 stimulated cross-border bank M&A activity between the U.S., Canada, and Mexico. Additionally, the abolition of the 1933 Glass-Steagall Act in the U.S. in 1999 is an attempt to level the ﬁeld of global competition in recognition of new market realities. Domestically, it put all ﬁnancial ﬁrms – commercial banks, securities ﬁrms, and insurance companies – on the same playing ﬁeld. Globally, it implies an application of more uniform global regulations. This is an extension of the earlier International Banking Facilities Act that places certain activities of U.S. banks on a par with foreign commercial banks operating in the Eurocurrency markets. The Single European Act places all European banks operating in the European Union under the same set of rules. Also, the uniform capital guidelines by the Bank of International Settlement require banks of 12 industrialized countries to meet the same capital constraints. U.S. regulators, however, are a bit more cautious about reducing capital requirements and will adopt Basel 2 more slowly than Europe. Nonetheless, cross-border M&A activity in the EU exhibits a very mixed picture. In the western European countries, banking M&A has been limited to domestic combinations, so that by 2004 only 23% of bank assets were controlled by foreign banks. In contrast, in eastern European countries, foreign ownership averaged about 70%. (Abdelal and Bruner, 2005) However, this may be changing. Despite attempts from the Bank of Italy to prevent it, the Banco Nazionale del Lavoro and Banca Antonveneta were taken over by foreign banks. The economic decline and deteriorating asset quality of the Japanese banks in the 1990s has also led to the dramatic increase in domestic and global combinations of the Japanese ﬁnancial service ﬁrms. Out of their need to compete globally as well as in their domestic markets that have been liberalized, the Japanese banks have consolidated rapidly and established strategic alliances with U.S. and international ﬁnancial institutions. For instance, J.P. Morgan set up a joint venture with Dai-Ichi Kangyo Bank to serve the retail end as well as corporate business in Japan, while Citigroup has set up a strategic alliance with the Bank of Tokyo-Mitsubishi and Nikko

Mergers and Consolidation of Financial Service Firms

407

Securities. Lately, the Asian ﬁnancial crisis of 1997 brought about the collapse and restructuring of the ﬁnancial sector in emerging market countries, which is still going on at this point in time.

4. MOTIVES FOR MERGERS Why are banks consolidating? One possibility is that they want to achieve market power through product differentiation. But the room for differentiating the most common banking activities is marginal at best. The most often cited reasons, however, are to achieve economies of scale and scope, and this is where size may be critical. Banks may also merge to erase excess capacity or to rectify poor performers, sometimes assisted by regulators. 4.1. Market Power Market power requires that ﬁrms differentiate their products or services from their competitor’s. Commerce Bank follows this strategy, however, by internal growth and not through merger or acquisition. By maintaining extended and weekend hours, Commerce pays about a half point less on deposits. However, any competitive advantage that an entity creates must be sustainable, that is, difﬁcult to copy or duplicate. What makes strategies difﬁcult to duplicate are the costs of acquiring skills, knowledge, or assets upon which the product differentiation rests. But, since banks are less regulated, and markets are more contestable, developing specialized resources and capabilities is increasingly important. Perhaps, this is the foundation for the strengths of Goldman Sachs. The desire to gain market power may also explain the importance of inmarket M&A activity globally. For example, U.S. banks have been innovative in providing swaps and other derivative-based products relative to foreign banks. To the extent that they maintain their comparative advantages in skills and technologies in these products, entry into foreign markets with such products could give U.S. banks market power in these product offerings. Citibank is one example. 4.2. Economies of Scale In ﬁnancial services there are several sources of efﬁciency that can be gained from in-market combinations. Duplicate branching in banks as well as back-ofﬁce activities can be consolidated. Some studies estimate scale

408

EDWARD C. BOYER AND JONGMOO JAY CHOI

economies of approximately 20% of costs on banking ﬁrms of sizes between $10 and $25 billion in assets (Berger & Mester, 1997). This supports the notion that acquiring ﬁrms are spreading excess managerial capacity, at least in ﬁrms of this size. Also, there is possibly an indirect diversiﬁcation beneﬁt from size: the larger the ﬁnancial institution, the more diversiﬁed would be the loan portfolio, and the lower the risk. In general, acquiring ﬁrms are more proﬁtable (Peristiani, 1993) and more efﬁcient (than target ﬁrms) in banking (Berger & Humphrey, 1992; Pilloff & Santomero, 1998) and in life insurance companies (Cummins, Tennyson, & Weiss, 1999). 4.3. Economies of Scope By diversifying across markets, that is, some combination of banking, insurance, and/or securities, the institution may be able to capture economies of scope. Sources of economies of scope are shared resources, capabilities, or customers. For example, customer proﬁles acquired on customers by a bank may become important information for providing securities services and insurance products to these same customers. It is less clear whether economies of scale can be realized for combinations involving ﬁnancial and nonﬁnancial institutions. In order to achieve economies of scope, ﬁrms can be expected to sell similar products (that exploit a similar technology) or sell to similar markets (exploiting similar customers). However, many companies (Union Carbide, Allison-Chambers, and Gulf and Western) follow neither of these strategies. Another possible source of scope economies could result from speciﬁc managerial skills (information systems, ﬁnance, and marketing) applied to diverse markets or products. Whether ﬁrms are successful in achieving synergies from their managerial skills is often difﬁcult to assess. In the ﬁnancial services industry, the popular strategic intent of diversiﬁcation is to offer one-stop shopping for customers. The question remains, however, whether this would achieve some economies of scope or scale and, therefore, be value enhancing. 4.4. Excess Capacity or Poor Performance Mergers may also occur when ﬁrms substitute acquisition for internal expansion or R&D, like in the case of Cisco Systems. In ﬁnance, consolidation also occurs due to the need to erase excess capacity. Mergers due to excess capacity may occur for several reasons: First, ﬁrms may accumulate excess managerial capability that can be applied to acquisitions, that is proﬁtable

Mergers and Consolidation of Financial Service Firms

409

ﬁrms acquiring less efﬁcient ﬁrms. Second, excess capacity may occur due to over-investment or industry retrenchment and competition. The latter occurrence may also lead to ﬁnancial distress of post-merger ﬁrms in the industry. Conversely, it is also possible that poor performance is a motivation for consolidation. In Japan as well as in Asian emerging countries, ﬁnancial service ﬁrms were forced to consolidate as a means of survival during the Asian ﬁnancial crisis because of inefﬁciency and poor performance. However, without an assurance of signiﬁcant restructuring and layoffs, which is resisted by labor unions, it remains doubtful whether a simple combination of two ailing ﬁrms would do the trick. A key for success for this type of combination is the resolution on the part of both management and labor and other interested parties concerned that restructuring is the only way to realize potential economies and synergies.

5. CORPORATE ACQUISITION STRATEGY AND VALUE CREATION Empirical studies of nonﬁnancial ﬁrms suggest that successful M&As correlate with the degree of ‘‘relatedness,’’ which is a measure of the focus of the ﬁrm’s corporate-level strategy. (Seth, 1990; Shelton, 1998; Singh & Montgomery, 1987) Relatedness is measured by the degree to which a ﬁrm’s revenues derive from different lines of business. Success (as measured by stock prices, accounting earnings, and long-term performance) is greater for strategies following a market focus. The strategic intent is to capture economies of scale or market power and can be achieved through a horizontal or vertically integrated M&A strategy. The literature suggests that because the focus of the ﬁrm is on a single product market, R&D is also likely to be more focused. This may account for the dominance of this strategy relative to other strategies, at least outside the ﬁnancial services industry. The evidence on improved proﬁtability in banking is not supportive. Amel, Barnes, Panetta, and Salleo (2004), in a review of the empirical evidence, report that ‘‘The empirical evidence suggests that commercial bank M&As do not signiﬁcantly improve cost and proﬁt efﬁciency and, on average, do not generate signiﬁcant shareholder value.’’ But, perhaps, the results of R&D have not yet kicked in and ﬁrms have only concentrated on certain operating efﬁciencies that may be gained by eliminating duplicating functions. And, it may be that failure to ﬁnd improved proﬁt or cost efﬁciency relates to the time that it takes to turn around a less efﬁcient target.

410

EDWARD C. BOYER AND JONGMOO JAY CHOI

Bank managers typically assert that the motive behind their merger activity is to achieve economies of scale and scope. The realization of scale and/or scope economies may clearly be related to whether M&As are inmarket or diversifying combinations. Furthermore, the degree to which business units are related is in turn a function of how they are related product-wise as well as geographically. DeLong (2001) categorizes bank mergers as diversiﬁed or focused and ﬁnds that the market rewards market concentration not diversiﬁcation. If the combination is not within the same geographic market but is within the same product market (a geographic extension combination), then the institution can achieve greater diversiﬁcation of risk, unless the new geographic region has a higher percentage of defaults. Diversifying combinations that stretch across product markets (such as a bank merging with an insurance company or a securities ﬁrm) are diversifying against cyclical swings in different sectors of the industry. Additionally, such mergers expect to capture economies of scope through cross-selling. Regardless of which form a combination assumes, the fundamental issue is whether it will create value. The empirical evidence is mixed. Most studies do not ﬁnd evidence of signiﬁcant scale economies. One exception is Hughes, Mester, and Moon (2001), who suggest that this is due to the fact that these studies ignore the ﬁrm’s capital structure and risk. These authors argue that large banks are more diversiﬁed and can, therefore, take on more risk. When this additional risk is considered within the production function, then bigger is clearly better. Related diversiﬁed strategies intend to create strategic competitiveness by extending their resources, capabilities, and core competencies. The cost savings due to this type of strategy emanate from economies of scope. Firms can follow two operational strategies in pursuit of these economies: transferring skills or sharing activities. The merging of pharmaceutical ﬁrms is intended to share the R&D laboratories and production facilities. Financial service ﬁrms may achieve economies of scope through sharing information technology. Transferring skills is another viable strategy and one that Philip Morris successfully pursued by sharing its marketing strength in the acquisition of Miller Brewing and also Brown Foreman in its acquisition of Southern Comfort. Research shows that sharing activities and skill transfer can create value (Brush, 1996). But these strategies also entail costs. Sharing activities requires strategic control and coordination between business-unit managers. This implies that there are important interrelationships that may impact the outcome of both units (Argures, 1996). Skill transfer often requires the moving of key personnel into the acquired business unit and can be met with

Mergers and Consolidation of Financial Service Firms

411

some resistance. These costs must be considered when identifying the predicted economies of shared activities and skill transfer. The measurement of economies of scope is perhaps the most difﬁcult, in particular because the single-ﬁrm benchmark rarely exists. As reported by Amel et al. (2004), very little evidence to support economies of scope exists. However, as ﬁrms continue to develop specialized resources and capabilities, and different strategic groups such as universal and specialized banks emerge, economies of scope may become more apparent. Empirically, the success or failure of M&As can also be identiﬁed by the performance of market prices for merging ﬁrms. Evidence from the U.S. in the 1990s does indicate some value creation that earlier studies did not ﬁnd (Amel et al., 2004). Several possible explanations emerge: First, deals in the 1990s were largely consummated with cash, as opposed to stocks in earlier periods. Second, the 1990s exhibited a more attractive regulatory environment for mergers. For Europe, a slightly different picture emerges. Again Amel et al. (2004) cite empirical studies that indicate that markets ‘‘do not value cross-border consolidations.’’ But, it remains to be seen whether, and how well, the market anticipates synergies. 5.1. Problems in Achieving Success There are a number of reasons why merging ﬁrms have difﬁculty achieving success. These are integration difﬁculties, inadequate valuation of the target, excessive debt, inability to achieve synergy, and too much diversiﬁcation. Integration difﬁculties arise for several reasons: Often, underlying disparate corporate cultures arise, a problem that Daimler-Chrysler experienced. Firms may also have difﬁculty linking ﬁnancial and control systems. Inability to develop effective working relationships among employees may prevent companies from integrating. Finally, it may take time to determine and coordinate the status of executives. Inadequate valuation of targets represents another problem associated with achieving success. Whether the reason for inadequately valuing the target is due to poor due-diligence, improper forecast for future cash ﬂows, or overexuberance on the part of managers, paying too high a price is something from which mergers have difﬁculty recovering. First Union paid dearly for several of its early acquisitions, and even effective strategy that followed failed to bolster the stock price (Tully, 1999). Similarly, Sony and Matsushita, a world-class Japanese consumer electronic acquired movie studios in Hollywood on a strategic ground at the peak of the Japanese economy and currency, only to unload them in the bubble-burst period at a huge loss.

412

EDWARD C. BOYER AND JONGMOO JAY CHOI

Too much diversiﬁcation may stretch managerial capability too far, causing management to rely on ﬁnancial rather than on strategic control. Strategic control requires in-depth understanding of the business, and too much diversiﬁcation may make this impossible. Relying on ﬁnancial control may also have important long-term consequences since it may turn management attention to the short-run rather than the long-term.

6. CASES Several cases illustrating the strategic goals of mergers and potential difﬁculties in realizing synergies follow. 6.1. BankAmerica Acquires Continental Bank The acquisition of Continental Bank by BankAmerica (BAC) represents a strategic diversiﬁcation strategy. At the time of the acquisition, announced in January of 1994, it would be one of the largest banking mergers with a price of approximately $1.9 billion. This merger is considered a diversiﬁed strategy since it brings together a bank with a strong commercial business (BAC) and a bank with a strong corporate clientele (Continental). BAC’s strength was in its corporate banking business in which it invested considerably. It was second only to Citibank in size and maintained a very high credit rating. BAC’s client base was centered on the West Coast, which limited its ability to grow. Continental was based in Chicago and, with an asset base of $22.6 billion, was the second largest bank in the Midwest. Its customers were primarily large ﬁrms headquartered in Chicago. In 1984, the bank had to be bailed out by the FDIC but has since focused on its commercial business to rebound signiﬁcantly. However, Continental was not able to fully regain its credit rating, limiting the services that it could provide to its customers. The anticipated synergies from the merger were founded in the strong client base of Continental and the excellent credit rating of BAC. In addition, BAC had a very large low-cost deposit base resulting from its multistate branching. Financial analysts applauded the merger for the strategic ﬁt of BAC’s product line, credit rating, and branching network. Both banks were viewed as efﬁcient and Continental would bring new talent to BAC. This merger represents one that anticipated both economies of scale and economies of scope. The major question for success of the merger was, and is, whether the very different corporate cultures could be meshed. The success of the merger is

Mergers and Consolidation of Financial Service Firms

413

hard to assess, since after the BAC–Continental merger, Bank of America and NationsBank merged, and then acquired FleetBoston Financial in 2003 and MBNA in 2005. Prior to these additional mergers, this was the largest in-market banking combination in the U.S., but it remains to be seen whether potential conﬂict in management and corporate culture as well as information technology systems can be efﬁciently integrated without incurring too much cost. 6.2. The Merger of U.S. Bancorp and Piper Jaffray On December 15, 1987 Piper Jaffray announced its intention to merge with U.S. Bancorp. The projected synergies would come from a shared customer base, specialized knowledge, and combined products. The stock price reaction to the announcement was positive, as U.S. Bancorp’s stock rose 0.5% and Piper Jaffray’s stock price increased by 21.22%. U.S. Bancorp was a diversiﬁed company with operations in ﬁve separate segments: retail banking, its largest business; business banking and private ﬁnancial services to commercial and wealthy clients; corporate banking and commercial lending; corporate trust and institutional ﬁnancial services, one of the largest in the industry, and mutual funds. The bank followed a strategy of offering standardized products at a low cost. U.S. Bancorp also had a dominant position in credit card and trust operations. The major appeal to its customers, however, was its excellent customer service, for which it enjoyed an excellent reputation. Piper Jaffray was one of the largest regional brokerage ﬁrms, with four subsidiaries: Piper Jaffray, which is individual brokerage services (about 60% of its business); Piper Capital, an investment banking operation; Piper Trust, which is corporate and municipal debt; and Piper Jaffray Ventures, a venture capital arm. The strategy that Piper Jaffray followed was to offer a limited range of products of high quality, and high levels of service. The ﬁrm also worked hard to develop strong customer ties. Until 1994, Piper Jaffray had an excellent reputation, particularly for service and quality products, when its mutual funds ran into trouble. The post-merger structure would build on the strengths of both ﬁrms. Piper Jaffray’s securities underwriting, ﬁnancial services, and brokerage services would assume U.S. Bancorp’s ﬁxed income and capital markets business and corporate ﬁnance group. Also, U.S. Bancorp’s retail brokerage would report into Piper Jaffray’s brokerage subsidiary. Finally, Piper Jaffray’s Capital Management and Piper Trust would report into U.S. Bancorp’s First Asset Management First Trust.

414

EDWARD C. BOYER AND JONGMOO JAY CHOI

The potential merger difﬁculties resided in the different corporate cultures and management styles. U.S. Bancorp placed emphasis on standardized products and cutting costs. Piper was more interested in employee satisfaction and a ‘‘family’’ atmosphere. Given that Piper Jaffray’s employees were a very valuable asset, could U.S. Bancorp impose its management style without losing these assets? In addition, could U.S. Bancorp manage ﬁrms in two distinct industries and be able to allocate capital efﬁciently? 6.3. The Creation of Citigroup The merger of Citicorp and Travelers into Citigroup in 1997 resulted in the largest ﬁnancial services ﬁrm with 1998 assets of $697 billion. Table 8 describes the combined company. In contrast to the Bank of America merger, the Citigroup is a prime example of diversiﬁed merger within a ﬁnancial service industry, creating a ﬁnancial conglomerate including commercial banking, securities and investment banking, and insurance companies. There were three major projected synergies from the combination. The ﬁrst is customer proﬁles that can be used across divisions to prepare products and sales approaches for customers. Second, the combination projected economies of scale in R&D relative to computer technology, such as internet-based marketing and transactions services. Third, there would be considerable cost savings in consolidating computer and network operations. Table 8. Ciricorp Private banking Consumer banking Credit cards Corporate banking Travelers Consumer ﬁnance Life insurance Property and casualty Investing services

Citigroup 1997.

$100 billion in assets serving wealthy clients; Net income 1997 of $283 million. Thousands of branches worldwide; Net income 1997 of $435 million. 36 million customers; Net income 1997 of $696 million. Net income 1997 of $2,177 million. 2 million customers; Net income 1997 of $222 million. Primerica and Travelers also sell annuities and long-term care policies; Net income 1997 of $850 million. Net income 1997 of $957 million. Salomon and Smith Barney do investment banking and mutual funds; Net income 1997 of $1,075 million.

Mergers and Consolidation of Financial Service Firms

415

There are some major differences between the strategic approaches of the two companies. Citicorp emphasized advanced technology; it was responsible for the ATM machines. Traveler’s focused on low-cost, plain-vanilla systems. But this was only the beginning of the difﬁculties in realizing immediate synergies. The biggest problem faced by the merger in realizing synergies is in integrating information technology systems. In 1998, it was reported that it would take at least a year to integrate the computing systems of Salomon and Smith Barney. But that is small in comparison to the prediction that it would take almost another 10 years to fully integrate the combination. As of 1998, Citicorp had already spent approximately 10 years to merge its computers linking its banks in 100 countries. In 1997, Citicorp took an $889 million charge for restructuring that would eliminate 9,000 jobs and centralize operations. It also hired AT&T to merge its data networks. Additionally, the computers of Salomon Smith Barney are not integrated to Traveler’s Commercial Credit ﬁnance company.

7. STRATEGIC IMPLICATIONS Since the 1960s, the banking industry sustained a number of challenges that altered or contested its core competencies, providing banks with the opportunity to develop new competencies and strengthen old ones. Fundamentally, banks accepted deposits, as a part of the payments function, pooled these into loans and, by so doing, diversiﬁed risk. This gave banks preeminence in information. In the 1960s and 1970s, information technology created opportunities for nonbanks to challenge this dominance in gathering and processing information on credit risk. Hampered by outmoded regulations and faced with new competition provided by information technology, banks had to ﬁnd new sources of loans. The money market mutual fund was one such innovation. However, events of the 1980s and the continued development of information technology, provided competition from money market mutual funds, insurance annuities, and equity mutual funds, and the share of total assets in the ﬁnancial sector controlled by banks fell to less than 20%. But deregulation in the 1990s and continued globalization of markets provided opportunities for banks to grow and expand into the provision of new products. If we assume that the goal of M&A activity is value creation, that is synergies, then motives for combinations include spreading excess core competencies to achieve market power or greater efﬁciency, that is, scale or

416

EDWARD C. BOYER AND JONGMOO JAY CHOI

scope economies. New technologies provide the resources for developing new core competencies, which includes derivatives and other risk management tools. As such, the ﬁrm’s capabilities are largely intangibles, which give companies like Goldman Sachs a competitive edge. There are many factors that prevent successful M&As, in spite of the complementarities that may exist between ﬁrms. These factors include difﬁculties with integration, inadequate valuation, and too much diversiﬁcation. The literature abounds with examples of inadequate valuation and too much diversiﬁcation. On the other hand, the deregulation that provides an impetus to the creation of an integrated ﬁnancial service ﬁrm only began formally in 1999. Despite difﬁculties of integration, globalized ﬁnancial markets underscore the importance of the potential corporate advantage of creating a global, integrated, comprehensive ﬁnancial service institution that is capable of competing across different functional market segments and across geographically diverse regions. At any rate, problems associated with integration are difﬁcult to predict. The merger of Chrysler and Daimler-Benz is an example of severe difﬁculties stemming from differences in corporate culture and international differences in management style. The integration difﬁculties that confronted Piper Jaffray and U.S. Bancorp have their roots in the vastly different domestic managerial styles. The difﬁculties that the Citigroup has faced may be those pertaining to integrating both technology and management. These considerations have strategic implications for the types of combinations, that is, strategic alliances versus merger or acquisition, which ﬁrms may undertake. The latter is particularly true for cross-border combinations. If the impetus for expansion continues, strategic alliances may become the prelude to more formal engagements between ﬁnancial service ﬁrms.

REFERENCES Abdelal, R., & Bruner, C. M. (2005). European Financial Integration. Harvard Business School Note. Amel, D., Barnes, C., Panetta, F., & Salleo, C. (2004). Consolidation and efﬁciency in the ﬁnancial sector: A review of the international evidence. Journal of Banking and Finance, 28, 2492–2519. Argures, N. (1996). Capabilities, technological diversiﬁcation, and divisionalization. Strategic Management Journal, 17, 395–410. Berger, A. N., Demstz, R. S., & Strahan, P. E. (1999). The consolidation of ﬁnancial services industry: Causes, consequences and implications for the future. Journal of Banking and Finance, 23, 135–194.

Mergers and Consolidation of Financial Service Firms

417

Berger, A. N., & Humphrey, D. B. (1992). Megamergers in banking and the use of cost efﬁciency as an antitrust defence. Antitrust Bulletin, 37, 541–600. Berger, A. N., & Mester, L. J. (1997). Inside the black box: What explains differences in the efﬁciencies of ﬁnancial institutions? Journal of Banking and Finance, 21, 895–947. Brush, T. H. (1996). Predicted change in operational synergy and post-acquisition performance of acquired businesses. Strategic Management Journal, 17, 1–24. Buch, C. M., & DeLong, G. (2004). Cross-border bank mergers: What lures the rare animal? Journal of Banking and Finance, 28, 2077–2102. Cumming, C. (1987). The economics of securitization, Federal Reserve Bank of New York, Quarterly Review, Autumn. Cummins, J. D., Tennyson, S. L., & Weiss, M. A. (1999). Consolidation and efﬁciency in the US life insurance industry. Journal of Banking and Finance, 23, 325–357. DeLong, G. (2001). Stockholder gains from focusing versus diversifying bank mergers. Journal of Financial Economics, 59, 221–252. Focarelli, D., & Pozzolo, A. F. (2001). The patterns of cross-border bank mergers and shareholdings in OECD countries. Journal of Banking and Finance, 25, 2305–2337. Hughes, J. P., Mester, L. J., & Moon, C. G. (2001). Are scales economies in banking elusive or illusive? Incorporating capital structure and risk-taking into models of bank production. Journal of Banking and Finance, 25, 2169–2208. Peristiani, S. (1993). The effects of mergers on bank performance. Federal Reserve Bank of New York, Studies on Excess Capacity in the Financial Sector, March. Pilloff, S. J., & Santomero, A. M. (1998). The value effects of bank mergers and acquisitions. In: Y. Amihud & G. Miller (Eds), Bank mergers and acquisitions (pp. 59–78). Boston, MA: Kluwer Academic Publishers. Seth, A. (1990). Sources of value creation in acquisitions: An empirical investigation. Strategic Management Journal, 11, 431–436. Shelton, L. M. (1998). Strategic ﬁts and corporate acquisition: Empirical evidence. Strategic Management Journal, 9, 278–287. Singh, H., & Montgomery, C. A. (1987). Corporate acquisitions and economic performance. Strategic Management Journal, 8, 377–386. Tully, S. (1999). First Union buys retail – and pays the price. Fortune, 139, 43.

This page intentionally left blank

CROSS-BORDER INVESTMENT IN THE LATIN AMERICAN BANKING SECTOR Jesu´s Arteaga-Ortiz, Harvey Arbela´ez and Wendy M. Jeffus ABSTRACT Cross-border investment has been a large part of merger and acquisition activity in the Latin American banking sector. Spain and the United States have been the largest investors, participating in almost 70% of the total transaction value. After an explanation of the importance of foreign direct investment and implications for cross-border investment, this paper focuses on the largest investor in the region’s banking sector and attempts to find an explanation for the increasing participation of Spanish banks. The paper alludes to a potential new reality: Latin America could be the geographical location where major contenders in banking will be engaged in battles for global dominance.

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 419–438 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07017-8 This paper was presented at the two International conferences: AIB 2005 and BALAS 2005.

419

420

JESU´S ARTEAGA-ORTIZ ET AL.

INTRODUCTION Foreign Direct Investment Foreign direct investment (FDI) represents a large part of the increasing and all-encompassing trend toward globalization. In this respect, it is estimated that the weight of FDI in the composition of capital at a world level tripled in the 1990s to reach an average level of above 15% (United Nations Conference on Trade and Development, 2001). FDI is considered to be the most stable form of international capital ﬂows; it is also credited with stimulating growth by providing access to technology as well as foreign markets (Borensztein, de Gregorio, & Lee, 1998). But as Hausmann and Ferna´ndez-Arias (2000) point out, a high share of FDI in total capital inﬂows may be a sign of a host country’s weakness rather than its strength. According to transaction cost theory, a company will opt for the most efﬁcient form of internationalization; the one that minimizes these costs (Kay, 1997; Williamson, 1985). While transaction cost theory is one justiﬁcation for FDI, the analysis of a company’s internationalization strategy must center on the development of its capabilities rather than on the level of transaction costs and the efﬁciency of the transaction (Madhock, 1998). Thus, when a company makes an investment abroad, it is internalizing its advantages or capabilities within its own limitations. In this respect, it can be argued that the internalization theory is the transaction costs theory of the multinational company (Dura´n, 2002). The set of resources possessed by a company and the analysis of supply and demand will shape the most convenient direction for expansion (Caves, 1982). Thus, a fall in the proﬁt rate or the absence of domestic economic growth may facilitate the search for foreign markets through exports, strategic alliances, or via direct investment. Furthermore, from the supply perspective, foreign entry abroad can be considered a way to obtain natural resources or to generate strategic assets, which is linked to a ﬁrm’s competitive advantage. Therefore, the basic decisions that lead to FDI are related to the acquisition of natural resources or to obtaining strategic resources or assets (Dura´n, 2002). The company that initiates FDI mobilizes speciﬁc strategic assets, both material and non-material, that are combined with advantages speciﬁc to the target country or geographical area. Moreover, it is the combination of capabilities and economic capital that enables products (or services) to be made and sold more efﬁciently (higher quality at a lower cost) than those of

Cross-Border Investment in the Latin American Banking Sector

421

rival companies. The next section of this paper focuses on the consequences of FDI in the banking sector. Banking Banks play a critical role in any economy inﬂuencing both consumers and producers such as shareholders, bank employees, customers, and business owners. This inﬂuence comes from the major functions of the banking sector, including transmitting payments, monitoring the business sector, facilitating the creation of money through the ‘‘multiplier effect,’’ and driving overall economic growth. The consequence of banking sector activities can be both positive and negative. Banks facilitate the transmission of payments through clearinghouse activities. As a clearinghouse, banks take on the responsibility of collecting payments from the parties involved. Banks are responsible for lending, payments and settlements of goods and services, securities transfers, foreign exchange, and other aspects of international capital ﬂows. The banking sector has access to the capital, the technological capabilities, and the international network to facilitate these activities. Banks monitor the business sector through the evaluation, pricing, and credit-granting functions (Verbrugge, Owens, & Megginson, 1999). The monitoring function of banks inﬂuences the efﬁciency of an economy. As banks seek to collect debt payments, they work to ensure that the ﬁrms to which they lend function in ways that will permit them to repay their debt obligations. As depository institutions, banks collect savings and simultaneously loan money and extend credit. These functions create a ‘‘multiplier effect’’ in the economy as banks lend money collected as deposits to institutions. Levine and Zervos (1998) ﬁnd that banking development is positively and robustly correlated with contemporaneous and future rates of economic growth, capital accumulation, and productivity growth. Banks can also inﬂuence the sectors of the economy that enjoy access to capital. The lending rates that the banks charge affect the ability of homeowners, small-business owners, and farmers to access capital for investment. The threat of failure and subsequent consequences are an essential part of the ﬁnancial sector’s role in any economy. For example, the failure of one bank may lead to a general crisis of conﬁdence in the reliability and integrity of the banking system. Additionally, the failure of an individual bank introduces the possibility of systemwide failures or systematic risk. The entire system is critically dependent on efﬁciency, transparency, and trust. A paper by Clarke and Cull (1999) gives three political reasons why private

422

JESU´S ARTEAGA-ORTIZ ET AL.

ownership might be more efﬁcient than public ownership: (1) Governments may expropriate investment from public enterprises. (2) They might impose objectives that have a stronger political rather than economic basis. (3) Pressure from interest groups might affect decisions. Cross-Border Implications in Banking Cross-border transactions, privatization, and mass consolidation are major trends in the ﬁnancial sector (Mody & Negishi, 2001). Just 20 years ago, Tschoegl wrote, ‘‘the current degree of integration across banking sectors around the world can only be compared to that existing at the eve of World War I.’’ (Clarke et al., 2001) Cross-border investment in the banking sector is unique because even after they have been purchased, banks are subject to a wide range of government controls (Prager, 1997). Some of the regulations include interest rates charged and received, margin requirements, portfolio allocation, and acceptable risk. The basis for such extensive regulation is the critical role the ﬁnancial sector plays in the domestic economy as well as its role in international economic relations. Cross-border transactions became a signiﬁcant part of all transactions in the banking sector since 1995. Capital inﬂow from foreign investment increases the host country’s balance of payments and can provide economic beneﬁts. The emphasis for privatization has stemmed from the belief that privatized ﬁrms perform better than stateowned enterprises. In fact, foreign-owned banks in developing countries have tended to have better-quality portfolios, higher net worth, and higher ratios of operating income to cost than domestic banks (Clarke, Cull, D’Amato, & Molinari, 2000). The ‘‘Washington Consensus’’ on banking transition called for the separation of commercial banks from the central bank, the abolition of restrictions on internal convertibility of money, the liberalization of interest rates, the restructuring and privatization of state banks, and the entry of new private banks regulated by minimum capital and licensing requirements (Fries & Taci, 2001). This program was driven by the problems with government ownership in the banking sector. For example, while non-bank governmentowned enterprises operate in a narrow industry, banks operate across the whole economy, providing politicians with the broader ability to transfer funds (Dinc- , 2005). Other economic costs of state-owned banks are the deterrence of prime-rate foreign lending and subsequent worsening of the prospects for competitive market development (Sherif, Borish, & Gross, 2003). Additionally, Kuczyinski (1993) points out that the funds in Latin America’s banks will eventually tilt toward investments in equities, rather than in debt,

Cross-Border Investment in the Latin American Banking Sector

423

for two reasons: (1) narrower interest margins between debt and international bonds, and (2) less ﬁscal pressure in Latin American countries. FDI in the Latin American Banking Sector The growth opportunity in Latin America is a result of a few interrelated factors such as the market size, proﬁt margins, and capital inﬂow. Latin America is much bigger than Europe, and the ratio of M3 to gross domestic product (GDP) is only 28% in Latin America, while it is 77% in the Euro area and 71% in the United States (Sebastian & Hernansanz, 2000). Proﬁt margins are higher in Latin America (net interest income over total assets). During 1988–1995, the average was 5.76% and only 2.80% in Organisation for Economic Cooperation and Development (OECD) countries (Classens, Demirguc-Kunt, & Huizinga, 1998). The ratio of operating costs to assets in Latin America was 5.5% on average between 1992 and 1997 compared with 1.7% in G3 countries, i.e., United States, Germany, and Japan (Hawkins & Mihaljek, 2001). In addition to growth, Latin America may prove to be a strategic location for ‘‘cross-subsidization’’ (Hamel & Prahalad, 1985). Consider the case of Citi Group (i.e., Citi Bank) and its major purchase of Bancomex in Mexico; once Banco Santander of Spain arrived to the summit of the global market, its owner, Emilio Botı´ n, has continued to expand its business presence in the region. This indicates that Latin American ﬁrms have been increasing their presence as a mirror of what has been occurring in the banking sector (for further evidence, see Milman, D’Mello, Aybar, & Arbela´ez, 2001). According to the International Monetary Fund (2003) International Financial Statistics and Balance of Payments Databases, and the World Bank Global Development Finance, the Latin American region received over 603 billion dollars of FDI inﬂows between 1985 and 2002, representing 8% of total FDI. During the same period, over 50 billion dollars of FDI inﬂows into Latin America were allocated to the banking sector, and Spain represented over 39% of FDI inﬂows into Latin America. The following 15 countries had merger and acquisition (M&A) activity within the time period we analyzed: Argentina, Bolivia, Brazil, Chile, Colombia, Dominican Republic, Ecuador, El Salvador, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay, and Venezuela. There were no reported transactions in the banking sector for Costa Rica, Cuba, Guatemala, or Honduras. In addition, the analysis of Spanish foreign investment and foreign investment in to Spain was performed using data for the period from 1933 to 2001, inclusive, that was provided by the Department of Trade at the Spanish Ministry of Finance.

424

JESU´S ARTEAGA-ORTIZ ET AL.

METHODOLOGY Resources and Statistical Analysis Data for this study were compiled from the Securities Data Company’s (SDC) worldwide M&A database for the period 1985 to 2002. SDC Platinum by Thomson Financial provides access to a collection of three ﬁnancial databases. The data cover initial public offerings, corporate mergers, and venture capital and acquisitions. This study examines 230 completed crossborder transactions in 15 Latin American countries that took place between January 1985 and December 2002. Descriptive statistics were used to analyze the data. The evaluation compiled an overall analysis of the data set, and as well as the identiﬁcation of the most active target nations and the most active acquiring countries. Only the four nations with a higher value of transactions (93%) were identiﬁed and analyzed separately. Finally, we emphasized the role of the most active investor in the region.

DISCUSSION OF RESULTS The ﬂow of capital to Latin America is becoming increasingly dominated by FDI (Hausmann & Ferna´ndez-Arias, 2000). Speciﬁcally, FDI through privatizations has been a means to reform the banking systems in Latin America. In Argentina, Mexico, and Venezuela the World Bank has helped these governments privatize in an attempt to restore conﬁdence in the ﬁnancial systems (Hausmann & Ferna´ndez-Arias, 2000). Table 1 shows the signiﬁcance of M&A activity in the banking sector in Latin America. For example, there were 696 transactions between 1985 and 2002, and almost half of these transactions were cross-border deals. Privatizations also represented a signiﬁcant percentage of the total transactions. While the value of every transaction was not available, the SDC database reports over 50 billion dollars in banking activity within the time period analyzed. The main determinants of the expansion of foreign banks into Latin America include diversiﬁcation, deregulation, potential gains from efﬁciency, and high intermediation margins. As the banking sector becomes increasingly competitive, foreign banks seek international opportunities to diversify earnings. Deregulation has also played a critical role in crossborder banking acquisitions. Through privatization programs and other

Cross-Border Investment in the Latin American Banking Sector

Table 1. Target Nation

425

Banking M&A Activity in Latin America.

Total M&A Activity

Privatizations

Cross-Border Transactions

Total Value of FDI

Argentina Bolivia Brazil Chile Colombia Dominican Republic Ecuador El Salvador Mexico Nicaragua Panama Paraguay Peru Uruguay Venezuela

132 4 149 51 53 5

6 0 16 2 7 0

70 4 64 30 22 4

5859.13 202.92 15551.87 5180.22 1075.75 n.a.

29 13 94 5 13 7 56 19 66

0 1 25 1 0 0 8 2 13

5 4 37 1 6 5 39 11 19

7.2 28.99 19878.76 11.05 66.65 10.8 767.9 110.5 1559.16

Total

696

81

321

50310.89

M&A, Merger and acquisition; FDI, foreign direct investment; n.a., not applicable. Source: SDC Worldwide Database (value in millions of U.S. dollars).

Table 2. Most Active Target Nations.

Argentina Brazil Peru Mexico Chile All countries

Transactions

%

Total Value

%

70 64 39 37 30

10 9 6 5 4

5859.13 15551.87 767.90 19878.76 5180.22

12 31 2 40 10

696

100

50310.89

100

Note: Cross-border transaction activity (value in millions of U.S. dollars). Source: SDC Worldwide Database.

reforms, Latin American markets have become more open to foreign ownership in the banking sector. Finally, the average margin on assets as measured by net interest over total assets is considerably higher than in the developed world (Paula, 2002). Table 2 shows the most active target nations in cross-border M&A activity in the banking sector. Ranked by value, they are Mexico, Brazil,

JESU´S ARTEAGA-ORTIZ ET AL.

426

Argentina, Chile, and Peru. Spain has been the most active in Argentina, Brazil, Chile, and Mexico. The next section takes a closer look at the banking sector in these four countries. Argentina Argentina’s banking sector has faced many challenges, including a shortage of credit, the government ownership of banks, and segmentation of private banks. The shortage of credit was the direct result of macroeconomic instability. This macroeconomic instability was a result of increasing price instability and periodic hyperinﬂation. Increased uncertainty in Argentina led to a reduction in both the level and the maturities of ﬁnancial instruments. In fact by 1990, most ﬁnancial instruments had maturities of less than 1 week (Carrizosa, Leipziger, & Shah, 1996). The scarce availability of credit can also be attributed to chronic public sector deﬁcits, which crowded out private sector credit needs. Government-owned banks have traditionally accounted for a large percentage of total loans and a large volume of non-performing loans that undermined conﬁdence and added to the ﬁscal burden of the country. To combat these problems, Argentina went through substantial reforms in the banking sector, including privatizations and foreign ownership of banks. After 1994, foreign banks became increasingly important in Argentina. In 1997, two Spanish banks (Banco Bilbao Vizcaya Argentaria [BBVA] and Banco Santander Central Hispano [BSCH]) bought two major domestic banks (Banco Frances and Banco Rio). Foreign-owned banks in Argentina extended more credit to the manufacturing sector than the agricultural, and lending was highly concentrated in Buenos Aires (Clarke et al., 2000). Brazil The banking sector in Brazil is one of the largest in Latin America and has faced its own set of challenges. While Argentina has been very open to foreign banks, Brazil has been more reluctant in opening its markets to foreign ownership in the banking sector. For example, until 1994, Brazil had a relatively small foreign subsidiary presence. In response to hyperinﬂation and lack of liquidity in the banking sector, the Brazilian government began to allow foreign entry on a case-by-case basis. In 1995, the Brazilian government announced that foreign banks would not be allowed to open new branches or acquire smaller banks unless they purchased one of the many ‘‘troubled banks.’’1

Cross-Border Investment in the Latin American Banking Sector

427

The rising foreign bank presence in Brazil has increased political discussions about the optimal level of bank penetration (Peek & Rosengren, 2000). The sale to a Spanish bank in 2000 of Banespa (Banco do Estado de San Paulo), a large government-owned bank with an extensive branch network, received a lot of criticism. Recently, in clear contrast to the rest of the region, the consolidation of banking activity in Brazil has been led by local banks, thereby reversing the trend observed since 1997. Two locally owned banks –Bradesco and Itau´– have expanded signiﬁcantly in the local market over the last 2 years. In 2003, Bradesco acquired two foreign operators: the local subsidiary of the Spanish bank, BBVA, for US$ 700 million, and the asset management subsidiary of JPMorgan for US$ 1.8 billion. Foreign banks, in contrast, have lost market share to local banks by divesting assets (Economic Commission for Latin America and the Caribbean [ECLAC], 2003). Chile Chile has experienced fewer privatizations than Argentina and Brazil (see Table 1). Until 1973, most banks in Chile had been government owned. Since that time Chile’s ﬁnancial development has been signiﬁcant. Chile has built a strong ﬁnancial sector that allowed the country to avoid the ﬁnancial turmoil observed during 1995 and 1997–1998 in other emerging market economies (Barandiaran & Hernandez, 1999). The United States (25%) and Spain (24%) have been the predominant investor countries in terms of total capital ﬂows in 1996–2003 (ECLAC, 2003). According to the ECLAC (2003), Chile has the most stable economy in Latin America and is the region’s best rated country in terms of corruption, competitiveness, general business climate, and risk rating. The banking sector in Chile is one of the economy’s most signiﬁcant sectors. Spanish banks have become signiﬁcant investors in this sector. Through cross-border investment, Spain’s BSCH controls two of the largest Chilean banks, Banco Santiago and Santander Chile. Mexico On September 1, 1982, most of Mexico’s banks were nationalized, and the constitution was amended to prohibit the private ownership of banks. In 1990, former President Carlos Salinas de Gortari submitted a constitutional amendment to re-privatize commercial banks in Mexico (Unal & Navarro, 1999). The privatization process included 18 banks, and the Mexican

JESU´S ARTEAGA-ORTIZ ET AL.

428

government went through great lengths to ensure the transparency throughout the process, resulting in a 45% premium paid to the government (Unal & Navarro, 1999). Mexico’s experience with bank privatization is considered to be very successful. The reasons that are given for Mexico’s success are clear objectives, credible procedures, and transparency (Barnes, 1992). It was not until 1995 that foreign banks were allowed to hold a controlling stake in Mexico, and the largest banks had a limit of 20% on foreign ownership until 1999 when ownership limitations were eliminated. Today, many of the largest banks are under foreign ownership. The Spanish bank BBVA has acquired Grupo Financiero Bancomer. Another Spanish bank, Banco Santander, outbid HSBC for Grupo Financiero. Table 3 shows the most active acquiring nations. Spain is the most active acquiring nation with 9% of total transaction activity and 39% of the total transaction value. The next section discusses Spanish investment abroad, gives a detailed overview of Spanish activity in the region, and attempts to explain the motivations for Spain’s active interest in the Latin American banking sector. After providing the theoretical framework and giving a thorough analysis of the connection between the banking sectors of Spain and Latin America, we offer conclusions and recommendations. Since the Spanish economy opened to other countries more than 40 years ago, Spanish business has been linked to direct investment. The trend has been from unidirectional investment to Spain to bilateral investment. Additionally, the ﬂow of direct investment in the Spanish economy has differed depending on the total amount, the geographical area, and the sector. Total investment has increased from 2,105 million dollars in 1993 to 66,380 million dollars in 2001.2 The average annual growth rate of Spanish foreign investment from 1993 to 2001 was 63.5% (as the investor) and 30% (as the recipient). Spain continues to be a net investor. In absolute terms, during the period 1993–2001, Table 3.

Most Active Acquiring Nations.

Transactions Spain United States Canada France United Kingdom Total

%

Total Value

%

62 29 14 11 9

9 4 2 2 1

19623.36 14949.73 1521.58 216.72 2430.24

39 30 3 1 5

696

100

50310.89

100

Note: Cross-border transaction activity (value in millions of U.S. dollars). Source: SDC Worldwide Database. Spanish Investment Abroad.

Cross-Border Investment in the Latin American Banking Sector

429

Spain invested signiﬁcantly more capital abroad than it received. In addition, when compared with GDP, it becomes evident that the multinationalization rate of the Spanish economy3 is higher than the world rate and notably higher than the average rate of OECD countries (Dura´n, 2002). With regard to the geographical distribution of investment received between 1993 and 2001, FDI was mostly from OECD countries, speciﬁcally the European Union (EU). On the other hand, the outgoing FDI is predominantly directed to Latin America, which received an average of almost 50% annually, with some years exceeding 55%. The sectors in which Spanish ﬁrms invested abroad, in terms of gross investment, and in line with the 16 sectors established in the database of the Directorate General of Investments of the relevant Spanish ministry were corporations and security holdings (42.75%); financial intermediation, banking, and insurance (18.26%); real estate activities and services (5%); and transport and communications (3%). The financial intermediation, banking, and insurance sectors, received 468 million dollars in 1993 (22.25% of total investments) as opposed to 1,266 million dollars in 1997, an increase of 173.68%.4 The ﬁgure reached 13,217 million euros in 2001, which represents a 188.47% increase on 1997.5 The next section discusses the justiﬁcation for Spanish investment in the Latin American banking sector. Spanish Investment in Latin America’s Banking Sector Spanish banks have been the main investors in Latin America. This is, in part, a result of the process of privatization that drove major Spanish service ﬁrms to expand abroad. Not only did the banking sector look to crossborder investment but also did transport and telecommunication and energy companies. Spanish banks were already ‘‘mature’’ when they decided to expand overseas as a result of a two-phase process. The ﬁrst phase was alliance driven and the second phase was a fast-paced expansion into the main Latin American markets. The largest Spanish banks are Banco Bibao Vizcaya (BBV), Banco Santander, and Banco Central Hispanico. BSCH is the leader among foreign banks in Argentina, Brazil, and Chile, while BBVA is the leader in Colombia and Venezuela and the second major bank in Mexico. BSCH and BBVA, as of September 2000, had more than $170B in assets in Latin America and around 55.8% of the total assets of the top six banks in the region (De Paula, 2002) Table 4. Two theoretical frameworks can be used to justify Spanish investment in Latin America. First, Berger, DeYoung, Genay, and Udell (2000) ﬁnd that similar culture and language might offer advantages to some foreign

JESU´S ARTEAGA-ORTIZ ET AL.

430

Table 4.

Biggest Foreign Banks in Latin America by Assets (September 2000) – USD million.

Bank

Origin Argentina

Brazil

BSCH Citibank BBVA BankBos HSBC ABN-

Spain USA Spain USA UK Neth.

28,682 8,798 5,004 9,315 9,126 15,581

Total

26,130 10,429 9,174 11,350 5,016 2,801 64,900

Mexico

Chile

Colombia Venezuela

Total

% 33.99 21.81 19.60 8.71 9.15 6.75

20,100 30,200 42,590 6,350 37,300 4,900 358 6,800 15,202 – 154 2,900

1,376 1,137 2,811 108 – 110

2,556 686 3,700 – – 95

109,044 69,990 62,889 27,931 29,344 21,641

76,5066 115,704 51,150

5,542

7,037

320,839 100

Source: De Paula (2002).

institutions. Likewise, Davidson (1980) ﬁnds that similarity, in terms of characteristics of culture, is a factor that inﬂuences the location decision. Characteristics of culture include language and the competitive environment. Another point that Davidson (1980) makes is that prior experiences with a country seem to facilitate continued strategic investment, which may explain subsequent investments in the banking sector. Prior experiences are important, because as the ﬁrm becomes established in the foreign market, the uncertainty of the new market decreases; and investment decisions are made on more fundamental analysis, such as market potential, production, and transportation costs. A second theoretical framework for Spanish investment in the Latin American banking sector is the idea of psychic distance. O’Grady and Lane (1996) use the following indicators: the level of economic development in the importing countries; the difference in the level of development between the countries; the difference in the level of education in the importing countries; the difference in ‘‘business,’’ cultural, and local languages; the existence of previous trading channels; and other business factors such as industry structure, competitive environment, and cultural difference. While not all of these variables are applicable, Spain can be considered ‘‘close’’ to Latin America in terms of psychic distance. Nevertheless, the geographical distribution of the Spanish investment abroad is quite different from the market destinations of the Spanish exports, where approximately 75% goes to EU-15 nations (European Union countries, before 2004 enlargement to 10 new nations); while Latin America has received an average of 5–7% over the past 10 years. Additionally, there is a historical justiﬁcation for Spanish investment in the Latin American banking sector. Spain, once the ‘‘mother country’’ of

Cross-Border Investment in the Latin American Banking Sector

431

much of Latin America, joined the then three European Communities in 1986, a single market was created in the EU, and Spanish foreign investment was liberalized by Royal Decree 372/92 after European legislation resulted in a directive.6 These events brought about numerous changes in the Spanish ﬁnancial sector. Although the Treaty of Accession gave special treatment (transition periods of up to 7 years for liberalization) to companies in sectors with a high level of state intervention or oligopolistic sectors such as banking, oil, gas, electricity, and communications, in reality, the intention of the EU was made quite clear in the mid-1980s. The intention was to liberalize the sectors considered strategic to Member States such as banking, which had a high level of public intervention (Pe´rez, 1997). Another detriment to the Spanish banking sector was the conversion to the euro; by replacing 12 currencies banks lost signiﬁcant revenue from currency conversion fees. With the new EU, the competition in each national market, and therefore that of Spain, became more intense and proﬁt margins fell considerably, especially in the banking and petroleum sectors (Guille´n, 2004). Thus, Spanish ﬁrms, conscious of being smaller in size than their European rivals, embarked on a search for new business opportunities and the possibility of increased proﬁts. Furthermore, and in line with the product life-cycle theory (Vernon, 1979), Spanish directors realized the need to neutralize the threat from their European competitors in other parts of the world. The opportunity to do so coincided with various Latin American countries launching privatization and liberalization programs that permitted foreign investors to take control of important assets and ﬁrms. As Guille´n and Tschoegl (2000) estimated, Spanish banks, led by the Santander Central Hispano7 (SCH) and the Banco Bilbao Vizcaya Argentaria8 were the ﬁrst acquirers in the region. According to our analysis, Spanish banks acquired up to 62 ﬁnancial institutions in 12 Latin American countries (see Table 5). From Table 5, it can be deduced that the Spanish banks expanded horizontally abroad, as opposed to other Spanish companies that opted for vertical expansion over the same period. Spanish banks were active acquirers in several countries, more than half of which were in Latin American countries. The BBVA and SCH became the largest banks in Central and South America and, in turn, two of the largest in Europe. Their horizontal expansion can be explained according to the oligopolistic and product lifecycle theories in the context of an increasingly saturated and liberalized home market (Broughton & Ripert, 1997; Maudos, Pastor, & Quesada, 1997). Thus, the Spanish banks, who frequently imitate one another, also

JESU´S ARTEAGA-ORTIZ ET AL.

432

Table 5. Target Country Argentina Bolivia Brazil Chile Colombia El Salvador Mexico Panama Paraguay Peru Uruguay Venezuela

Spanish Cross-Border Acquisitions. Transactions

Total Value

14 1 8 9 7 2 11 1 1 3 1 4

2791.75 179.99 7795.21 2348.12 878.02 n.a. 4592.1 60 n.a. 166.69 55.5 755.98

Note: Cross-border transaction activity (value in millions of U.S. dollars). n.a., not applicable. Source: SDC Worldwide Database.

competed between them, forcing acquisition costs up. Their Latin American subsidiaries marketed products more typical of the Spanish market, such as certain types of mortgage and bank accounts. This led to the accumulation of signiﬁcant proﬁts, which enabled them to recover the cost of their acquisitions, even taking into account the above-mentioned increased costs. In fact, by June 2001, the percentage of total group proﬁts obtained in Latin America was 52% in the case of SCH and 31% in that of BBVA.9 Consequently, and in line with the postulates of the Scandinavian school (e.g., Johanson & Vahlne, 1977; Johanson & Wiedersheim-Paul, 1975), their greater size and international stature permitted the large Spanish banks to compete in the European market. Hence, according to Guille´n (2004), the SCH exchanged equity with the Royal Bank of Scotland, Commerzbank, and Grupo Champalimaud. Moreover, it purchased shares in the Societe´ Generale, San Paolo IMI, Banque Come´rciale du Maroc, and the more recent acquisition of the British Abbey National Bank;10 while the BBVA acquired stock in the Cre´dit Lyonnais, Banca Nacional del Lavoro, and Banco Internacional de Andorra. Moreover, this growth permitted ﬁnancial ﬁrms to explore and think about other new target markets a few years ago. In that sense, we have to consider the 2004 acquisitions of two North American banks by BBVA: Valley Bank (California), and Laredo National Bank (Texas). However, we understand that membership in the EU is not the only factor determining the multinationalization of Spanish ﬁrms, more speciﬁcally

Cross-Border Investment in the Latin American Banking Sector

433

those in the banking sector. There is also a series of favorable institutional factors, such as a context favorable to FDI (favorable legislation, privatizations, promotion and reciprocal investment protection agreements, double rate agreements, incentives, regional integration, etc.) coinciding with the technological process and the ﬁrms’ competitive advantages (Dura´n, 2002). All of that helps explain Spanish direct investment in Latin American countries, while there is also the strategic option that the geographical area represents for Spanish ﬁrms; an option that becomes all the more evident when the cultural, political, and historical variables are considered (Dura´n, 1999). Finally, Aguilera, Dencker, and Escandell (2004) point out that ‘‘countrylevel institutional factors, policy and regulatory factors, cultural distance, industry-level institutional factors, industry relatedness, industry diversiﬁcation, ﬁrm-level institutional factors, organizational size, M&A experience’’ continue to be determinants of FDI decisions. In effect, the authors indicate, for example, that ‘‘if a Spanish ﬁrm has the choice to invest in Latin America –a main location of Spanish foreign investment – the likelihood that an announced deal will complete in Argentina is 24% higher than in Brazil, two of the most attractive Latin American markets.’’

CONCLUSION Our research sets forth a potential new reality: Latin America could be the geographical place where major contenders in the banking industry worldwide will be engaged in battles for global dominance. For example, consider the case of Citi Group (i.e., Citi Bank) and its major purchase of Bancomex in Mexico; then, once Banco Santander of Spain arrived to the summit of the global market, its owner, Emilio Botı´ n, has expanded its business presence in the region. Based on this, it is possible to argue along the lines of Hamal and Prahalad (1985) that the battle will occur in a relatively ‘‘neutral’’ territory by exercising what these authors call ‘‘cross-subsidization.’’ The Latin American banking sector has been a target for FDI, with 8% of the worldwide FDI between 1985 and 2002 directed to the Latin America region and 9% of this quantity allocated in the banking sector. The most active acquiring nations in the Latin American banking sector are Spain (39%) and the United States (30%). This paper focused on Spanish investment. The most active target nations are Argentina (10%), Brazil (9%), Peru (6%), Mexico (5%), and Chile (4%). By total value, four countries represent more than 90% of total investment: Mexico (40%), Brazil (31%), Argentina (12%), and Chile (10%). Foreign cross-border investment in

434

JESU´S ARTEAGA-ORTIZ ET AL.

Latin America peaked in 1999 and recently has remained somewhat steady. Spanish investment in the region started in 1992 and continues to remain strong with over US$1 billion invested in 2002. Although it is true that all developed countries can become important foreign investors at a particular moment (Dunning & Narula, 1994), in the case of Spain, the increase in the volume of FDI has been spectacular. The multinationalization of the Spanish economy has been slightly greater than the average of the OECD countries during the 1990s. The justiﬁcation of the geographical orientation of Spanish foreign investment can be supported by at least two theoretic postulates. The ﬁrst of these is ‘‘the psychic distance paradox’’ in line with the proposals of O’Grady and Lane (1996), and the second is the ‘‘country characteristics and experiences’’ proposed by Davidson (1980). Additionally, a great part of the marked growth in Spanish investment is due to Spain’s membership in the EU, which involves the acceptance of l’acquis communitaire, or community patrimony; which, in turn, entails the recognition of European legislation and adapting national laws to that legislation, including the foreign exchange regulations that also cover foreign transactions. However, there are also other determining factors to be considered, such as favorable institutional factors. These may include the fact that a context favorable to FDI coincides with the technological process and the generation and accumulation of competitive advantages by companies, as well as cultural, political, and historical variables. The Spanish banks expanded horizontally in Latin America, which strengthened their internal, national, and European positions. Spanish outgoing FDI during the period analyzed is predominant in Latin America (approximately 50%), but it is not connected to Spain’s export geographical scope. The disparity could be due to the existence of a single market inside the EU that facilitates the export of goods and, therefore, could have great inﬂuence in that point. In spite of that, Spanish ﬁrms show that Latin American markets are more proﬁtable than the EU market (Arteaga, 2003). Nonetheless, the ﬁndings of this study need to be interpreted with caution within its appropriate context to suggest both limitations and new directions for further research. The database used in this study provides comprehensive information concerning initial domestic and international issues of debt and equity securities. The limitations to this database include missing data such as transaction values and percentage acquired, which lead to some comparability problems. Nevertheless, this paper attempts to give an overview of the M&A activity in the Latin American region; and for these purposes, we found the data provided by SDC to be sufﬁcient. It can be

Cross-Border Investment in the Latin American Banking Sector

435

safely stated that this study has positively contributed to an important topic of research at different levels. For future research, it could be suggested to also take into account the ﬁrst ﬁrms involved in that activity within the region and analyze its evolution worldwide.

NOTES 1. Legislative Intent (‘Exposicao de Motivos’) no. 311, August 8, 1995. 2. A total of 1,893 million euros were converted at the 12/31/93 spot rate of 0.898876EUR/USD and 74,577 million euros were converted at the 12/31/01 spot rate of 1.12347EUR/USD. 3. Measured by the country’s total incoming and outgoing foreign investment in relation to its GDP. 4. A total of 421 million euros were converted at the 12/31/93 spot rate of 0.898876EUR/USD and 1,153 million euros were converted at the 12/31/97 spot rate of 0.910746EUR/USD. 5. A total of 14,850 million U.S. Dollars were converted at the 12/31/01 spot rate of 1.12347EUR/USD. 6. 88/361 of the European Council (EEC) reﬂected at the Spanish Royal Decree 1816/91. 7. Initially, the Banco Santander, which later merged with the Central Hispano – itself the result of a previous merger between the Banco Central and the Banco Hispanoamericano. 8. The result of the original merger of the Banco Bilbao and the Banco Vizcaya and the later acquisition of Argentaria, a public company until its privatization almost 10 years previously. 9. Wall Street Journal Europe (18/6/01, p. 11), in Guille´n (2004). 10. This was the ﬁrst cross-border banking operation in the EU, with the Santander, with assets worth more than 52,000 million euros, becoming the world’s eighth largest bank in terms of market capitalization (62 U.S.$ billion) and the ﬁrst continental European one (The New York Times, July, 26, 2004, p. 10).

REFERENCES Aguilera, R. V., Dencker, J. C., & Escandell, X. (2004). Left at the altar: An institutional analysis of global merger and acquisition announcements in the 1990s. Paper presented at the 2nd JIBS/AIB/CIBER International Business Research Conference held in Lansing, MI, September 15–20, 2004. Arteaga, J. (2003). La actividad exportadora y las barreras a la exportacio´n: una aplicacio´n empı´ rica a las pequen˜as y medianas empresas espan˜olas. Ph.D. thesis. Universidad de Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain. Barandiaran, E., & Hernandez, L. (1999). Origins and resolution of a banking crisis: Chile 1982–86. Working Paper, page 57, Central Bank of Chile.

436

JESU´S ARTEAGA-ORTIZ ET AL.

Barnes, G. (1992). Lessons from bank privatization in Mexico. Working Paper no. 1027, World Bank, November, 45, June, 115. Berger, A. N., DeYoung, R., Genay, H., & Udell, G. F. (2000). Globalization of ﬁnancial institutions: Evidence from cross-border banking performance. Brookings Papers on Economic Activity, 2, 23–158. Borensztein, E., de Gregorio, J., & Lee, J. W. (1998). How does foreign direct investment affect growth? Journal of International Economics 45, June, 115–135. Broughton, A., & Ripert, A. (1997). European banking: Spain, the Spanish banking system. New York: Morgan Stanley. Carrizosa, M., Leipziger, D., & Shah, H. (1996). The Tequila effect and Argentina’s banking reform. Finance and Development, March, 22. Caves, R. E. (1982). Multinational enterprise and economic analysis. Cambridge, UK: Cambridge University Press. Clarke, G. R. G., & Cull, R. (1999). Bank privatization in Argentina: A model of political constraints and differential outcomes. World Bank Working Paper no. 2633. Clarke, G., Cull, R., D’Amato, L., & Molinari, A. (2000). On the kindness of strangers? The impact of foreign entry on domestic banks in Argentina. In: S. Classens & M. Jansen (Eds), The internationalization of financial services: Issues and lessons for developing countries (pp. 331–354). Boston: Kluwer Academic Press. Clarke, G. R. G., Cull, R., Martinez P., Maria S., Sanchez, S. M., (2001) ‘‘Foreign Bank Entry: Experience, Implications for Developing Countries, and Agenda for Further Research’’ World Bank Policy Research. Working Paper No. 2698. Classens, S., Demirguc-Kunt, A., & Huizinga, H. (1998). How does foreign entry affect the domestic banking market? World Bank Discussion Paper, Washington, DC: World Bank. Davidson, W. (1980). The location of foreign direct investment activity: Country characteristics and experience effects. Journal of International Business Studies, 11-2, 9–22. De Paula, L. F. (2002). The recent wave of European banks in Brazil: Determinants and Impacts. Sao Paulo, Brazil: Banco Santos. Dinc- , S. (2005). Politicians and Banks: Political Inﬂuences on Government-Owned Banks in Emerging Markets. Journal of Financial Economics, 77, 453–479. Dunning, J. H., & Narula, R. (1994). Transpacific foreign direct investment and the investment development path: The record assessed. South Carolina Essays in International Business, no. 10, Columbia, South Carolina. Dura´n, J. J. (1999). Multinacionales espan˜olas en Iberoame´rica, Valor estrate´gico. Madrid: Pira´mide. Dura´n, J. J. (2002). Estrategias de localizacio´n y ventajas competitivas de la empresa multinacional espan˜ola. Informacio´n Comercial Espan˜ola, 799, 41–53. Economic Commission for Latin America and the Caribbean. (2003). Foreign investment in Latin America and the Caribbean – 2003 Report. Santiago, Chile. Fries, S., & Taci, A. (2001). Banking transition: A comparative analysis. In: L. F. Bokros & C. Alexander Votava (Eds), Financial transition in Europe and Central Asia: Challenges of the new decade (pp. 173–187). Washington, DC: World Bank. Guille´n, M., & Tschoegl, A. (2000). The internationalization of retail banking: The case of the Spanish banks in Latin American. Transnational Corporations, 9(3), 63–98. Hamel, G., & Prahalad, C. K. (1985). Do you really have a global strategy? Harvard Business Review, July–August, 138–148.

Cross-Border Investment in the Latin American Banking Sector

437

Hausmann, R., & Ferna´ndez-Arias, E. (2000). Foreign direct investment: Good cholesterol? Inter American Development Bank Working Paper no. 417, April. Hawkins, J., & Mihaljek, D. (2001). The banking industry in the emerging markets economies: Competition, consolidation and systemic stability. BIS Papers no. 4. Basel, Switzerland: Bank of International Settlements. International Monetary Fund. (2003). International financial Statistics and Balance of payments database (5th ed.) (BPM5). Washington. Johanson, J., & Vahlne, J. E. (1977). The internationalization process of the ﬁrm: A model of knowledge development and increasing foreign market commitments. Journal of International Business Studies, 8, 23–32. Johanson, J., & Wiedersheim-Paul, F. (1975). The internationalization of ﬁrm: Four Swedish cases. Journal of Management Studies, October, 305–322. Kay, N. M. (1997). Pattern in corporate evolution. Oxford, UK: Oxford University Press. Kuczyinski, P. P. (1993). International capital ﬂows to Latin America: What is the promise?’’ Proceedings of the World Bank Annual Conference on Development Economics, pp. 323–348. Levine, R., & Zervos, S. (1998). Stock markets, banks, and economic growth. American Economic Review, 88(3), 537–558. Madhock, A. (1998). Cost, value, and foreign market entry mode: The transaction and the ﬁrm. Strategic Management Journal, 18, 39–61. Maudos, J., Pastor, J. M., & Quesada, J. (1997). Technical progress in Spanish banking, 1985–1994. In: J. Ravell (Ed.), The Recent Evolution of Financial System. London: Macmillan. Milman, C. D., D’Mello, J. P., Aybar, B., & Arbela´ez, H. (2001). A note using mergers and acquisitions to gain competitive advantage in the United States in the case of Latin American MNCs. International Review of Financial Analysis, 10, 323–332. Mody, A., & Negishi, S. (2001). Cross-border mergers and acquisitions in East Asia: Trends and Implications. Finance Development, 38(1), 1–8. O’Grady, S., & Lane, H. W. (1996). The psychic distance paradox. Journal of International Business Studies, 27(2), 309–333. Peek, J., & Rosengren, E. S. (2000). ‘‘Implications of the globalization of the banking sector: The Latin American experience.’’ New England Economic Review, Federal Reserve Bank of Boston, September, 45–62. Pe´rez, S. (1997). Banking on privilege: The politics of Spanish financial reform. Ithaca, NY: Cornell University Press. Prager, J. (1997). Banking privatization: How compelling is the case? In: A. Bennett (Ed.), How does privatization work? Essays on privatization in honour of Professor V.V. Ramanadham. Studies in the Modern World Economy (Vol. 9. pp. 33–50). London: Routledge. Sebastian, M., & Hernansanz, C. (2000). The Spanish bank’s strategy in Latin America. SUERF Studies no. 9. Vienna, Austria: Societe Universitaire Europeenne de Recherches Financieres. Sherif, K., Borish, M., & Gross, A. (2003). State-owned banks in the transition: Origins, evolution, and policy responses (pp. vii, 122). Washington, DC: World Bank. Timmons, H. (2004). Santander set to buy Abbey National of U.K. The New York Times, July 26, p. 10. Tschoegl, A. E. (1985). Ideology and changes in regulations: The case of foreign bank branches over the period 1920–1980. In: T. L. Brewer (Ed.), Political risks in international business (pp. 95–114). New York: Praeger.

438

JESU´S ARTEAGA-ORTIZ ET AL.

Unal, H., & Navarro, M. (1999). The technical process of bank privatization in Mexico. Journal of Financial Services Research, 16(1), 61–83. United Nations Conference on Trade and Development. (2001). Estrategias de localizacio´n y ventajas competitivas de la empresa multinacional espan˜ola. Informacio´n Comercial Espan˜ola, 799, 41–53. Verbrugge, J., Owens, W., & Megginson, W. (1999). State ownership and the ﬁnancial performance of privatized banks: An empirical analysis. Conference Proceedings of a Policy Research Workshop Held at the World Bank, March 15–16, Federal Reserve Bank of Dallas. Vernon, R. (1979). The Product Cycle Hypothesis in a New International Environment. Oxford Bulletin of Economics and Statistics, 41, 255–267. Williamson, O. E. (1985). The economic institution of capitalism. New York: Free Press.

PART VII: THE INFLUENCE OF THE STATE

This page intentionally left blank

GOVERNANCE AND POLITICAL RISK Arvind K. Jain ABSTRACT Political risk should be seen as arising from renegotiation of implicit or explicit contract under which foreign investors enter a host country. Governments will legitimately enter into renegotiation to increase the share of rents earned by the society. Corrupt political leaders, however, will use their powers to extract rents from foreign investors for personal gains rather than for the good of the society. Political risk assessment, therefore, should assess the intentions of government as well as the strengths of political and social institutions that keep leaders under control. Firms should also understand that their own actions may contribute to creating political risk.

Two events in Bolivia in the spring of 2006, reminded international investors that ‘‘political’’ or ‘‘country’’ risk had not become a thing of the past. Only one of the two events affected foreign investors in the country directly but together these events provide very useful windows on understanding the sources of what investors worry about under the rubric of political risk. On May 1, 2006, the newly elected president of Bolivia, Evo Morales ordered the military to take control of the country’s natural gas ﬁelds. The military take over was to prevent any manipulations of the companies’ Value Creation in Multinational Enterprise International Finance Review, Volume 7, 441–460 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07018-X

441

442

ARVIND K. JAIN

accounts. He asked the companies, all foreign, that had controlled the gas production to agree to turn over higher proportion of the revenue from the natural resources of the country to the government. The government expected to raise its revenues from $300 million to $780 million. These companies had acquired control of the gas ﬁelds in the 1990s when these ﬁelds were privatized without the approval of the congress. A few weeks later, the president took the ﬁrst step in the government’s plans to redistribute about 20 million hectares of land – almost one ﬁfth of the country’s total area – to landless indigenous communities. Much of the country’s agricultural land is currently owned by families of European descent and mestizo elite. Landowners resisted the move with threats of production stoppages. President Morales’s action stemmed from application of the 1953 law that allowed redistribution of large land holdings. At about the same time, in a development unrelated to Bolivia, Coca-Cola was facing international legal challenges arising from accusations that it had sided with the corrupt president of Uzbekistan to protect its long-term interests at the cost of its minority shareholders.1 Facing potential of erratic actions from a repressive and corrupt regime, the company had taken what it thought were protective steps. In the process, it had created other risks for itself. In relation to the concept of political risk, these events raise some questions that we will attempt to answer in this paper. What motivated the Bolivian Government, and especially President Morales – the ﬁrst ever indigenous leader of Bolivia – to undertake these moves? Did the strategy followed by Coke increase or decrease its risks of operating in a turbulent political environment? Do these events imply that politics in emerging economies tends to be chaotic or erratic? Were the moves in Bolivia merely frustrations of the poor against the rich landowners and/or (rich) foreign companies? Who will lose and who will beneﬁt from these moves? Is the country better off due to these moves? Do these types of actions represent special risks for foreign investors in a country? The aim of this paper is to help the reader answer some of these questions. As international investments become more important for the global economy, investors have to assess risks associated with various types of investments. Assessment of risk serves two purposes. First, it helps the investor identify projects whose risk may be above a threshold level. Second, investors have to ensure that the rewards associated with the project adequately compensate them from risks associated with the project. Political risk is one of these risks that must be understood and assessed. International investors are not the only ones who have to assess the risks associated with foreign investments. Recipients of these investments,

Governance and Political Risk

443

especially the less-industrialized economies, face the potential or least the perception, of being exploited by foreign investors who tend to be better informed about global economic developments and tend to have the backing of powerful industrialized countries’ governments. For the emerging economies, the risk is that the societies’ returns will fall below those that could have been obtained from the alternate use of the countries’ resources. It is in the interest of both the parties to understand and assess all risks accurately. Political risk is one of the risks that have been difﬁcult to assess. While scholars and policymakers have deﬁned and measured this risk for a long time, our knowledge about political risk is still quite rudimentary. Political risk is generally seen as arising from ‘‘political’’ behavior of governments and foreign investors tend to put a heavy price for bearing this kind of risk especially because it is perceived as being erratic (Moran 1998, 2001). In this paper, we view a government’s actions as rational responses to the choices it faces in order to meet its objectives. Perhaps, President Morales was doing what any good leader in his position would do. We explicitly recognize that these objectives may include improvement of the welfare of the society that the government is supposed to manage as well as personal objectives of the decision makers themselves. Political risk for investors, in this context, arises when the fulﬁllment of these objectives becomes detrimental to the private investors. In the ﬁrst section below, we review the past understanding of this concept and propose a somewhat new deﬁnition of the term political risk. The essence of this deﬁnition is that a government will attempt to create conditions that increase the host country’s share of the rents that might be generated by the activities of the foreign investor. The second section discusses some constraints that the governments will face in fulﬁlling their objectives. These constraints will determine the range of options that might be available to a government that wishes to bargain with the foreign investor. Foreign investors tend to view the exercise of these options as political risk. The last section outlines some implications for the foreign investors.

A DEFINITION OF POLITICAL RISK The term political risk has been used in many ways to describe the inﬂuence of political developments within a country on the operations of private, especially foreign, enterprises in the economy. A typical deﬁnition of political risk views it as ‘‘the probability that political forces will negatively affect a multinational enterprise’s (MNE’s) proﬁt or impede the attainment

444

ARVIND K. JAIN

of other critical business objectives’’ (Rugman & Hodgetts, 2000, p. 362).2 Others note that ‘‘managers in MNEs are concerned and frustrated by governments that force unwanted changes in their preferred method of operations’’ (Beamish, Morrison, & Rosenzwieg, 1997, p. 199). A broader deﬁnition of political risk is provided by Wells (1998): ‘‘y risks that are primarily the result of forces external to the industry and which involve some sort of government action, or occasionally, inaction’’ (p. 15). These deﬁnitions and prescriptions for managing political risks that often follow these deﬁnitions,3 tend to view the political developments by host nations as having a nuisance value for foreign (as well as domestic) investors without serving any real purpose other than strengthening or consolidating the powers of the political leaders. There is unwillingness in these views of political risk to distinguish between legitimate actions of governments that follow from their assessment of the best strategies for their economies and actions that target foreign investors due to their visibility as dominant and powerful forces that affect a society. These views of political risk also absolve foreign investors from any responsibilities in creating conditions that may force host governments to take certain actions to protect their societies. Wells (1998), for example, considers a government’s change in its development strategy (which affect all the participants in the economy, not just the foreign ﬁrms) and development of local competition as sources of political risk. Wells also seems to include restrictions on entry to a market as a form of political risk (pp. 18–19). These views of political risks appear to be self-serving in that they make no attempt to understand why host governments undertake the steps that are seen as creating risks for the investors. The unstated assumption behind these views of political risks appears to be that the governments’ actions are essentially erratic and hence irrational.4 These analyses, carried out in a partial equilibrium framework, fail to see any connection between, ﬁrst, the ‘‘political forces’’ and other economic events taking place in the economy, and second, on the actions of the foreign investors themselves that lead to certain responses from the governments. These views are also objectionable in that they consider national goals of a country to be secondary to the commercial and ﬁnancial interests of the foreign enterprises. Attempts to quantify political risk as deﬁned above tend not to be very successful. Henisz (2002) has reviewed a number of attempts to quantify political risk and has identiﬁed both conceptual as well as quantitative problems with these attempts (pp. 8–11). He concludes that ‘‘most of the problems with conventional political risk measures ultimately stem from their lack of focus on the political systems that they purport to measure’’ (p. 9).

Governance and Political Risk

445

In contrast to the views that political risk represents some type of erratic behavior of host governments, it may be fruitful to view the host governments as rational players in their economies and analyze their political actions as attempts to improve the welfare of their societies. These actions may take the form of economic development policies that assign a different role to foreign investors than one that these investors would consider desirable, or by following policies or negotiating contracts that increase the beneﬁts that the host economies derive from the presence of foreign (or domestic) investors. In developing a new model for understanding political risk, we will focus almost exclusively on the second of these actions of host governments, namely their attempts to improve the beneﬁts that host economy players derive from the projects that employ resources of the economy. Beneﬁts accruing to the host economies can be increased by implementing economic policies that change the distribution of rents inherent in the activities of private investors in favor of the society. Such redistribution requires a redeﬁnition of laws and rules under which investors operate within the economy. Governments may, of course, also change the rules and laws of the society that are not directed at the private investors but which affect these investors along with others. Political risk, in this view, may be deﬁned as the risk that the ﬁrm will be subjected to renegotiation of the explicit or implicit contract under which the ﬁrm entered the host economy, when the objective of the renegotiation is a redistribution of the rents inherent in the operations of the ﬁrm such that the ﬁrm’s share of the rents is reduced. Governments, or better the government leaders, however, are not always motivated by honest intentions of improving the welfare of the societies that they control. Political leaders may follow policies that will favor voters who are likely to reelect them. Such redistribution of rents and wealth within a society is consistent with the principles of a democratic political system.5 Corrupt government leaders may also be motivated to change the laws when they themselves, rather than some of the members of the society, may beneﬁt directly from the changes. Academics and practitioners have tended to include such corruption as one of the factors that give rise to political risk (Henisz, 2002; Wells, 1998). Such opportunistic behavior, however, will not go unchecked within the political system. The extent to which the leaders can get away with such behavior will depend upon the strength of political and social institutions in the society that can keep such self-serving behavior in check. Assessment of political risk of a country, therefore, involves assessment of the motives of the government as well as the policy options available to the government. In an ideal political system, the only motive of the government

446

ARVIND K. JAIN

would be the maximization of citizens’ welfare. In a corrupt political system, government leaders will assess the extent to which they can appropriate the public power for their private use. The objectives of the government will be a combination of welfare of the public and maximization of the private wealth of government leaders. The policy options available to the government depend upon the constraints faced by the leaders. In a democratic system, these constraints arise from all the political institutions that operate in a society in order to inﬂuence and control actions of government leaders. Existence and strengths of such political institutions will determine the range of actions government leaders can take in order to fulﬁll their goals. Political risk will be determined by an interaction of the motives of the governments and the options available to the governments. In the following section, we model political risk as a function of economic and political variables within a country and of the characteristics of the market in which an investor may operate.

BARGAINING MODEL OF NEGOTIATIONS To understand the importance of a government’s objectives and policy options, we extend the bilateral monopoly model of negotiations between host governments and foreign investors popularly known as ‘‘obsolescing bargaining’’ model. This model was ﬁrst proposed by Vernon (1971) and developed by others, for example, Fagre and Wells (1982) and Poynter (1982).6 Host governments with a monopoly over access to domestic economy negotiate with a foreign investor over the conditions under which the investor can enter the domestic market. Before the entry, the foreign investor offers a combination of factors of production – technology, markets, management, capital – that the host country desires. The host country offers access to its markets and/or resources that the foreign investor desires. In this context, foreign investors are seen to have a strong bargaining position before the investment is made when the host government needs the factors of production that the foreign investors possess. Before making their resources available to the country, foreign investors may be able to negotiate favorable terms for their entry into the host country. Once, however, the investment has been made, these investors may be considered as captives of the host governments, especially when the capital that has been invested is largely immobile and cost of divestment is high. The obsolescing bargaining model is based on the perceptions and assessments of a society and the government about the costs and the beneﬁts

Governance and Political Risk

447

of foreign investments. These perceptions are compared in Fig. 1. When host countries desire the foreign investors to enter the market, they may follow policies that attract foreign investments and may offer incentives and tax concessions to entice the foreign investors. Over time, as the host country absorbs the technology offered by the foreign investor, or as other offers of investments become available, the perception of the beneﬁts versus the costs of the foreign investment may change and host governments may wish to renegotiate the terms at which the foreign investor operates in the economy. This model predicts increasing political risk for ﬁrms on the premise that host governments’ will attempt to take advantage of the postinvestment captive nature of the foreign investors and will change the terms of operations to extract higher rents from the foreign investors once the investment has been made. This formulation of bargaining powers views foreign investment as a one-shot investment and does not explicitly account for future potentials. Foreign investor will attempt to shift the perception of beneﬁts by offering updated technology and designs in order for the host country to view foreign investment as an attractive proposition even after the initial investment has been made. At the time the initial contract is made, both the parties also offer the other, without explicit promises, potential for sharing of future developments of their respective assets. Reasonable negotiators will take such potential into account while setting the initial conditions for entry and while changing the conditions for foreign investors to continue to operate in a host economy. Traditional formulation of the obsolescing bargaining model also assumes implicitly that governments do not face any constraints on their Value of foreign investment for host economy

Perceptions

Benefits Host country wants to change conditions under which investors operate

Host country wants to attract investors

Costs Time

Fig. 1.

Obsolescence Bargaining Model.

448

ARVIND K. JAIN

behavior. In reality, political systems within a society put constraints on actions that government leaders can take. Stronger the democratic institutions in a country, stronger will be these constraints. With some exaggeration, it can be claimed that past discussions of political risk have tended to ignore implications of these constraints and future developments of respective assets owned by the two sides.7 In the discussion of political risk in this paper, actions of governments are modeled as rational choices for attaining the objective of maximization of societal welfare. First, however, we have to examine the factors that constrain the bargaining behavior of each party – the host government and the foreign investor.

COMPETITIVE STRUCTURE OF THE INDUSTRY Competitive structure of an industry will determine the scope of excess rents that can be negotiated between the two parties. The structure of an industry can range from a monopoly at one end to a perfectly competitive market at the other end. A monopolistic structure allows a ﬁrm to earn the highest level of returns in the industry. A monopolistic ﬁrm, in the absence of government intervention, chooses its price and quantity in order to maximize its proﬁts. If the government wishes to reduce the welfare losses implicit in this structure, it will set conditions (e.g., the price) for the operations of the monopoly. In effect, the government is imposing a contract on the ﬁrm – a contract that redistributes some of the excess proﬁts from the ﬁrm to the society. At the other extreme, a ﬁrm operating in a perfectly competitive industry has no excess rents. As the environment of an industry changes from a monopoly to a more competitive one, excess rents tend to be reduced such that in a perfectly competitive market, there are no excess rents. In a perfectly competitive industry, there is no room for any negotiation between the ﬁrm and the government. If the government attempts to extract any rent from the ﬁrm, the ﬁrm will have no choice but to withdraw from the market. Markets, however, are rarely perfectly competitive. Less than perfectly competitive markets are characterized by excess rents. These excess rents create a tension between the ﬁrm and the government – both of whom would like to capture as much as possible of the excess rents for themselves. In the obsolescing bargaining model much of the dispute between the ﬁrm and the government will be on the estimate of the excess rents that the ﬁrm is able to generate.

Governance and Political Risk

449

Market structures of industries diverge from a perfect competition for a number of reasons. There are economic as well as strategic reasons why ﬁrms will create monopolistic advantages for themselves. Government’s desire to intervene in the industry will depend upon the extent to which the industry’s market structure deviates from a perfectly competitive one. Just because the government may want to intervene does not imply that it has the power to control the behavior of the ﬁrm. The government’s ability to intervene will depend on the tools it has available, as well as its bargaining power vis-a`-vis the ﬁrm. Political risk, therefore, would depend upon the factors that cause industries to deviate from a perfectly competitive market structure and upon the factors that determine the strengths and weaknesses of two protagonists – the ﬁrm and the government.

OBJECTIVES OF THE GOVERNMENT A government’s evaluation of its options will begin with the deﬁnition of its objectives. Ideally, these objectives ought to be the maximization of citizen’s welfare. Since expenditure policy is one of the most important tools for achieving that objective, government would maximize its tax revenues – the proportion of the rents from economic activities that it expropriates for the provision of public goods in the economy – from private investors. While it may be possible for government to extract beneﬁts in other ways, we can assume, without any loss of generality, that all such beneﬁts can be analyzed as a tax. For the remainder of this paper, we will treat the tax rate as the instrument that captures all the beneﬁts that the government is able to extract from the investment. The tax rate, however, affects the level of private investment in the economy. Investors will assess the possibilities of opportunistic behavior – of raising taxes once the investment has been made – by the government and adjust their investment levels accordingly. In reality, government’s motive may be modiﬁed for at least two reasons. First, a democratic government will limit its objective to maximization of the welfare of the segment of the society that is likely to keep it in power. This will have two effects on the selection of tax rates. The government will ﬁrst estimate the likely beneﬁts of the investments for the sector of the population that supports the government in power. It will also estimate the effectiveness of its ability to redistribute income to its target supporters through public policy. Both these estimates will inﬂuence the selection of tax rates. Second, government leaders may give priority to their own welfare especially when they can circumvent society’s control mechanisms that are

450

ARVIND K. JAIN

supposed to constrain the behavior of political leaders. This situation represents a corrupt government that is able to use its political power to obtain personal gains from its position.8 The planning horizon of a corrupt government will be limited to its expectations about how long it will stay in power. Whereas noncorrupt democratic governments will set tax rate to maximize the beneﬁts of investments and its tax-generated-public-goods expenditures over a number of periods with the expectations of being reelected, a corrupt government may have a shorter planning horizon since it may not expect to be reelected.

POLICY OPTIONS FOR THE GOVERNMENT The tax policy of the government and the ability of the foreign investor to resist government’s demand will be constrained by a number of characteristics of the bargaining situation. Competitive Situation of the Industry As discussed above, more competitive the industry, less will be the bargaining room and hence less will be the ability of the government to change the tax rates. Range of bargaining within an industry will inﬂuence the political risk that a ﬁrm within that industry is likely to face. We expect that less competitive the structure of an industry higher will be the level of political risk. Demonstration Effect on Future Commitments Higher tax revenue in the current period resulting from renegotiating the existing contract will impact the level of future investments. A regime that unilaterally changes the terms of its contract with foreign investors, for example, by imposing a special tax on foreign investors, is giving a signal to all future investors in the country that the terms under which they enter the country may be changed unilaterally and unexpectedly by the host government. The potential investors will have to assess whether the government’s actions are a justiﬁed response to changed economic circumstances or do they constitute a breach of contract (even if legal) and hence should be treated as a deterrent to future investments. Governments that wish to attract investment in the future will have to create an environment in which ﬁrms consider government’s promises about future policies to be credible.

Governance and Political Risk

451

To evaluate the government’s assessment of the joint impact of the two opposing effects of taxation, consider a two-period framework in which the government announces its tax rate at the beginning of the second period having observed the level of private investment in the ﬁrst period. Investors then decide the level of investment in the second period that will then determine the tax revenue in the second period (assume a constant rate of proﬁtability). Let the investment in the ﬁrst period be I1. We assume that investments have a life longer than one period. Investments are, therefore, subject to some uncertainty about future tax rates. Based on this level of investment, the government announces the tax rate for the second period to be t2. Based on the announced tax rate, the total investment in the second period (including undepreciated portion of I1) will be I 2 ¼ f ft2 ; EðDt3!a Þg þ fð1 depreciation rateÞ I 1 g

(1)

where E(Dt3-a) is the expected change in future tax rates. Expectations about future tax rate inﬂuence the investment decision since investments have a life longer than one period. EðDt3!a Þ ¼ Sai ðDti Þ

(2)

where i refers to the next n periods and 0oao1 such that expectations of future changes in tax regimes give higher weights to recent changes. Government sets the tax rate to maximize the tax revenue over the period that it expects to stay in power. Two variables can be used to assess the deterrent effect of loss of future investments on the actions of the governments. First, the annual ﬂow of foreign direct investment in the country as a proportion of the pool of accumulated value of the foreign direct investment in the country provides an indication of how long the foreign investors have been investing in the country. The longer that investors have been around, more it is an indication of the faith that foreign investor have in the credibility of government promises. In addition, the longer that foreign investors have been around in a country, the stronger will be the domestic vested interests that would like the foreign investors to continue to operate in the country. Second, volatility of the foreign direct investment ﬂows in the country may be partially motivated by the uncertainty about the future policies of the government. In summary, we expect that a long experience with foreign direct investment in a country is both an indication of the perception of political risk as well as a deterrent of that risk. Thus, political risk will vary inversely with

452

ARVIND K. JAIN

importance of current levels of foreign direct investment to the total value of cumulative foreign direct investment in the country. Inherent Conflict between the Motives Foreign investors and host economies are in a fundamental conﬂict with each other in how they view market imperfections. Foreign direct investment theory is quite clear about the motives and comparative advantage of foreign investors. The raison d’eˆtre for foreign investors is the existence and the internalization of market imperfections. Given that foreign investors have a disadvantage of being foreigners in a host economy, they can undertake foreign investment only when they can create sufﬁcient value by exploiting, in the form of foreign investment rather than in the form of licensing or exporting, some comparative advantage that they have created. If there were no market imperfections, foreign investment would not be the most efﬁcient vehicle for the exploitation of that comparative advantage.9 The interest of the foreign investor is, therefore, maintenance and enhancement of that market imperfection that justiﬁed the foreign investment. Elimination of these imperfections spells the end of the foreign investment. The host country, on the other hand, has a strong interest in eliminating market imperfections that impede the growth of domestic economy. Raison d’eˆtre for government is to compensate for market imperfections. As a corollary, a good government will attempt to reduce these imperfections. The conﬂicting objectives of the government and the foreign investors enhance political risk. The greater the extent of market imperfections in the economy – that is, less industrialized an economy – higher will be the difference between the objectives of the government and the foreign investor and hence higher will be the incentive for the government to force the foreign investor to share in the rents being generated in the economy. This argument may not apply to a corrupt government. A corrupt government will attempt to maintain these imperfections because existence of these imperfections allows corruption rents to be extracted.10 Hence it could be argued that corrupt governments have an incentive to maintain these imperfections. We can estimate the level of imperfections in the economy or in an industry within that country from the stage of the product life cycle of the industry. Industries at the early stages of their product life will be characterized by high degrees of market imperfections and entry barriers and those at the mature stages by fewer imperfections. We expect that political

Governance and Political Risk

453

risk will be higher for industries at the early stages of their product life cycles. Information Asymmetry A corollary to the above argument will be that information asymmetries that contribute to development of market imperfections will further add to the incentives of the government to tax the foreign investors. We can expect that the larger the difference between the levels of development of the host country and the home countries of foreign investors, the larger will be the political risk faced by the foreign investors. Social versus Economic Costing Market imperfections may force a government to use shadow prices and engage in social accounting and costing of projects rather than economic or ﬁnancial accounting and costing that the foreign investors may undertake. The larger the extent of market imperfections, the larger will be the disparity between the values and costs of a project as perceived by the investors versus the governments. The larger the disparity between the perceptions, the higher will be the political risk. Effects of social versus economic accounting are reﬂected in the shadow prices of factors of production. Since such prices cannot be observed, we hypothesize that countries at the lower end of the development spectrum, that is, countries with lower per capital GDPs will exhibit higher political risk than those at the higher end of the development spectrum. Visibility of a Foreign Investor Merely being a foreigner can be a cost. Foreign direct investment theories, starting from Hymer (1976), have recognized that a foreign investor has a cost of operating in a foreign environment – a cost that a domestic investor does not have to incur. In the context of obsolescing bargaining model of political risk, this cost may take the form of resentment against a foreign investor which may (1) cause the costs of a foreign investor being perceived as being higher than they may be, (2) cause the perceived beneﬁts of a foreign investment to be lower than they may be, or (3) allow the government to justify stricter controls (and hence transfer of rents to the domestic economy) on foreign investors than could not have been imposed on a domestic investor undertaking a similar investment.

454

ARVIND K. JAIN

Life Expectancy of Regimes Regimes that do not expect to last for a long time will have incentives to renegotiate contracts with foreign investors. Such regimes would not be concerned about the loss of investment in the future since they expect not to be around to bear the cost. They can effectively transfer the cost to the future regimes while retaining the beneﬁts of spending the tax revenues that may increase their chances of reelection. In fact, the possibility of nonreelection may motivate the regime to squeeze foreign investors both as a populist measure and as a measure to extract extra tax revenues to spend on vote-buying expenditures. We, therefore, expect that regimes that have lost popular support since coming to power, and hence perceive their chances of reelection to be small, will create higher political risk for foreign investors than regimes that are secure and expect to be reelected. Corruption The objective function of corrupt regimes will not be constrained by costs associated with renegotiations. One of these costs is the possibility of nonreelection. A corrupt regime may expect not to remain in power for too long and hence its objective will be to maximize tax revenue in the current period without worrying about the costs of lost investments in the future. We can expect that corrupt regimes will be associated with higher political risks than noncorrupt regimes. Maturity of Political and Social Institutions within the Country The opportunistic behavior of corrupt regimes will be constrained, however, by the strength of political and social institutions within the country. More developed the institutions that can exert a pressure on the government in the decision-making process, less likely it is that government will take decisions that cater to its self-interest rather than those of the public. Various interest groups, where such groups exist, will be able to counter the propensity of a corrupt regime to ignore the future consequences of their actions in addition to exerting pressure to minimize the level of corruption. Triesman (2000) ﬁnds that longer a country has followed a democratic tradition, lower will the level of corruption within the country. Long tradition with democracy allows social and political institutions to develop that can prevent gratuitous behavior of the politicians.

Governance and Political Risk

455

Accordingly, we expect that political risk will decrease with (1) the length of time that a country has been a democratic country and (2) an increase in the degree of social, political, and journalistic freedom in the country.

ERRATIC VERSUS RATIONAL GOVERNMENT POLICIES AND POLITICAL RISK How should ﬁrms assess political risks for their foreign investments? We have argued that political risk arises when governments change the conditions under which foreign investors operate in order to reduce the share of rents from the investments that the foreign investors get to keep. What leads governments to change their implicit contracts with foreign investors? In an ideal system – by that we mean a political system that operates under known and broadly accepted rules – governments will undertake such actions only when they believe that there will be a net beneﬁt to their constituents from such a change. They would have taken into account the beneﬁts from rise in current rate of taxation (remember that we are using the term taxation to capture all beneﬁts that a society derives) and the losses from lower future investments. In such a system, the foreign investors will have the means to inﬂuence the government’s decision through various interest groups that operate in the society. The political risk in such a situation may be undistinguishable from the ordinary business or commercial risk. All political systems, however, do not work according to well-accepted rules. One of the main areas where foreign investors will face a risk is when the decision makers are motivated by their self-interest – or corruption. Corrupt leaders may change the conditions for the operations of foreign (as well as domestic) investments even when such changes do not serve the society’s interest – these leaders are only interested in what is good for them. Such leaders may assess that their regime will have a short life span. Such a perception may motivate them to act quickly – thus increasing the risks for foreign investors even more. Even corrupt regimes, however, will face opposition from domestic interest groups that lose out when corruption increases economic uncertainty and may oppose the corrupt regimes. Foreign investors will ﬁnd that economies in which civic groups are active tend to have lower levels of corruption and corruption tends to be less harmful for economic activities. Societies with well-developed political institutions and civic groups, hence,

456

ARVIND K. JAIN

offer lower levels of political risk. The longer a country has had a democratic tradition higher are the chances that it has a large range of the necessary political and social institutions that keep a check on the behavior of the political leaders. Due to the real or perceived asymmetries between the powers of emerging economies and the foreign investors, governments are more likely to change their ‘‘incentive package’’ for monopolistic industries than for competitive ones. Foreign investors that operate in monopolistic industries will ﬁnd a bigger change in the host society’s perceptions over time than ﬁrms that operate in competitive industries. Foreign investors will ﬁnd that government’s perceptions of beneﬁts of foreign investments change more when there is big gap between the level of development of the host country and the home country of the investor. Foreign investors originating in richer countries will be perceived to have information advantages which will create resentments when investments become successful and begin to either dominate the host economy or begin to expatriate proﬁts back to the home country. Countries that rely heavily on foreign investments are less likely to change conditions under which foreigners operate in the country than countries that rely very little on foreign capital. Presence of other investors and a large number of foreign ﬁrms in the country is a good signal that the country does not change its policies regarding the foreign investors too frequently.

BOLIVIA, UZBEKISTAN AND POLITICAL RISK How can we understand the recent event in Bolivia in view of this discussion? How do these events help us understand the concept of political risk? To understand President Morales’ actions, we must separate the interests of Bolivia from those of the foreign investors. Let us assume that Bolivia will go ahead and nationalize the gas ﬁelds. Who will beneﬁt in Bolivia and who will lose? The ﬁrst reaction of the rest of the world is likely to be to cut off ﬂow of any future technology and investments, at least for a foreseeable future.11 The general opinion is that Bolivia does not have sufﬁcient knowhow to develop and manage the ﬁelds by itself. It will, therefore, not be able to develop any new ﬁelds. Why is the government willing to accept that future loss? To answer that question we must understand who supports President Morales and what would these supporters want from him. He is the ﬁrst indigenous person to have been democratically elected to lead Bolivia and

Governance and Political Risk

457

like most of his supporters, he used to be a coca farmer. His election platform included a promise to regain control of Bolivia’s natural resources. His support comes from the poorest sections of the Bolivian society. In 1999, the most recent year for which the data are available, Bolivia had the worst Gini coefﬁcient in the world (63.0) for the income distribution in the rural area.12 What would the peasants lose if more oil ﬁelds could not be developed? What have they gained from the past revenues generated by the oil ﬁelds? If the majority of the people – peasants who elected President Morales – see a net beneﬁt from receiving a larger share of what may become a stagnant pie then that is the rational policy that the government should follow. In simple welfare terms, the marginal beneﬁt to the majority exceeds the marginal loss to the minority. In terms of our political risk model, government wants to renegotiate the contract with the foreign investors and its motivation is welfare of the society. In this case, there is no reason to believe that President Morales is motivated by economic self-interest in the form of corruption. In this situation, there is reason to believe that the foreign investors have contributed to creating risks for themselves. They acquired the gas ﬁelds in the 1990s with the connivance of the then president of the country circumventing the legal process that would have required the approval of the congress. The process helped create resentment against foreign investors – providing President Morales the justiﬁcation for his actions. Similar actions regarding the redistribution of land did generate opposition from the affected interest groups – as is expected in a democratic society. Coke’s actions in Uzbekistan may also have created a situation where the ﬁrm contributed to creating risk for itself. About ﬁve years ago, the ﬁrm sided with one of the most corrupt regimes to come out of the break up of former Soviet Union in order to safeguard its market position in the country. In the process it ignored the interests of other stakeholders – including some minority shareholders. Unfortunately for the ﬁrm, these shareholders have been able to use the international legal system to ﬁle claims of negligence against the ﬁrm. The political risk, in this case, arises from the collaboration between the ﬁrm and the corrupt leaders.

POLITICAL RISK: A RATIONAL PERSPECTIVE This paper takes a different view of the actions of a government than has been the norm in the literature on political risk. Taxation of foreign investors, like the taxation of domestic investors, is a balancing act: Current

458

ARVIND K. JAIN

revenue must be traded off against future revenues. Governments will try to extract the highest level of beneﬁts from investors in the society that they can, while trying to ensure that tax rate do not scare the future investors away. Bargaining between a host economy and foreign investors is a continuous process. Within the legal boundaries, governments will try to maximize the share of rents they can extract from foreign investors. Governments, however, consist of individuals who, as agents of the public, have a tendency to place their self-interest above those of principals they represent.13 Displacements of the objectives of the government will lead to a different taxation, and hence political risk, level than would happen with an honest government. Assessment and measurement of political risk, therefore, involves an assessment not only of the objectives of the government but also of the strength of political institutions that keep the government’s behavior under check. In this paper, we have attempted to identify some variables that will determine the range within which host governments can negotiate with foreign investors. Honest governments will take a long perspective, partly because it will lead to good solutions and partly because a well-developed political and social structure will provide incentives to the politicians to take the correct decisions. When corruption becomes an important characteristic of a society, governments will deviate from the ideal behavior to satisfy their personal goals. Political risks in such a situation will be much higher.

NOTES 1. ‘‘Bottled up: Why Coke stands accused of being too cozy with Karimovs.’’ (Alden, 2006, June 14). Financial Times, p. 8. 2. Daniels, Radebaugh, and Sullivan (2004, p. 89) offer a similar deﬁnition. 3. Contributors to Moran (1998) suggest varieties of techniques to manage political risk. Contributors to Moran (2001) focus mostly on insurance as a technique for managing this risk. 4. A reading intended to educate managers about political risk informs that ‘‘political risk arises from the vagaries of governmental action’’ (Harvard Business School, 1997, p. 1). Authors of a text book on international business advise ﬁrms to assess political risk by seeking answers to questions such as ‘‘When making changes in its policies, does the government act arbitrarily or does it rely on the rule of law?’’ (Grifﬁn & Pustay, 2002, p. 73.) 5. Tax policies followed by the Bush government since 2001 would clearly fall in this category. 6. Moran (1998, pp. 9–12) provides a brief summary of the concepts.

Governance and Political Risk

459

7. Empirical estimates of bargaining powers estimated by Fagre and Wells (1982) and Poynter (1982) fall in this category. Wells (1998, pp. 28–30) recognizes the dynamic nature of obsolescing bargaining but claims that host governments do not take future costs of renegotiations into account (p. 27). 8. See Jain (2001) for a full discussion of what constitutes corruption. 9. For the importance of internationalization, see Dunning (2003). 10. See Jain (2001, pp. 77–80) for a discussion of the importance of market imperfections for continuation of corruption. 11. We must remember that foreign investors tend to suffer from disaster myopia. There are too many examples of ﬁrms and banks investing in countries that had – not too ago – defaulted on their obligations. To some extent, this merely reﬂects a futureoriented view. See Jain (2005). 12. See www.wider.unu.edu/wiid/wiid.htm 13. See Jain (1987) for a model of political leaders as agents and its impact on the objective of the government.

REFERENCES Alden, E. (2006). Bottled up: Why coke stands accused of being too cosy with the Karimovs. Financial Times, 14 June, 8. Beamish, P. W., Morrison, A., & Rosenzwieg, P. (1997). International management: Text and cases (3rd ed.). Burr Ridge, IL: Irwin. Daniels, J. D., Radebaugh, L. H., & Sullivan, D. P. (2004). International business: Environment and operations (10th ed.). New York: Pearson Education. Dunning, J. H. (2003). Some antecedents of internalization theory. Journal of International Business Studies, 34(1), 108–115. Fagre, N., & Wells, L. T., Jr. (1982). The bargaining power of multinationals and host governments. Journal of International Business Studies, 13(2), 9–24. Grifﬁn, R. W., & Pustay, M. W. (2002). International business: A managerial perspective (3rd ed.). Upper Saddle River, NJ: Prentice Hall. Harvard Business School. (1997). Note on political risk analysis. Boston: Harvard Business School. Henisz, W. J. (2002). Politics and international investment: Measuring risk and protecting profits. Cheltenham, UK: Edward Elgar. Hymer, S. (1976). The international operations of national firms. Boston: MIT Press. Jain, A. K. (1987). Agency problem and the international debt crisis. Proceedings of the fourth symposium on money, banking, and insurance (geld, banken und versicherungen) (Vol. 1, pp. 367–391). West Germany: Karlsruhe. Jain, A. K. (2001). Corruption: A review. Journal of Economic Surveys, 15(1), 71–121. Jain, A. K. (2005). Investor behavior and global ﬁnancial crises. Unpublished paper. (An earlier version of this paper appeared. In: F. Michael, H. Ulrich, & R. Markus, (Eds), 2000. Risk management: Challenge and opportunity, essays in honor of Gunter Dufey (pp. 231– 245). Berlin: Springer-Verlag.) Moran, T. H. (Ed.) (1998). Managing international political risk. Malden, MA: Blackwell. Moran, T. H. (Ed.) (2001). International political risk management: Exploring new frontiers. Washington, DC: The World Bank Group.

460

ARVIND K. JAIN

Poynter, T. A. (1982). Government intervention in less developed countries: The experience of multinational companies. Journal of International Business Studies, 13(1), 9–26. Rugman, A. M., & Hodgetts, R. M. (2000). International business: A strategic management approach (2nd ed.). New York: Pearson Education. Triesman, D. (2000). The causes of corruption: A cross-national study. Journal of Public Economics, 76(3), 399–457. Vernon, R. (1971). Sovereignty at bay. New York: Basic Books. Wells, L. T., Jr. (1998). Good and fair competition: Does the foreign direct investor face still other risks in emerging markets? In: T. H. Moran (Ed.), Managing international political risk (pp. 15–43). Malden, MA: Blackwell.

STRATEGIC TRADE POLICY FOR SMALL STATES: A POLITICAL ECONOMY PERSPECTIVE Jaleel Ahmad ABSTRACT Much of the literature on strategic trade policy deals with industries and sectors characterized by international rivalry for market shares, and the struggle to capture ‘‘rents’’ over and above normal factor rewards. The present paper explores the validity and implications of strategic trade policy for small ‘‘states’’ and small firms that are not major players in international markets. The smallness of the firms may, in fact, be an advantage rather than a hindrance. The implications of smallness for strategic behavior are examined in the framework of a simple gametheoretic framework. These insights become sharper when extended to intra-industry trade in differentiated products. The desirable policy interventions for small countries and firms are quite different from those for large firms.

1. INTRODUCTION Among recent theoretical discussions of trade policy, the approaches generally known as strategic trade policy have aroused considerable attention as Value Creation in Multinational Enterprise International Finance Review, Volume 7, 461–473 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07019-1

461

462

JALEEL AHMAD

well as controversy. Two innovative aspects of these approaches are particularly noteworthy. First, strategic trade policy analyses have attempted to incorporate as an integral feature the insights from recently reﬁned framework of international trade in imperfectly competitive markets – a framework missing from much of the traditional trade theory. The inclusion of non-competitive market structures brings into focus the ‘‘strategic interaction’’ between ﬁrms in their price and quantity decisions. Second, at the same time and perhaps more signiﬁcantly, it also highlights the role of ‘‘strategic intervention’’ by policy-makers with a view to change the market outcome in the desired direction. Strategic interventions in this framework go beyond the domain of customary trade policy. Given that a signiﬁcant proportion of world trade takes place in markets characterized by varying degrees of non-competitive behavior, some novel arguments for trade intervention (protection and industrial policy) have been developed. It should be noted, however, that while the theoretical structure of strategic trade policy is generally uncontested there is no unanimity on the need or the efﬁcacy of such policy interventions. The implicit context for much of strategic trade policy is believed to be in industries characterized by international rivalry for market shares and aggressive producer behavior. The oft-quoted examples are the struggle for a dominant market share between Boeing and Airbus Industrie (a European consortium) for commercial passenger aircrafts, and between American and Japanese ﬁrms producing 16 K Random access Memory chips, or between the United States and Japan to be the ﬁrst in commercial production of high deﬁnition television (HDTV). Accordingly, strategic trade policy would seem to be relevant only to countries populated by large, monopolistic ﬁrms with sizable stakes in the world market. Even though smaller rivals may continue to produce similar goods or close substitutes, they remain largely on fringes, unable to contest the market supremacy of the larger ﬁrms. Nevertheless, the insights from strategic trade policy literature raise broader questions, and thus may have a potential relevance to countries that are not necessarily large, but may have ﬁrms with signiﬁcant shares in the international market for some products. The relevance of strategic trade policy for small ‘‘states’’ (to be deﬁned momentarily) has so far not been the subject of systematic analysis. The purpose of the present paper is to explore without a formal model the implications and the validity of strategic policy for small states that are not major players in the international market. The paper is divided into the following sections. Section 2 deﬁnes ‘‘smallness’’ of countries (or states) which are the focus of this paper. Section 3 contains a brief interpretation of the main arguments of strategic trade

Strategic Trade Policy for Small States

463

policy that are relevant to the discussion at hand. Section 4 is devoted to an analysis of the main points of relevance of strategic trade policy to small states. Some broader implications of strategic trade policy for small states are the subject of Section 5. Section 6 discusses the role of policy interventions.

2. SMALL STATES Smallness of countries in international trade is an elusive concept and does not permit an unambiguous deﬁnition. The assumption of ‘‘small’’ open economy is standard in much of trade policy to denote price-taking behavior in international markets, and thus enables one to work with the analytically simpler framework of perfect competition. This underscores the difﬁculties of incorporating large parts of non-competitive market structures in the analysis of international trade. Smallness in this connotation has, however, nothing to do with the geographic or economic size of countries. Many otherwise large economic entities are deemed ‘‘small’’ in the international trade to the extent that they do not dictate the terms at which they trade with the rest-of-the-world. The notion of smallness in this paper is somewhat different. It encompasses, but goes beyond, the mere absence of monopoly power in international trade. Small countries in this paper denote countries that are small economic entities in terms of geography and population, endowments of land and natural resources, capital stocks and skilled labor, and economic infrastructure. But above all, they have a limited domestic market that forces an outward orientation. Some small countries may, nevertheless, have large ﬁrms with signiﬁcant presence in the specialized international markets for some products. Some of these small-country ﬁrms may, in fact, have residual monopoly power in price-setting that they share with their counterpart ﬁrms from large countries. Examples of Switzerland, Sweden, South Korea, Singapore, and Malta readily come to mind – countries whose ﬁrms often have a signiﬁcant presence in world markets. These ﬁrms are, however, not in the same league as Boeing in aircrafts, General Motors in automobiles, Dell in personal computers, or Sony in electronics. Despite notable stakes in international markets, small-state ﬁrms are seldom in a position to lock horns with powerful international duopolies. Nevertheless, small domestic markets mean that a signiﬁcant proportion of their output is destined for international markets and, accordingly, they are often aggressive competitors with other large and small

464

JALEEL AHMAD

ﬁrms. Within this broad focus, it is difﬁcult to be precise about the role of per capita income in deﬁning smallness of countries. Flexibility dictates that we include countries with high and medium per capita income that otherwise meet the smallness criterion.

3. ELEMENTS OF STRATEGIC TRADE POLICY Beginning with the seminal work of Brander and Spencer (1983, 1984, 1985), strategic trade policy has spawned a growing literature of its own.1 Any substantive review of strategic trade policy is outside the scope of this paper.2 The discussion here is conﬁned to the main elements of strategic trade policy that make its conclusions signiﬁcantly different from those of standard neo-classical trade policy. Strategic trade policy represents a radical departure from its traditional counterpart, in that it employs a market structure signiﬁcantly different from perfect competition. In line with the many newer developments in trade theory, the strategic trade policy has employed a wide variety of monopolistic competition models. They range from duopoly (in both Cournot and Bertrand versions), and oligopolistic coordination (on the lines of Stackelberg ‘‘ﬁrst mover’’) to the Chamberlin large group monopolistic competition in differentiated products. Even these extensive applications barely scratch the surface of all possible market conﬁgurations that are ‘‘imperfect’’. The introduction of market structures other than perfect competition appears to have two signiﬁcant implications. One is that international trade in some products (aircrafts, for example) takes place in an environment where strategic interaction between the small number of ﬁrms determines equilibrium quantities and prices. The core of the ﬁrm strategy consists of attempts to garner a higher market share at the expense of rival ﬁrms. In the extreme case, as highlighted in the game-theoretic studies on the subject, the strategic outcome ends up as an either-or-proposition: either the domestic or the foreign ﬁrm emerges the only viable producer for the world market. While rivalry and exclusion are the hallmarks of strategic behavior, there is a tolerance of (or coexistence with) peripheral or ‘‘fringe’’ producers as long as they are unable or unwilling to ‘‘contest’’ the market and, accordingly pose no threat to the monopoly position of big players. This last observation has great potential signiﬁcance for the relevance of strategic trade policy to small states, which we will discuss in detail.

Strategic Trade Policy for Small States

465

The other implication of non-competitive market structure is that, by and large, ﬁrms earn substantial excess return or ‘‘rents’’, over and above what can be attributed to normal factor rewards. Excess returns may eventually be competed away and prices equal average costs due to entry of new ﬁrms if perfect competition prevails or, in extreme cases, if anti-trust actions break up the monopoly. When effective barriers to entry exist, however, a few ﬁrms may continue to retain signiﬁcant monopoly power in segments of market and continue to earn rents indeﬁnitely. In case of international duopoly, such as the production of large passenger aircrafts, the possibility of high rents to incumbent ﬁrms is magniﬁed. The existence of these rents and the struggle to capture them is what lies behind the appeal of strategic trade policy.

3.1. Strategic Trade Policy Intervention The strategic trade policy postulates that intervention by domestic governments (by means of tariffs, domestic taxes or subsidies) can alter the market outcome in favor of domestic ﬁrms. In other words, the purpose of strategic trade policy is to shift potential rents from the foreign to domestic ﬁrms, or to prevent domestic ﬁrms from losing them. Such trade policy initiatives can be preemptive, with a view to deter foreign ﬁrms from entering the market in the ﬁrst place, or they may arise as a ‘‘countervail’’ to subsidies or other ﬁnancial support received by foreign ﬁrms. Strategic trade policy, it is argued, can use such tactical measures as, targeting, pre-commitment, assurance, threats and, often, bullying as part of its arsenal. Clearly, sectors and industries differ in their potential for producing ‘‘excess’’ returns. Strategic trade policy literature has, therefore, given prominence to the notion of ‘‘strategic sectors’’ as those where potential rents and, hence international stakes, are high. High potential rents can arise from a total dominance of the world market, or, can be acquired gradually from economies of scale and learning curves, ‘‘clustering’’ or other externalities that may spark reactive growth of other sectors. But like a concept that invoked to explain so many different things cannot explain much, the notion of strategic sectors in strategic trade policy remains imprecise and open-ended. If domestic ﬁrms by their actions alone are able to expropriate industry rents to themselves, there would seem to be no role for strategic intervention by means of trade policy. The advocates of strategic trade policy argue, not quite convincingly, that strategic options available to ﬁrms lack sufﬁcient credibility and ‘‘pre-commitment’’ to deter entry of foreign ﬁrms, or to

466

JALEEL AHMAD

reduce their number, and thus, may have to forego the rents. More convincing is the argument that foreign government subsidies can often disturb established patterns of production and trade, and ﬁrms’ actions alone are not sufﬁcient to counteract them. Hence, strategic interventions in the form of production subsidies or protective barriers are required to swing the outcome in favor of domestic ﬁrms. In the well-worn illustration of the wide-bodied jet aircraft market, for instance, a government subsidy to airbus Industrie (a consortium of European governments) may snatch away the long-standing competitive advantage that the U.S. producers had because of their early lead. In the case of international oligopoly (more than two producers) strategic intervention may ‘‘target’’ an increase in the number of domestic ﬁrms at the expense of foreign ﬁrms. In some versions of strategic trade policy, rentshifting is proxied simply by a change in the relative market share of domestic ﬁrms vis-a`-vis foreign ﬁrms.3 It is clear, therefore, that in strategic trade policy models, policy intervention in the form of subsidies or trade barriers is assumed to have a benign role in comparison with other types of trade intervention where ‘‘deadweight’’ losses are common. Trade policy interventions of the strategic kind are not to be confused with ‘‘large country’’ tariffs or quota restrictions for terms-of-trade improvement. Strategic interventions in appropriate circumstances, it is argued, far from reducing national welfare may actually increase it by increasing the number of domestic high-rent ﬁrms. The original Brander–Spencer formulation avoided the problem of consumer surplus losses from trade intervention by assuming that international rivalry between domestic and foreign ﬁrms is conﬁned to markets in third countries. Even in cases of international rivalry in markets of third countries, it is recognized that potential rent-shifting as a result of trade policy intervention must be high enough to offset the cost of any subsidization. A considerable part of the appeal of strategic policy lies in the assertion that strategic actions aimed at imperfectly competitive industries may be in the national interest. The national interest is, of course, being equated with the capture of international monopoly rents by domestic ﬁrms. But in a recent assessment of trade policies undertaken for the National Bureau of Economic Research (NBER), Feenstra (2001) gives the strategic trade policy a failing grade. The widespread claim that strategic trade policy is in the national interest, says Feenstra, ‘‘has been shown quite convincingly to be false’’ (p. 5). This judgement may be a little premature, since as Feenstra acknowledges, each new type of market structure and conduct needs to be investigated carefully.

Strategic Trade Policy for Small States

467

4. STRATEGIC TRADE POLICY AND SMALL STATES In this section, we undertake a preliminary assessment of the relevance of strategic trade policy for small states. Largely because of the use of Cournot, quantity-based, competition in early strategic trade policy modeling, there is a presumption that such policies can be relevant only for large countries populated with large monopolistic ﬁrms. This is because in a Cournot-type duopoly, a ﬁrm can appropriate larger proﬁts by threatening to produce a larger quantity than its rivals. In other words, the ﬁrms’ dominant strategy consists of a threat to become larger. But this is only one possible strategy. In a model of Bertrand-type ‘‘price’’ competition, Eaton and Grossman (1986) demonstrate that a ﬁrm can actually garner higher proﬁts by remaining small and thus contributing to higher prices. In this variant, it is clear that the smallness of the ﬁrms is actually an advantage, rather than a handicap. Moreover, strategic trade policy turns out to be the opposite of the one in Cournot-type duopoly, i.e., a production or an export tax, rather than a subsidy. While the international struggle for rents may indeed be more appropriate for large countries, other implications of strategic behavior in non-competitive market structures apply equally to small-state ﬁrms that compete in the international market. When international markets for goods and services are characterized by elements of oligopoly, ﬁrms from small states may be accommodated (or rather, tolerated) as ‘‘fringe’’ suppliers, as long as they do not threaten the hegemony of large ﬁrms. But such an outcome is not automatic; it presupposes an appropriate strategic behavior on the part of small-state ﬁrms. The ‘‘puppy-dog ploy’’ of being a small non-aggressive competitor may, for instance, dictate a strategy of deliberate under-investment in order to remain small.4 The economies of scale thus foregone may be less of a handicap than is commonly supposed. This is because the higher average industry selling price resulting from proﬁt-maximizing strategy of large ﬁrms will allow a small-state ﬁrm to operate at a higher cost margin and still be proﬁtable.

4.1. A Game-Theoretic Interpretation Much of the strategic behavior described above can be understood conveniently in Fig. 1. We assume that both the large-country ﬁrm (designated the L-ﬁrm) and the small-state ﬁrm (the S-ﬁrm) choose their prices to maximize proﬁt, taking the price of the other ﬁrm as given – a standard noncooperative Nash behavior. In other words, we rule out collusion between

468

JALEEL AHMAD

Fig. 1.

A Game-theoretic Interpretation.

the two types of ﬁrms. The prices set by L-ﬁrm and the S-ﬁrm for their products, which may be assumed as imperfect substitutes for each other, are denoted as p and p, respectively. The reaction curve labeled p (p) shows the L-ﬁrm’s proﬁt-maximizing prices for given values of the S-ﬁrm’s price, p. The S-ﬁrm’s reaction curve p (p) is similarly deﬁned. The assumption that an increase in L-ﬁrm’s price raises the proﬁts of its smaller rival, thereby increases the marginal returns from increasing p, means that the reaction curves are positively sloped in the (p, p) plane. A rise in p, for instance, will increase the marginal return (or reduce the marginal loss) from increasing p, and would, therefore, induce the S-ﬁrm to raise its price as well. We assume further that the L-ﬁrm does not base its pricing decision on what the S- ﬁrm does. Hence, its reaction curve, p (p), is assumed to be vertical, or nearly so, as long as the strategy of the small ﬁrm to remain small is credible. The L-ﬁrm may ﬁnd the strategy to reduce its price to compete with the S-ﬁrm too costly. In other words, while the L-ﬁrm can effectively ignore the pricing decisions of its smaller rival, the S-ﬁrm has little choice but to follow the price strategy of the L-ﬁrm. Finally, we note that the iso-proﬁt contours are downward sloping to the left of the S- ﬁrm’s reaction curve and upward sloping to the right, crossing

Strategic Trade Policy for Small States

469

the curve horizontally. This is because to the left of p (p), p is below its proﬁt-maximizing value which it attains only on its reaction curve. An increase in p in that region, therefore, increases proﬁts. Similarly, to the right of p (p) higher p reduces proﬁts. Because lower p reduces S-ﬁrm’s proﬁts, higher iso-proﬁt contours correspond to higher proﬁts. What is important in this discussion is that the S-ﬁrm is actually better off because of the fact that in an international market dominated by the large country ﬁrms, its set of choices is shrunk. The price ﬂoor set by the L-ﬁrm commits the S-ﬁrm to set the same price, even though it may be higher or lower than it otherwise would. If the large ﬁrm were to raise its price from p to p1, the optimal strategy for the S-ﬁrm would be to raise its price as well. This is so not only because S-ﬁrm’s proﬁts are higher but also because such a move does not provoke an aggressive reaction from its large-country competitors. These assumptions imply that Cournot competition is not a relevant strategy for the small-states ﬁrms. It also shows that the presence of large monopolistic ﬁrms in the international arena is not necessarily a handicap to the proﬁtability of small-state ﬁrms. In many cases, it may be an advantage. The intersection of the two reaction curves at N yields the Nash equilibrium level of prices set by the two types of ﬁrms. The Nash equilibrium depicted in Fig. 1 makes it clear that when the large country ﬁrms follow a strategy of maintaining higher prices, the gains accruing to the small-state ﬁrms are, in fact, higher. In this respect, the welfare implications for the small country in an oligopolistic international market are not too different from the gains in perfectly competitive markets, but with an important dissimilarity. Small country ﬁrms are consigned to a price-taking behavior in both types of market structure. But while the small country ﬁrms can sell indeﬁnitely large quantities of their output at the world price, as argued by the neo-classical trade theory, a small ﬁrm’s free rider gains in duopolistic and oligopolistic markets are contingent on the ﬁrm remaining ‘‘credibly’’ small. A larger production by the S-ﬁrm that threatens to encroach on the sales of the L-ﬁrm may prompt the latter to pursue an aggressive strategy to protect its market share. Such an action would clearly not be in the interest of the S-ﬁrm.

5. FURTHER IMPLICATIONS The implications of strategic behavior on the part of small-state ﬁrms come into sharper focus in the context of intra-industry trade in differentiated products, where Chamberlin-type monopolistic competition is the relevant market structure. A possible strategic behavior for the small-state ﬁrm in

470

JALEEL AHMAD

this type of market structure would be to ﬁnd a ‘‘niche’’ (or ‘‘boutique’’) in the diversiﬁed, multi-product international market for manufactured goods. The niche market strategy would seem to be within the grasp of even the smallest ﬁrm, since such specialization is not tied to the existence of a ‘‘primary’’ factor. The niche strategy can clearly be acquired and promoted by appropriate policies, such as technology transfer, minor technical adaptation, and superior service. But while the discussions of niche markets in business strategy literature assume that the pricing decisions of niche ﬁrms are not tied to the pricing decisions of large ﬁrms, the strategic policy analysis in this paper points to a radically different view. Not only is an active pricing strategy for similar products strictly the prerogative of the large ﬁrms, the small country ﬁrms must follow these price signals to make this strategy proﬁtable for both. This is not to suggest that the strategy proﬁle of small-state ﬁrms is an empty set. The strategic ‘‘game’’ described in the previous section is interesting fundamentally because it exhibits externality – that the large-country ﬁrms do affect the payoff of the small-state ﬁrms, not always negatively, as commonly supposed. We have conﬁned our attention in this paper to competition between the large-country ﬁrms and small-state ﬁrms in the larger international market and not necessarily in each other’s domestic markets. Allowing for trade between large and small countries introduces some new elements. Krugman (1980) has argued that when a large and a small country trade with each other, the small country tends to specialize in the constant returns-to-scale sector, while the large country specializes in the production of increasingreturns goods. In a similar vein, Amiti (1998) has shown that small country ﬁrms tend to specialize in goods with low transportation costs. In a continuum of industries ranging from zero minimum efﬁcient scale to high minimum efﬁcient scale, Holmes and Stevens (2002) show that goods in the high minimum efﬁcient scale category are not produced in the small country. Furthermore, as is typical in oligopoly games, there are multiple equilibria with respect to the country in which a given industry locates.

6. POLICY INTERVENTIONS The possible roles of strategic policy interventions on the part of governments in small states appear less clear. Subsidies to small-state ﬁrms would seem to be counter-productive if they encourage small ﬁrms to become large and, thus, lose their strategic advantage. The role of tariffs is similarly ambivalent. Tariffs may be part of a broader strategy if they restrain output

Strategic Trade Policy for Small States

471

and keep the small ﬁrms small, in the connotation employed in this paper. On the other hand, tariffs may encourage entry of new ﬁrms to produce the protected output and may provoke retaliation from large ﬁrms. The role of economies of scale is equally equivocal. While they enable a reduction in average costs and thus enhance competitiveness of the ﬁrm, they do so only by increasing the output of the ﬁrm and the industry. By contrast, the outlook for exploiting the economies of ‘‘scope’’ would appear to be more promising. Diversiﬁcation of the ﬁrm’s activities within its scope would preserve the notion of ‘‘smallness’’ of ﬁrms for any given product, while at the same time allowing for a more diversiﬁed and more ﬂexible industrial structure for the country in question. These other segments within the scope of a given ﬁrm may, in turn, begin to have the strategic relationship with their large-country counterparts on lines analogous to the one indicated in this analysis. In small states that are not major players in international trade there would seem to be a good case for a lump-sum subsidy to ﬁrms directed speciﬁcally to the development of a subset of activities that may lie within the scope of a ﬁrm and for which complementary inputs may be available within the ﬁrm. This sort of intervention would be more in line with strategic policy than an output subsidy which would lead inevitably to a larger output of broadly similar variety. Such a course of action essentially means that government policies (both trade and non-trade) should be directed toward ‘‘strategic’’ sectors and not across-the-board, a point we take up in the next section.

6.1. Strategic Sectors The strategic trade policy literature has given prominence to strategic-ness of the sectors, such as in the case of duopolistic rivalry in the international market for large passenger aircrafts. The implicit (and largely unexamined) assumption here is that aircrafts manufacture confers strategic advantages on the country that produces them, i.e., it is in the ‘‘national interest’’. But in reality, the role of strategic sectors appears to be more critical and decisive in the case of small states than in large ones. Large countries, by virtue of their economic size, have a large number of traded-goods sectors with high value-added per worker, and high rents. In fact, the number of such sectors may be so large that the notion of strategic sector itself loses much of its cogency. Even in a narrower connotation, the number of high-rent sectors with potential for international rivalry may be quite large and varied with the result that no clear ordering may emerge. If so, an interventionist policy

472

JALEEL AHMAD

that attempts to choose between them as ‘‘targets’’ may be both unnecessary and counter-productive. Attempts to ‘‘target’’ one sector may crowd out others if all sectors compete for similar inputs in inelastic supply, as argued by Dixit and Grossman (1986). By contrast, the notion of strategic sectors in small states may have a far more analytical and practical justiﬁcation. Small states, by virtue of their small size, are unlikely to have more than a few sectors (and in extreme cases, a single sector of international exposure, for example, electronics in South Korea) that are crucial to economic viability and progress. These sectors may have a high payoff for the country as stakes in international competition and may make the difference between the country’s presence on the international scene or its obscurity. Small-state ﬁrms competing with big rivals in the international market may not be able to redistribute a signiﬁcant proportion of monopoly rents in their favor, but at least, they can sit at the same table!

7. CONCLUSION This paper has examined the relevance of strategic trade policy to small states. A preliminary analysis suggests that small-state ﬁrms can share, to a limited extent, in the rents in international oligopoly markets as long as they remain small players. The strategic action that turns out to be critical for them is to remain on the sidelines and avoid the temptation to become large. The case for strategic policy intervention to inﬂuence the market outcome is less clear in small states. Such a role would seem to be limited to strategic support for small-state ﬁrms to make them efﬁcient peripheral producers in a market dominated by large ﬁrms. In addition, we ﬁnd that paradoxically the notion of strategic ‘‘sectors’’ appears to have more relevance to small states than to large ones. These sectors have to be selected carefully to prevent proliferation which will likely dissipate their strategic advantage.

NOTES 1. Also see Dixit (1984, 1987a), Dixit and Grossman (1986), Eaton and Grossman (1986), Krugman (1987), Katz and Summers (1989) and other references in these papers. 2. Topical reviews of strategic trade policy are contained in Grossman and Richardson (1985) and Krugman (1986).

Strategic Trade Policy for Small States

473

3. Rent-shifting should be distinguished from ‘‘rent extraction’’ if foreign ﬁrms have monopoly power. In the latter case, the relevant policy would be an optimal tariff. Also see Dixit (1987b). 4. See Fudenberg and Tirole (1984) for ‘‘puppy-dog ploy’’ of small ﬁrms.

REFERENCES Amiti, M. (1998). Inter-industry trade in manufactures: Does country size matter?. Journal of International Economics, 44, 231–255. Brander, J. A., & Spencer, B. J. (1983). Strategic commitment with R&D: The symmetric case. Bell Journal of Economics, 14, 371–389. Brander, J. A., & Spencer, B. J. (1984). Tariff protection and imperfect competition. In: H. Kierzkowski (Ed.), Monopolistic competition in international trade (pp. 194–206). Oxford: Oxford University Press. Brander, J. A., & Spencer, J. (1985). Export subsidies and International market share rivalry. Journal of International Economics, 18, 83–100. Dixit, A. (1984). International trade policy for oligopolistic industries. Economic Journal, 16, supplement. Dixit, A. (1987a). Issues of strategic trade policy for small countries. Scandinavian Journal of Economics, 89, 349–367. Dixit, A. (1987b). Strategic aspects of trade policy. In: T. Bowley (Ed.), Advances in economic theory (pp. 329–362). Cambridge, UK: Cambridge University Press. Dixit, A., & Grossman, G. M. (1986). Targeted export promotion with several oligopolistic industries. Journal of International Economics, 21, 233–250. Eaton, J., & Grossman, G. M. (1986). Optimal trade and industrial policy under oligopoly. Quarterly Journal of Economics, 101, 383–406. Feenstra, R. C. (2001). International trade and investment. Program Report, National Bureau of Economic Research, Cambridge, MA. Fudenberg, D., & Tirole, J. (1984). The fat-cat effect, the puppy-dog ploy, and the lean and hungry look. American Economic Review, papers and proceedings, 74, 361–366. Grossman, G. M., & Richardson, J. D. (1985). Strategic trade policy: A survey of issues and early analysis. Special Papers in International Economics, no. 15. Princeton University, Princeton, NJ. Holmes, T. J., & Stevens, J. J. (2002). The home market and the pattern of trade: Round three. Finance and Economics Discussion Series, no. Z.11, Federal Reserve Board, Washington, DC. Katz, L. F., & Summers, L. H. (1989). Can Inter-industry wage differentials justify strategic trade policy. In: R. Feenstra (Ed.), Trade policies for international competitiveness (pp. 85–116). Chicago, IL: University of Chicago Press. Krugman, P. (1980). Scale economies, product differentiation and the pattern of trade. American Economic Review, 70, 950–959. Krugman, P. (Ed.) (1986). Strategic trade policy and the new international economics. Cambridge: MIT Press. Krugman, P. (1987). Strategic sectors and international competition. In: R. M. Stern (Ed.), U.S. trade policies in a changing world economy (pp. 207–232). Cambridge: MIT Press.

This page intentionally left blank

STRATEGIC IPO UNDERPRICING: THE ROLE OF CHINESE STATE OWNERSHIP Yong Wang and Xiaotian (Tina) Zhang ABSTRACT Initial public offering (IPO) underpricing remains a puzzle after decades of investigation. The stock markets in emerging economies are attractive to international investors but their unique characteristics need to be examined. Chinese stock markets experienced much more significant IPO underpricing than most other stock markets in the world. This paper offers a two-period wealth maximum model to explain the strategic IPO underpricing by state owners. Given the fact that the entire IPO procedure, including IPO price, is regulated and controlled by state owners, we argue that state owners strategically underprice the IPO, because they care less about the IPO proceeds but more about the wealth gain after IPO. The empirical finding of a positive relationship between IPO underpricing and state ownership in Chinese stock market is consistent with the wealth maximization hypothesis of IPO pricing. The paper offers better understanding for IPO procedure of state-owned enterprises in emerging markets.

Value Creation in Multinational Enterprise International Finance Review, Volume 7, 475–495 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07020-8

475

476

YONG WANG AND XIAOTIAN (TINA) ZHANG

1. INTRODUCTION Although national borders are losing their signiﬁcance both as a psychological and as a physical barrier, international investors today are still challenged by the diversity of various local business environments, which as a result inhibit their optimal global investment. Even though a substantial body of research has documented the so-called ‘‘home bias’’ for investors, there is limited understanding of the uniqueness of the ﬁnancial market of emerging economies and its impact on foreign investors. This paper tries to explain the abnormally underpricing phenomena of Chinese IPO market and offers better understanding of Chinese stock market to international investors. With an average growth rate of 8.5% from 1990 to 2003, China’s rapidly expanding economy has attracted considerable enthusiasm from foreign investors. As a result, foreign investors poured enormous amount of capital into the stock market of China. On the initial public offering (hereafter IPO) front, business analysts predict that IPO market in Asia (except Japan) will continue to be robust in 2006 and more than 60% of the IPO proceeds would be raised from companies based in China (Wall Street Journal, 2006) Chinese IPO companies, dominated by state-owned enterprises (hereafter SOEs), have increasingly attracted investors’ attention. Foreign investors are even prompted to put stakes in some Chinese SOEs. As demonstrated by the recent negotiation between Bank of China (state-owned and won approval to launch an IPO in 2006) and four famous foreign banks, many foreign investors attempt to access to the potentially large market and use IPOs as a way to obtain the beneﬁts. In China, after two decades of reform and privatization, roughly one third of China’s economy is still controlled by the government through SOEs. State ownership is the majority shareholder both before and after the IPO, therefore state ownership dominates IPO price setting decision. Furthermore, the IPO procedure is completely regulated by the government, which is the legal entity of all state owners in China. The totally controlled IPO procedure seems to leave no room for underpricing because the state ownership will bear the loss when it occurs. However, China’s stock market is one of the most striking markets in terms of the magnitude of IPO underpricing. This controversy begs for explanation. The underpricing of IPOs1 is one of the unsolved puzzles worldwide. Prior studies attempt to explain it from different perspectives. Jenkinson and Ljungqvist (2001) offer a comprehensive review of the existing theories and suggest that the offer price does not fully incorporate all available

Strategic IPO Underpricing: The Role of Chinese State Ownership

477

information. Given the fact that issuers rarely get upset about leaving money on the table, an alternative explanation is that issuers strategically underprice IPO. Issuers may use underpricing as a signal of ‘‘high-quality ﬁrms’’ to lower their cost of capital in subsequent funding exercise,2 or as a signal of ﬁrms value.3 The other stream of literatures (e.g., Brennan & Franks, 1995) links the IPO underpricing with the agency problem between managers and shareholders. Loughran and Ritter (2002) focus on wealth effect and argue that issuers would not feel upset leaving substantial money on the table if they simultaneously ﬁnd an increase in wealth. Issuers add up the loss from leaving money on the table with the wealth increase coming from a price jump on their retained shares and perhaps are better off than break even. Our study attempts to examine the total wealth change of state ownership in IPO price determination procedure. We argue that the wealth gain of state ownership comes from both the IPO proceeds and the wealth increase after IPO. A low IPO price may decrease the ﬁrst component but at the same time increase the second component. The state owner would adjust IPO price and trade off the two components. When the state owner holds a large proportion of ﬁrm’s total share un-circulating, he places more focus on the wealth gain after IPO. Consequently and rationally, the state owner would strategically reduce IPO price. Therefore, we expect that a great proportion of state ownership un-circulating would be associated with a low IPO price and a large IPO underpricing. This paper studies 408 IPOs on China’s Shanghai and Shenzhen stock exchanges, from January 1, 1998, to December 31, 2002. The empirical results show that the proportion of state ownership un-circulating is signiﬁcantly and positively related to the IPO underpricing, which is consistent with our prediction. As with the existing literatures, the time lag between issuing date and listing date is signiﬁcantly and positively related to the IPO underpricing. To our best understanding, this paper makes the following contributions. First, it points out that the IPO procedure in China is different from that in Anglo-American market; therefore ﬁrms’ behavior is consequently different. The uniqueness of IPO could help global investors to better understand the gain and risk associated within the procedure and further evaluate their investment. Second, it offers an alternative explanation of the striking underpricing phenomenon in Chinese stock markets. The state owner strategically underprices IPOs to maximize its total wealth. Third, as a complement to the agency theory explanation originated by Jensen and Meckling (1976), which appears to be particularly pertinent to ﬁrms with

478

YONG WANG AND XIAOTIAN (TINA) ZHANG

separation of ownership and control, our model well explains the scenario of little dispersion in ownership and control, where state ownership controls the ﬁrm by holding majority shares. The remainder of the paper is organized as follows. In Section 2, we describe the Chinese stock market and introduce the unique features in IPO process. Section 3 presents the model explaining state owner’s strategic underpricing to maximize wealth. Section 4 describes data and provides the empirical results. Conclusion is in Section 5.

2. CHINESE STOCK MARKETS AND IPO PROCEDURE After Shanghai Security Exchange opened in 1990, the stock market capitalization and the number of listed companies in China increased from $12.6 billion and 53 in 1992 to $516.79 billion and 1,205 in 2002. Stock market introduced alternative investment vehicle for Chinese citizens in addition to Treasury bill. The huge saving deposit4 can ﬁnally become a stimulus instead of a threat to the economy. On the other hand, going public was viewed as the best remedy for improving efﬁciency and expanding scale of the SOEs.5 With pressure on both demand and supply sides, China’s stock markets boomed in the last 13 years and became second only to Japan in Asia. China Securities Regulatory Commission (CSRC) functions as an authorized ministry of central government regulating the stock markets, with its goal to reform unproﬁtable and inefﬁcient SOEs. The IPO procedure in China has vestiges of China’s transitional economy with socialist planning. First, the new share issue quota is determined jointly by the State Council Securities Committee (SCSC), the State Planning Commission (SPC), and the People’s Bank of China (PBOC; central bank of China). The quota is allocated based on provinces as well as on municipalities. This process is signiﬁcantly different from market economy, where the stock exchange decides the qualiﬁcation of the ﬁrms. Second, new issues reﬂect only a small proportion of outstanding shares. The majority of shares are still owned by the state (e.g., China Mobile Phone and Sinopec, which is an oil and chemical producer), but these state-owned shares cannot be traded on stock exchanges. Tian (2001) reports that Chinese government and state-owned legal entities are the largest shareholders of 44% of Chinese public listed companies. More strikingly, the largest shareholders hold more than 40% of the equity in these Chinese public listed companies, playing a role of majority

Strategic IPO Underpricing: The Role of Chinese State Ownership

479

shareholders in equity markets. Third, share allocation is not based on market mechanism. Only the winners of IPO lottery, which accounted for small numbers of applying investors, were eligible to buying shares at IPO price. The winners could easily ﬂip IPO shares to other investors at higher prices and earn free money. Although an auction mechanism was introduced in recent years, the price was ﬁxed to a certain extent so that investors could only bid on the quantity of shares they would buy, but not on the price of all shares. Fourth, the IPO price is determined by the regulation. Before 1999, the offering price was set to be the product of net earnings per share and an issuer-chosen multiplier. After 1998, the Security Law was stipulated, which allows the issuers and underwriters to decide IPO prices, but it should still be authorized by the regulator. Fifth, after initial public offering, ﬁrms still need to apply for a date for their shares to be listed onto the stock exchanges. There is consequently a time gap between ﬁrm going public and shares being traded on the secondary market. According to CSRC data, with a median of 34 days, this listing time lag can range from 3 days to 12 years. During this time period, IPO subscribers’ investment is locked up.6 The magnitude of IPO underpricing on China’s stock market is striking. As noticed by Mok and Hui (1998), the underpricing of A-share in Shanghai stock exchange was 289% from 1990 to 1993. If the early years of China’s stock markets are taken into account, Su and Fleisher (1999) show that the underpricing could go as high as 948%. This paper attempts to explore the reasons behind the underpricing. Using La Porta, Lopez-de-Silance and Shleifer’s (1999) method, Tian (2001) shows that the Chinese government directly owns 28% of all the shares of China’s public listed companies and ultimately controls 44% of China’s public listed companies through pyramids, cross-holdings and reciprocal shareholdings. The state owner is not only the regulator (like CSRC) but also the stakeholder SOEs. Considering the severe underpricing of IPOs, the striking question is why issuers, here mostly state ownership, would leave huge amount of money on the table? Why would the CSRC allow severe underpricing which harms the interest of state ownership? Intuitively, because there are few alternative investment opportunities available for investors, and because CSRC regulates the IPO, the state owner would have made the most favorable price for itself and not voluntarily sacriﬁce the beneﬁt in going public. This paper develops a model to demonstrate that IPO underpricing is consistent with state owner’s total wealth maximization. The state owners may not be upset to have left huge amount of money on the table because

480

YONG WANG AND XIAOTIAN (TINA) ZHANG

their total wealth increases accordingly after the IPO. The IPO price is strategically underpriced to achieve a higher total wealth increase both during and after the IPO procedure. Previous studies in China IPO underpricing mainly concentrate on the investment risk. Tian (2003) states that there is a time value of the investment forgone during the lock-in period before the shares are listed on the stock exchange, and it must be compensated in the ﬁrst day return after listing. Under this hypothesis, the length of the time lag between the issuing date and listing date will be positively related to the magnitude of IPO underpricing. Our empirical study also controls this effect.

3. TWO-PERIOD WEALTH MAXIMUM MODEL The main point this paper attempts to make is that underpriced IPO may decrease state owner’s proceeds from IPO procedure, but at the same time it enlarges the wealth increase after going public. The issuer faces a trade off between wealth in the two periods because the objective of going public is not only to obtain the capital, but also to increase ﬁrm’s total value. Because of the limited and controlled supply of IPO shares, investors invariably rush to new IPOs whenever they are available. Underpriced IPOs with a high short-term return will consequently attract more investor attention and analysts’ coverage, which will in turn drive up the stock price even further in the short run. Since a large stake of shares is not circulating, the SOE will accordingly enjoy the huge increase in the market value of their total wealth. Therefore, like as Loughran and Ritter (2002), we argue that the total wealth effect makes state owner not upset by leaving money on table. Suppose the state owner holds a proportion a of ﬁrm’s total share Q uncirculating. It will sell the shares of (1a) Q at price P, with the proceeds of P (1a) Q. There exists the potential to sell the remaining shares after lock-in period. Suppose the owner can sell the shares of a Q at the highest price in the certain future period, Phigh, the gain is a Q Phigh : The state owner would attempt to achieve the combined wealth maximization in two periods by choosing the appropriate P level. Without the selling potential, the state owner will still care about a Q Phigh because one of the goals in deciding to go public is to maximize ﬁrm’s market value. Zingales (1995) argues that IPO underpricing is the ﬁrst stage for ﬁrm to ﬁnal sale, and consequently, IPO underpricing will result in a good price at ﬁnal sale. We make the additional assumption that a great IPO underpricing will lead to a higher price in future short period. Aggarwal, Krigman and

Strategic IPO Underpricing: The Role of Chinese State Ownership

481

Womack (2002) also found the similar result that manager’s strategic IPO underpricing brings the price up in the lock-in expiration sales. This assumption ensures that the revenue from selling shares in secondary market will be large when IPO is underpriced. Therefore, the state owner, as the largest shareholder, may attempt to decrease IPO price in order to obtain more proceeds or a higher market value in the secondary market. The momentum theory, which indicates that a high ﬁrst day return would lead to a high holding period return in new issuing market, is also consistent with our assumption. Booth and Chua (1996) demonstrate that IPO underpricing drives the liquidation in secondary market. Ibbotson, Ritter and Sindelar (1994) show that initial returns lead volume by 6-12 months. On the other hand, Aggarwal et al. (2002) argue that the large trading volume attracts investors’ interest, and the enhanced interest shifts the demand curve for IPO stock outwards and furthermore brings higher price in the certain period after IPO. Based on these studies, we argue that in the short run, underpriced IPO has higher ﬁrst day return and catches more investor and analyst attention. At the same time, such IPO will also incur higher trading volume. All these effects will work together and push the price even higher in the short run. When stock price could reach a higher level after IPO, the un-circulating shares would also appreciate accordingly. Even if a ﬁrm does not have a speciﬁc plan of selling additional shares in the near future, the total market value of the ﬁrm, both circulating and un-circulating shares, would already been measured at the market price. Considering the large proportion of a ﬁrm’s total shares which can’t be sold in IPO procedure, the state owner would place more attention on the beneﬁt after IPO and eventually strategically reduce the IPO price to maximize its proﬁt after IPO wealth increase. Therefore, we expect a negative relationship between IPO underpricing and the proportion of ﬁrms going public for circulating.

4. DATA AND EMPIRICAL RESULTS The data include the IPOs on both Shanghai and Shenzhen stock exchanges from January 1, 1998, to December 31, 2002. The IPOs data are from Lido Finance Database, Shanghai, China. We also manually crosschecked the completeness of the list with the data obtained from Stock Star Web site (www.stockstar.com). The company proﬁle, issuing date, circulating share volume, total share volume, stock price and trading volume were obtained from the China Publicly Trading Company Information Web site

482

YONG WANG AND XIAOTIAN (TINA) ZHANG

(www.cnlist.com). Originally we got 336 IPOs on Shanghai stock exchange and 135 IPOs on Shenzhen stock exchange. We deleted 63 of the IPOs due to the unavailability of the lottery rate or P/E ratio; the ﬁnal sample includes 408 IPOs. We test the model developed in Section 3 by examining the following empirical implications: ﬁrst, a great magnitude of IPO underpricing leads to a high trading volume, which in turn cause a higher price in the short run; second, the larger the proportion of state ownership un-circulating, the larger the IPO underpricing. Table 1 provides the summary statistics. Un-circulating state ownership proportion is deﬁned as the ratio of the sum of share of government directholding shares plus the cross-held by other SOEs (which cannot be circulated according to security law of China) to the total share of the ﬁrm. Ownership proportion is very high, with a mean of 71.87% and a median of 71.94%. The circulating shares are only a small proportion of the total shares of the IPO ﬁrms. The result is consistent with the hypothesis that the state owners care less about the IPO proceeds but much more about the wealth effect for the non-circulating shares in control. Consistent with the ﬁndings in earlier study, the IPO underpricing in our sample is huge, with a mean of 114.54% and a median of 78.58%. Given a mean market index return of 2.82%, if someone is lucky enough to win IPO lottery or bid, his investment will almost certainly be doubled in the ﬁrst day in ‘‘non-hot’’ market. This larger IPO underpricing is partly because of the extremely low lottery rate, with a mean of 0.62% and a median of 0.34%. Very few investors can get the allocations from IPO and only a tiny percent of the huge demand can be met at IPO stage. Therefore, the trading activities on secondary market would be tremendous. In a 25-day window after ﬁrst day, the holding return for the IPO stock, 8%, is quite larger compared with a market return of –2.82%. The volatility of price is relatively constant, with a mean of 0.609, implying that after going public IPO stock price keeps constant. Overall, the results suggest that the wealth increase after going public is stable in the short run. If we consider the highest price in the 25-day window, the wealth gain of the state owner is larger than considering the average price. The time lag between the issuing date and the listing date is very long, with a mean of 114 days and a median of 32 days. This means that after issuing data, the IPO ﬁrms will have to wait about a month for their shares to be listed in the stock exchange. Therefore, the issuers have to lower the IPO price to compensate the initial buyers for any risk contained in this time period. Table 2 shows a simple correlation matrix for selected variables. The state owner’s un-circulating proportion is positively, although insigniﬁcantly,

Strategic IPO Underpricing: The Role of Chinese State Ownership

Table 1. Variables Average price in 25 days Average volume in 25 days Circulating share (million) Highest price in 25 days Index return IPO UP Holding return Lottery rate Lowest price in 25 days Owner proportion P/E Time lag Total share (million) Volatility of price in 25 days Volatility of volume in 25 days

483

Summary Statistics.

Mean

Median

St. Dev.

Maximum

Minimum

13.667 4834.47 71.348 14.888 2.82% 114.54% 0.0861 0.6217 12.628 0.718 20.245 114.152 552 0.609 7993.164

12.096 4288.58 52.200 13.160 6.83% 78.58% 0.1148 0.3451 11.116 0.719 18.68 32 185 0.446 6230.75

8.217 3706.80 91.615 9.005 1.3872 1.9356 0.0503 0.9846 7.624 0.088 52.451 509.305 4350 0.612 8387.522

67.020 579.76 1540.000 75.778 702.69% 2056.8% 0.6072 11.1567 63.472 0.982 71.45 4541 86,700 6.002 133 620

2.306 55710.00 11.250 2.402 475.14% 53.29% 0.4317 0.0066 2.179 0.346 1 35 50 0.056 810.62

Note: The sample consists of 408 ﬁrms whose IPOs are between January 1998 and December 2002. The variable deﬁnition is as follows: Average price in 25 days is the average daily closing price in 25 days following IPO. Average volume in 25 days is the average trading volume (unit: millions of RMB) in 25 days following IPO. Circulating share is the share number in circulating after IPO. Highest price in 25 days is the highest daily closing price in 25 days following IPO. Index return is the Shanghai/Shenzhen stock market index return. IPO UP is the IPO underpricing deﬁned as the ﬁrst closing price minus IPO divided by IPO. Holding return is the holding return from the IPO’s ﬁrst day closing price through 25 days following IPO, deﬁned as the highest closing price in 25 days following IPO minus ﬁrst closing price divided by ﬁrst closing price. Lottery rate is reported in IPO procedure. Lowest price in 25 days is the lowest daily closing price in 25 days following IPO. Owner proportion is the proportion of state ownership un-circulating to the ﬁrm’s total share. P/E is the reference price to earning ratio of similar ﬁrms on the stock market, which is selected by SEC before IPO. Time lag is the difference of issuing day and listing day. Total share is the number of total shares of the ﬁrm. Volatility of price in 25 days is the daily closing price volatility in 25 days following IPO. Volatility of volume in 25 days is the daily trading volume volatility in 25 days following IPO.

related to the IPO underpricing. IPO underpricing is also positively related to the holding period return from the ﬁrst day to 25 days after IPO, with a value of 0.012. The two correlations show that a large state owner proportion un-circulating is associated with a large IPO underpricing, and the large IPO underpricing is also associated with a large holding period return after going public. The ﬁndings are consistent with our hypothesis that state owners strategically underprice IPO to obtain a large wealth in the period after going public. What’s more, we ﬁnd that the lottery rate is negatively related to IPO underpricing, with a value of 0.278 at a 1% signiﬁcant level.

Pearson Correlation Matrix of Selected Variables.

Average Volume

Circulating Share

Index Return

1

0.849 1

P/E Ratio

Time Lag

0.254

0.195

0.099

0.291 0.145 0.061

0.248

0.123

0.071

0.106 0.070 0.278 0.076 1 0.012 1

0.006 0.236 0.071 0.052

0.009 0.040 0.129 0.105

Lottery Rate

IPO UP

Holding Return 25

Owner Proportion

0.031

0.274

0.177

0.161

0.050 1

0.064 1

1

0.177 1

0.028 0.009 0.717 0.072 0.068 0.194 1

Note: This table reports Pearson correlation. The sample consists of 408 IPOs whose IPOs are between January 1998 and December 2002. The variable deﬁnition is as follows: Average volume is the average volume in 25 days following IPO. Circulating share is the share number in circulating after IPO. Index return is the Shanghai/Shenzhen stock market index return. Lottery rate is reported in IPO procedure. IPO underpricing (IPO UP) is deﬁned as the ﬁrst closing price minus IPO divided by IPO. Holding return is the holding period return from the IPO’s ﬁrst day closing price through 25 days following IPO, deﬁned as the highest closing price in 25 days following IPO minus ﬁrst closing price divided by ﬁrst closing price. Owner proportion is the proportion of state ownership un-circulating to ﬁrm’s total share. P/E ratio is the reference price to earning ratio of similar ﬁrms on the stock market, which is selected by SEC before IPO. Time lag is the difference of issuing day and listing day. Total share is the total share number of the ﬁrm. signiﬁcant at 10% level. signiﬁcant at 5% level, Signiﬁcant at 1% level,

YONG WANG AND XIAOTIAN (TINA) ZHANG

Average volume Circulating share Index return Lottery rate IPO UP Holding return Owner proportion P/E Time lag

484

Table 2.

Strategic IPO Underpricing: The Role of Chinese State Ownership

485

This relation is also expected because a large lottery rate indicates that more public investors can get the IPO allocation and consequently the demand on secondary market after IPO will be less, which would cause a less ﬁrst day price jump. The correlation of time lag and IPO underpricing is positive, with a value of 0.717 and signiﬁcant at 1% level, which is consistent with the ﬁndings in literature. The market index is uncorrelated with other variables to certain extend, implying that the IPO pattern in this sample is independent of the market trend. Similar to Aggarwal et al. (2002),7 we used a two-stage least square estimate to test the second implication: greater IPO underpricing leads to higher trading volume, and the large trading volume will further cause a high level of holding period return. We ﬁrst investigated the relationship between trading volume8 and IPO underpricing, and then in the second stage, we investigated the relationship between trading volume and holding period return.9 We followed Loughram and Ritter (2002) and chose 25-day window to analyze the effect and also performed the robustness in 7-day window. The estimation takes the form: Stage 1 : Log ðVolumeÞi ¼ a þ b1 IPO UPi þ b2 OwnerProportioni þ g1 TimeLagi þ g2 Log ðProceedsÞi þ g3 P=E i þ g4 LotteryRatei þ g5 LowestPricei þ g6 DummyMKTi þ g7 DummyYEARi þ i Stage 2 : Holding Returni ¼ j þ y1 LogðVolumeÞi þ l1 VolatilityðVolumeÞi þ l2 IndexReturni þ ui In the ﬁrst stage, the independent variable is the IPO underpricing (IPO UP); and the dependent variables include the natural log of the average trading volume (Log (Volume)) and the proportion of state ownership uncirculating (OwnerProportion). We introduced several variables into the ﬁrst stage analysis to control for offering and market characteristics. Offering characteristics variables include the time lag between issuing day and listing day (Time Lag) and the lottery rate (LotteryRate). Offering size is the natural log of the proceeds (Log (Proceeds)), and quality of offering is proxied as the reference P/E ratio of similar ﬁrms on the stock market, which is selected by SEC before IPO (P/E). We also captured the market factors by measuring the lowest price in 25 days after IPO (LowestPrice), the individual market effect (DummyMKT) and the calendar year effect (DummyYEAR).

486

Table 3.

Investigation of Large IPO Underpricing Leading to Two-Period Wealth Gain. Log (Volume) 25-Day Window

IPO UP Owner proportion Time lag Log proceeds P/E Lottery rate Lowest price Year dummy Market dummy Adj R2

Non-Parsimonious

Parsimonious

Non-Parsimonious

Parsimonious

4.313 (o.0001) 0.256 (o.0001) 0.269 (0.324) 0.001 (0.434) 0.658 (o.0001) 0.006 (0.010) 0.004 (0.868) 0.051 (o.0001) Yes Yes

4.001 (o.0001) 0.243 (o.0001) 0.376 (0.164) 0.00001 (0.983) 0.636 (o.0001) 0.006 (0.008) 0.001 (0.962) 0.052 (o.0001) No No

3.879 (o.0001) 0.289 (o.0001) 0.197 (0.436) 0.001 (0.037) 0.679 (o.0001) 0.003 (0.130) 0.010 (0.679) 0.046 (o.0001) Yes Yes

3.299 (o.0001) 0.2665 (o.0001) 0.315 (0.209) 0.001 (0.221) 0.642 (o.0001) 0.004 (0.042) 0.020 (0.407) 0.047 (o.0001) No No

0.521

0.517

0.544

0.536

YONG WANG AND XIAOTIAN (TINA) ZHANG

Panel A: First Stage Estimates Intercept

7-Day Window

Holding Return 7-Day Window

Non-Parsimonious

Parsimonious

Non-Parsimonious

Parsimonious

0.295 (0.298) 0.049 (0.095) 2.71 (0.135) 0.006 (0.136)

0.294 (0.292) 0.048 (0.096) 2.71 (0.135) 0.006 (0.133)

0.581 (0.002) 0.079 (0.001) 2.36 (0.001) 0.003 (0.250)

0.580 (0.002) 0.079 (0.001) 2.36 (0.001) 0.003 (0.248)

0.003

0.003

0.034

0.034

Panel B: Second Stage Estimates Intercept Log volume Volatility of volume (million) Index return Adj R2

Note: The sample consists of 408 ﬁrms whose IPOs are between January 1998 and December 2002. The ﬁrst stage estimates the relationship between IPO underpricing and trading volume. The second stage estimates the relationship between volume and holding period return. In the ﬁrst stage, the dependent variable is the logarithm of average trading volume (Log Volume); the independent variable is the IPO underpricing (IPO UP); the control variables are as follows: the proportion of state ownership un-circulating (OwnerProportion); time lag between issuing day and listing day; size effect (Log Proceeds); issuing characteristic (Lottery Rate); ﬁrm quality (P/E); stock price pattern (Lowest Price); individual market effect (Dummy MKT, deﬁned as 1 for Shanghai Stock Market and 0 For Shenzhen Stock Market) and calendar year effect (Dummy YEAR, including 2002, 2001, 2000,and 1999). In the second stage, the dependent variable is holding period return (Holding Return), deﬁned as the highest closing price in 25 days following IPO minus ﬁrst closing price divided by the ﬁrst closing price; the independent variable is the logarithm of average trading volume (Log Volume); the control variables include the volatility of trading volume in 25 days after IPO (Volatility of Volume) and the entire stock market trend (Index Return, deﬁned as the Shanghai/Shenzhen stock market index return). pvalues are in parentheses. Estimation Form: Stage 1 :

Strategic IPO Underpricing: The Role of Chinese State Ownership

25-Day Window

Log ðVolumeÞi ¼ a þ b1 IPO UPi þ b2 OwnerProportioni þ g1 TimeLagi þg2 Log ðProceedsÞi þ g3 P=E i þ g4 LotteryRatei þ g5 LowestPricei

Stage 2 :

Holding Returni ¼ j þ y1 LogðVolumeÞi þ l1 VolatilityðVolumeÞi þl2 IndexReturni þ ui

487

þg6 DummyMKTi þ g7 DummyYEARi þ i

488

YONG WANG AND XIAOTIAN (TINA) ZHANG

In the second stage, the dependent variable is the holding period return (Holding Return); and the independent variable is the logarithm of the average trading volume (Log Volume). The control variables include the volatility of trading volume in 25 days after IPO (Volatility of Volume) and the market index return from Shanghai and Shenzhen stock market, respectively (IndexReturn). Panel A of Table 3 reports the results of ﬁrst stage estimate, and panel B provides the results of second stage estimate. The results indicate that the IPO underpricing is signiﬁcantly positively related to the holding period return through the channel of trading volume. These ﬁndings provide evidence consistent with the hypothesis that state owners strategically underprice IPO in order to get a larger wealth gain after IPO. Column 2 reports the coefﬁcient estimate on a 25-day window. In ﬁrst stage estimate, the coefﬁcient of IPO underpricing (0.256, with a p-value of 0.0001) has both statistic and economic signiﬁcance. It implies that a low IPO underpricing predict a high trade volume in secondary markets – a 1% IPO underpricing will lead to a 25.5% increase in trade volume in the secondary markets. In the second stage estimate, results show that volume has a signiﬁcant positive impact on holding period return, with a coefﬁcient of 0.049 and a p-value 0.095. The two main results generally support the hypothesis that a large trading volume leads to a greater highest price, which implies a high holding period return. The coefﬁcient of proportion of state ownership un-circulating is positive but insigniﬁcant. This result can be explained by the fact that the state owners usually own shares of large ﬁrms, and the large ﬁrms would have a large trading volume in secondary market. Most of the control variables are also consistent with our expectation: Offering size (Log(Proceeds)) has a signiﬁcantly positive effect on volume while the variables of time lag, P/E ratio, lottery rate and the lowest price have signiﬁcantly negative coefﬁcients. The coefﬁcients of volatility of volume and index return are insigniﬁcant, indicating that both individual stock behavior and market behavior have no inﬂuence on holding period return. A parsimonious speciﬁcation without variables of market effect and calendar year indicator has also been investigated and similar results are obtained (Column 3 of Table 3). Columns 4 and 5 further show the similar results of non-parsimonious estimate and parsimonious estimate in a 7-day window. We notice that in the second stage, the impact of volume on holding period return is larger and more signiﬁcant, with the coefﬁcient of 0.079 and a p-value 0.001. This ﬁnding reﬂects the fact of a great wealth increase after going public in a short period. In a short period, state owners would be happy to see their wealth growth and would not be upset to leave money on the table.

Strategic IPO Underpricing: The Role of Chinese State Ownership

489

Finally, we test our hypothesis that if majority of shares are not circulating and kept in state owner’s hand, state owner would strategically lower the IPO price and try to gain a larger wealth increase after IPO. Therefore, the higher the ratio of un-circulating shares, the heavier the underpricing would be. We used ordinary least square (OLS) and general least square (GLS) to explore the underlying relationship between the magnitude of underpricing and the ratio of un-circulating shares. The estimation takes the following form and the results are shown in Table 4: IPO UPi ¼ a þ b1 OwnerProportioni þ b2 TimeLagi þ g1 logðProceedsÞi þ g2 IndexReturni þ g3 LotteryRatei þ g4 P=E i þ g5 DummyMKTi þ g6 DummyYEARi þ i The dependent variable, IPO underpricing (IPO UP), is the offer to ﬁrst day return. The independent variables of interest are the proportion of state ownership un-circulating (OwnerProportion) and the time lag between issuing and listing (TimeLag). We also controlled the size of the offering (the logarithm of Proceeds), the characteristic of the offering (the LotteryRate), and the quality of the offering (P/E ratio). We used the market index return from Shanghai and Shenzhen stock market, respectively (IndexReturn), to control for IPO market condition, such as the effects of any hot and cold IPO markets. We also included individual market indicators (Dummy MKT) to control for individual market effect and the calendar year indicators to control for the time effects (Dummy YEAR). Panel A of Table 4 reports the OLS and GLS of the entire sample estimate. The speciﬁcation can be used to discover the relationship between ownership structure and IPO underpricing phenomena. Our argument is that state owners strategically underprice the IPO to maximize their total wealth from two periods; therefore, we expect a positive sign between the proportions of state ownership un-circulating and the level of IPO underpricing. The ﬁndings are consistent with our prediction: In OLS estimate, the coefﬁcient of proportion of state ownership un-circulating 0.569, with a p-value of 0.097, is signiﬁcantly positive, consistent with our hypothesis that state owners strategically underprice IPO to achieve their maximum wealth. It implies that an increase of 1% in the share of state owners is associated with a 57% ﬁrst day price rise for the new issuing stocks. Given the fact that state ownership controls the majority ownership of China’s public listed ﬁrms, our ﬁndings provide some explanation of the large IPO underpricing in Chinese stock markets. The offering proceeds are negatively correlated with IPO underpricing, which can be explained in two ways. On one hand,

490

Panel A: Whole Sample Estimate Intercept

Investigation of Large State Owners Leading to IPO Underpricing. Expected Sign

OLS

GLS

?

13.676 (o.0001) 0.569 (0.097) 0.005 (0.001) 0.714 (o.0001) 0.051 (0.098) 0.032 (0.529) 0.002 (0.627) Yes Yes

13.634 (o.0001) 2.039 (0.001) 0.013 (o.0001) 0.810 (o.0001) 0.020 (0.548) 0.011 (0.847) 0.009 (0.115) Yes Yes

0.327

0.595

Owner proportion

+

Time lag

+

Log proceeds

?

Index return

?

Lottery rate

P/E

+

Market dummy Year dummy Adj R2

? ?

YONG WANG AND XIAOTIAN (TINA) ZHANG

Table 4.

Quartile 1 Panel B: Sub-sample Estimate Intercept Owner proportion Time lag Log proceeds Index return Lottery rate P/E Market dummy Year dummy Adj R2

Quartile 2

Quartile 3

Quartile 4

18.231 (o.0001) 1.879 (0.359) 0.001 (0.665) 0.983 (o.0001) 0.185 (0.014) 0.151 (0.551) 0.037 (0.958) Yes Yes

12.443 (0.008) 0.703 (0.893) 0.002 (0.462) 0.649 (0.001) 0.088 (0.132) 0.079 (0.565) 0.003 (0.776) Yes Yes

10.802 (0.089) 2.240 (0.073) 0.002 (0.562) 0.598 (0.0002) 0.043 (0.449) 0.040 (0.632) 0.003 (0.827) Yes Yes

8.405 (0.003) 3.069 (0.061) 0.016 (o.0001) 0.571 (o.0001) 0.078 (.221) 0.007 (0.926) 0.0003 (0.650) Yes Yes

0.370

0.248

0.254

0.561

Note: The sample consists of 408 ﬁrms whose IPOs are between January 1998 and December 2002. Dependent variable is IPO underpricing (UP); Independent variable is the proportion of state ownership un-circulating(Owner Proportion); the control variables include time lag between issuing and listing (Time Lag), size effect (Log Proceeds), issuing characteristic (Lottery Rate), quality of the offering (P/E); (Dummy MKT, deﬁned as 1 for Shanghai Stock Market and 0 for Shenzhen Stock Market) and calendar year indicators to control for any time effects (Dummy YEAR, including 2002, 2001, 2000 and 1999). Panel A is the entire sample estimate; Panel B is the four sub-samples estimate. The sub-samples are sorted as quartiles by the proportion of state ownership un-circulating. p-values are in parentheses. Estimation takes the form:

Strategic IPO Underpricing: The Role of Chinese State Ownership

The Proportion of State Ownership Un-Circulating

IPO UPi ¼ a þ b1 OwnerProportioni þ b2 TimeLagi þ g1 logðProceedsÞi þ g2 IndexReturni þ g3 LotteryRatei þ g4 P=E i þ g5 DummyMKTi þ g6 DummyYEARi þ i

491

492

YONG WANG AND XIAOTIAN (TINA) ZHANG

keeping the number of issuing shares ﬁxed, a higher IPO price implies a smaller underpricing and will simultaneously generate higher proceeds. On the other hand, keeping IPO price ﬁxed, less shares issued could not meet market demand and would therefore generate heavier underpricing. Consequently, fewer shares issued would also generate fewer proceeds. Chan, Wei, and Wang (2004) also ﬁnd that underpricing is negatively related to the number of shares being issued, which conﬁrms our ﬁnding. Most control variables have the expected signs. Consistent with prior literature, we ﬁnd that the longer the waiting time from the issuing date to the listing date, the larger the magnitude of IPO underpricing. Partial of the IPO underpricing in this case could work as the compensation to investors’ patience in IPO procedure. P/E ratio, as a proxy of ﬁrm’s quality, is found to be positively related to IPO underpricing. One potential explanation is that the prudential investors in secondary market favor the good quality ﬁrms and drive up the price of ﬁrms with high P/E ratios. The variable of lottery rate is found to be negatively related with IPO underpricing, and this ﬁnding is also consistent with our expectation: a small lottery rate means that a large proportion of stock shares have been held by government or state owners and few investors are permitted to purchase the stock in the primary market. In addition to the IPO underpricing by state ownership, limited IPO allocation in the primary market will ignite the purchasing interest in the secondary market , which further causes a great ﬁrst day price jump IPO. We also obtain similar results when we use GLS estimation. Endogeneity is a concern in any corporate ﬁnance research. We argue, however, that the endogenous problem in our study is not severe. Ownership structure is stable and usually predetermined before IPO procedure. It is not possible that the causality runs from the IPO underpricing to state ownership. Another concern is that both state ownership and IPO underpricing are determined by some underlying factors. We control this problem by analyzing four sub-samples classiﬁed by the quartile of the proportion of state ownership un-circulating. If there were an endogenous problem, we would ﬁnd similar relationship between state ownership and IPO underpricing. On the other hand, if the relationship between ownership and IPO underpricing is non-constant, then the endogenous problem is not severe. The results are shown in Table 4, Panel B. The coefﬁcients of the proportion of state ownership un-circulating in the two sub-samples of ownership ranked in quartiles 1 and 2 are insigniﬁcant but positive; but that in quartiles 3 and 4 are both positive and signiﬁcant. This ﬁnding conﬁrms that ownership structure does play a role in the IPO underpricing. In addition, the

Strategic IPO Underpricing: The Role of Chinese State Ownership

493

magnitudes of the coefﬁcients are increasing from quartile 1 to quartile 4. This ascending relationship between state ownership and IPO underpricing shows that the endogeneity would not be a concern and it offers further support to our model of strategic IPO underpricing by state owners. We also get the expected signs of control variables. To sum up, the empirical results are generally consistent with our model in which state owners strategically underprice IPO to achieve the maximum of two-period wealth. When majority of shares are kept un-circulating, state owner would lower IPO price and sacriﬁce IPO proceeds in exchange for a larger wealth gain after IPO. Consistent with our ﬁndings, Chen, Firth and Kim (2004) also ﬁnd that high government and legal entity shareholdings are associated with underpricing in Chinese markets. Although our paper does not offer any prediction about long-term performance after IPO, Chan et al. (2004) offer some evidence that the long run underperformance after IPO dominating the U.S. and other developed markets are much moderate in Chinese stock market. This also lends some support to the after market wealth beneﬁt.

5. CONCLUSION This paper provides an alternative explanation of IPO underpricing in Chinese stock markets: State owners focus on wealth effect and trade off between the IPO proceeds and wealth gain after going public. Given the recent huge amount of capital invested by global investors in China’s IPO market, this study helps foreign investors to better understand the uniqueness of China’s IPO market and the process of SOEs going public. Our model implies that the state owner strategically underprices IPOs to drive a larger trading volume, which further results in a higher price. The higher the maximum price, the more wealth gains for SOE ownership after IPO. The state owner eventually achieves the maximum wealth, which consists of IPO proceeds and wealth increase after going public. In contrast with Mok and Hui (1998), who suggest that higher equity retention by state owner induce less IPO underpricing, we expect a positive relation between the state ownership and the magnitude of IPO underpricing. This paper provides an explanation to IPO underpricing in the scenario of little dispersion in ownership and control, where state owners are the largest shareholder and also controls management of ﬁrms. The analysis differs from previous literatures in that it does not assume any information asymmetry. It avoids the difﬁculty of joint testing market inefﬁciency and IPO

494

YONG WANG AND XIAOTIAN (TINA) ZHANG

underpricing. We don’t regard IPO underpricing as an outcome of or a solving method to agency problem. The paper actually argues that the IPO underpricing is an optimum strategy, leading to the maximum wealth of the majority of existing shareholders. This paper, with caveat, does not explain the magnitude of IPO underpricing. Second, further study should concentrate on the long-term performance of SOEs. Third, the role of foreign investors in the process of SOEs going public needs direct investigation. All of these are open research questions.

NOTES 1. The underpricing of IPO is deﬁned as the difference between the ﬁrst day close price and the IPO price divided by the IPO price. In this paper, the term IPO underpricing and ﬁrst day return are interchangeable. 2. See, for example, Allen and Faulhaber (1989), Grinblatt and Hwang (1989) and Welch (1996). 3. See, for example, Su and Fleisher (1999) and Lowry and Schwert (2004). 4. According to Chinese government report in 2002, the balance of savings deposits rose from US$554.2 billion in 1997 to US$1.05 trillion in 2002, while at the same time, China’s GDP rose from US$891.6 billion to US$1.2 trillion. 5. Sun and Tong (2003) offer some evidence about the effectiveness of SOEs’ going public. 6. For more detail about IPO procedure in Chinese market, please see Aharony, Lee, and Wong (2000). 7. They use a two-stage model to test the information momentum in IPO ﬁrms. 8. We use the average trading volume (unit: millions of RMB) in 25 days following IPO as the proxy of volume in 25-day window and the average trading volume (unit: millions of RMB) in 7 days following IPO as the proxy of volume in 7-day window to perform the robustness test. 9. The proxy of Holding Return is the holding return from the IPO’s ﬁrst day closing price through 25 days following IPO, deﬁned as (the highest closing price in 25 days following IPO minus ﬁrst closing price)/ﬁrst closing price. We use Holding Return in 7 days to perform the robustness test.

ACKNOWLEDGMENT We would like to thank David Reeb and J. Jay Choi for their helpful comments and discussions. Xiaotian (Tina) Zhang acknowledge ﬁnancial support from Temple University CIBER Research Grant. All errors are solely ours.

Strategic IPO Underpricing: The Role of Chinese State Ownership

495

REFERENCES Aggarwal, R. K., Krigman, L., & Womack, K. L. (2002). Strategic IPO underpricing, information momentum, and lockup expiration selling. Journal of Financial Economics, 66, 105–137. Aharony, J., Lee, C.-W., & Wong, T. J. (2000). Financial packaging of IPO ﬁrms in China. Journal of Accounting Research, 38, 103–126. Allen, F., & Faulhaber, G. R. (1989). Signaling by underpricing in the IPO market. Journal of Financial Economics, 23, 303–323. Booth, J. R., & Chua, L. (1996). Ownership dispersion, costly information and IPO underpricing. Journal of Financial Economics, 41, 291–310. Brennan, M. J., & Franks, J. (1995). Underpricing, ownership and control in initial public offerings of equity securities in the UK. CEPR Discussion Paper no. 1211, London. Chan, K., Wei, K. C., & Wang, J. (2004). Under-pricing and long-term performance of IPOs in China. Journal of Corporate Finance, 10, 409–430. Chen, G.-M., Firth, M., & Kim, J. (2004). IPO underpricing in China’s new stock markets. Journal of Multinational Finance Management, 14, 283–302. Grinblatt, M., & Hwang, C. (1989). Signaling and the pricing of new issues. Journal of Finance, 44, 393–420. Ibbotson, R., Ritter, J., & Sindelar, J. (1994). The market’s problems with the pricing of initial public offerings. Journal of Applied Corporate Finance, 7, 66–74. Jenkinson, T., & Ljungqvist, A. (2001). Going public: The theory and evidence on how companies raise equity finance. Oxford: Clarendon Press. Jensen, M., & Meckling, W. (1976). Theory of the ﬁrm: Managerial behavior, agency costs and owners structure. Journal of Financial Economics, 3, 30–360. La Porta, R., Lopez-de-Silance, F., & Shleifer, A. (1999). Corporate ownership around the world. Journal of Finance, 54, 471–517. Loughran, T., & Ritter, J. (2002). Why don’t issuers get upset about leaving money on the table in IPOs? The Review of Financial Studies, 15, 413–443. Lowry, M., & Schwert, W. (2004). Is the IPO pricing process efﬁcient? Journal of Financial Economics, 71, 3–26. Mok, H., & Hui, Y. (1998). Underpricing and aftermarket performance of IPOs in Shanghai, China. Pacific-Basin Finance Journal, 6, 453–474. Su, D., & Fleisher, B. (1999). An empirical investigation of under-pricing in Chinese IPOs. Pacific-Basin Finance Journal, 7, 173–202. Sun, Q., & Tong, W. (2003). China share issue privatization, the extent of its success. Journal of Financial Economics, 70, 183–222. Tian, L. (2001). Government shareholding and the value of China’s modern firms. William Davidson Working Paper no. 395, April 2001. Tian, L. (2003). Financial regulations, investment risks, and determinants of Chinese IPO underpricing. Working Paper, Peking University Management School and London Business School. Welch, I. (1996). Equity offerings following the IPO: Theory and evidence. Journal of Corporate Finance, 2, 227–259. Wall Street Journal (2006). Year-end review of markets & ﬁnance 2005. Wall Street Journal, January 3, R3. Zingales, L. (1995). Insider ownership and the decision to go public. Review of Economic Studies, 62, 425–448.

This page intentionally left blank

MULTINATIONAL VERSUS STATE POWER IN AN ERA OF GLOBALIZATION: THE CASE OF MICROSOFT IN CHINA, 1987–2004 Jean-Marc F. Blanchard ABSTRACT There are two leading paradigms about the power balance between multinational corporations (MNCs) and states. The MNCs in Command approach takes the perspective that MNCs dominate states. The States in Command perspective assumes that states lord over MNCs. Each perspective suffers from noteworthy flaws. I advocate a modified bargaining power (MBP) approach to understanding the relative power of MNCs and states. I test the value of this approach by examining Microsoft’s experience in China between 1987 and 2004. My study shows that that a MBP approach sheds considerable light on the aforementioned case, whereas the two leading paradigms do not.

INTRODUCTION As a result of improvements in telecommunications and transportation, the elimination of barriers to foreign investment, and global economic growth, Value Creation in Multinational Enterprise International Finance Review, Volume 7, 497–534 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07021-X

497

498

JEAN-MARC F. BLANCHARD

multinational corporations (MNCs) have proliferated.1 Whereas there were 7,000 in 1970 and 30,000 in 1990, there were more than 60,000 multinational ﬁrms operating in 2000. Not only have MNCs proliferated, but the largest multinational enterprises have become much larger. Moreover, MNCs are taking a growing share of the world economy and hold a greater percentage of global productive and capital assets ($7.1 trillion as of 2002). Lastly, MNCs now operate in many more countries as well as many more venues within states (Gabel & Bruner, 2003, pp. 2–8; Roach, 2005, pp. 19–44; Cohn, 2006, p. 313). For some, the spread and growth of MNCs constitute a daunting challenge for states. Using their wealth, technology, and prestige, multinational enterprises, or so it is argued, are forcing states to open their borders. In addition, they are driving states to liberalize their economies while concurrently withdrawing from involvement in them. Worse, MNCs are causing states to engage in a race to the bottom in terms of labor and environmental standards. On a more abstract level, this ‘‘Multinationals in Command’’ perspective takes the view that MNCs are making it difﬁcult for states to deﬁne their national interests and are usurping control over a growing number of economic and social realms (Strange, 1996; Mathews, 1997, pp. 50–66; Friedman, 2000; Barber, 2001; Hall & Biersteker, 2002a). Yet this school of thought is challenged by another line of thinking that argues that states maintain considerable power, even in an era of globalization. This camp, which I term the States in Command approach, contends that states possess considerable leeway to determine the institutional framework in which MNCs operate, maintain a considerable ability to tax and regulate MNCs, and still have an important role to play in, for instance, the provision of public goods (Drezner, 1998; Weiss, 1998; Pauly, 2002). Indeed, the strongest versions of the States in Command approach assert that even the smallest states can hold sway over larger multinational ﬁrms (Waltz, 1979, pp. 94–95). Problematically for the Multinational in Command camp, there are numerous cases where multinationals have lost. For example, against the wishes of big energy ﬁrms such as ExxonMobil, Petrobras, and Total, Bolivia, Ecuador, and Venezuela have been increasing the former’s tax burdens, imposing various operating restrictions on them, and seizing their assets (Weitzman, 2006; Webb-Vidal, 2006; Catan & Webb-Vidal, 2006). Worrisomely for the States in Command camp, there are numerous instances where multinationals have won. To illustrate, in the 1970s, MNCs routinely subverted Nigerian policies designed to increase local ownership

Multinational versus State Power: Microsoft in China

499

shares and local representation in management (Biersteker, 1980). More recently, the Japanese automobile ﬁrm Suzuki obtained special privileges such as tax concessions and free land from the Hungarian government because it had the backing of the government of Japan, Hungary’s largest creditor (Bartlett & Seleny, 1998). Not only do the aforementioned schools of thought fare poorly in a number of cases, they also have various conceptual defects. One of the most noteworthy is their assumption that the mere possession of resources endows MNCs or states, as the case may be, with power. In tandem, these empirical and theoretical shortcomings suggest a pressing need for an alternative analytical framework. In this article, I offer such a theory, drawing upon the bargaining power literature. My modiﬁed bargaining power model highlights three factors as playing a decisive role in shaping the power relationship between states and multinationals: (a) the balance of needs; (b) allies/adversaries; and (c) the institutional environment. To demonstrate the explanatory value of my model, I examine the experience of Microsoft in China. This case is an appropriate one because it represents a favorable case for the MNCs in Command and the States in Command approaches. Put differently, it should provide empirical support for both schools – that is, if they are correct. Yet contrary to what the MNCs in Command perspective would lead us to expect, we see that Microsoft did not always get what it wanted. The States in Command perspective, however, does not successfully illuminate the case either. China did not always get what it wanted. In contrast, my modiﬁed bargaining power model fares well. As will be shown, Microsoft got some of what it wanted and China got some of what it wanted. This is what my modiﬁed bargaining power model would lead us to expect, given the constellation of needs and allies as well as the institutional environment. This paper consists of eight sections. In the second section, I review the literature on the power balance between states and multinationals. In the third section, I present my theory of the balance of power between states and multinationals in an era of globalization. In the fourth section, I provide background on Microsoft and the levers that it can wield against host governments. In the ﬁfth section, I offer a brief overview of the People’s Republic of China and the resources it can bring to bear against multinational ﬁrms. In the sixth section, I discuss the factors in my model as they apply to the Microsoft–China case. In the seventh section, I examine the experience of Microsoft in China. In the eighth section, I summarize my ﬁndings, discuss their implications, and identify some future avenues of research.

500

JEAN-MARC F. BLANCHARD

TWO SCHOOLS ON THE POWER OF MNCS VERSUS STATES IN AN ERA OF GLOBALIZATION In international relations, there are two dominant lines of thinking about the power balance between multinational ﬁrms and states. I label one the MNCs in Command approach and the other the States in Command perspective. The former ﬁnds its intellectual antecedents in the dependencia and Marxist schools of thought and is closely tied to the contemporary globalist approach to international relations. The latter owes its roots to the international relations schools of realism and statism. In this section, I review and critique both schools of thought. The MNCs in Command Approach For those in the MNCs in Command camp, a skewed ‘‘balance of resources’’ is the primary reason why MNCs can inﬂuence or dominate states (Fagre & Wells, 1982; Kobrin, 1987; Eden & Molot, 2002). One resource that MNCs have is their ﬁnancial might – that is, their revenues and/or proﬁts. MNC in Command adherents are fond of pointing out that in 1998 the individual revenues of Wal-Mart, Mitsubishi, and General Electric were all higher than the individual GDPs of the Czech Republic, Egypt, and New Zealand (Stopford, 1998, p. 16). More shockingly, perhaps, Domino’s Pizza Corporation alone earned enough revenue in 1991 to fund the combined government expenditures of Bolivia, Iceland, Senegal, and Uganda (Barber, 2001, p. 24). Unfortunately, as former Supreme Court Justice Louis Brandeis observed, while ‘‘businesses may keep growing biggeryhuman beings come in the same size’’ (The Economist, 2002, p. 32). MNC revenues or proﬁts are a potential source of leverage over host countries for a number of reasons. One reason is that they endow MNCs with the ability to upgrade or expand a host country’s capital stocks. More speciﬁcally, they give MNCs the resources to build plants, import advanced manufacturing equipment, and develop transport facilities. Furthermore, large revenues and proﬁts suggest that MNCs may serve as engines of job growth. Finally, they imply the potential of a given multinational ﬁrm to augment a country’s tax revenues. The access to and control over markets that MNCs have also represents a formidable source of power. Access to markets gives MNCs power because they can use it to facilitate or impede the ability of countries to sell their goods to businesses and consumers abroad as well as to import goods.

Multinational versus State Power: Microsoft in China

501

Control over markets is a source of power because it affects the processing, shipping, marketing, distribution, and pricing of materials. In addition to these two sources of power, MNCs have power because of their control of two other resources: managerial and operational expertise and technology/knowledge. In terms of managerial and operational knowledge, multinational ﬁrms typically employ staff who have special skills in functional areas such as advertising, distribution, and production. They also know how to exploit advanced transnational business arrangements such as cross-border licensing ventures. Regarding R&D, MNCs, rather than states, are at the frontier of inventing the aerospace, communications, and biotechnological and pharmaceutical products that will power the economy of the 21st century. A ﬁfth resource that MNCs have is their multimillion dollar public, government relations, and legal budgets. These moneys support the employment of hundreds of lobbyists in key locales such as Washington and Brussels, the molding of public opinion, and the setting of the policy agenda. On a different, but related, note, MNCs frequently put former high-level policy makers and senior bureaucrats on their payrolls to conduct community outreach and to build strategic alliances with other MNCs. These individuals may also provide invaluable access into the policy making and implementation process (Sklair, 2002). MNCs in Command Perspectives on MNCs in an Era of Globalization It is widely accepted among those embracing the MNCs in Command viewpoint that globalization only magniﬁes these sources of corporate power (Strange, 1996; Mathews, 1997; Friedman, 2000; Barber, 2001; Hall & Biersteker, 2002b; Cutler, 2002; Sassen, 2002). For this school of thought, globalization makes MNCs more footloose. Possessing greater mobility, MNCs can easily relocate plants, outsource jobs, reduce their tax burdens, escape regulation, or transfer capital, all at the proverbial push of a button. The greater the mobility of MNCs, the greater will be the ability of MNCs to pressure policy makers to acquiesce to MNC preferences, whether relating to speciﬁc governmental policies or the overall operating environment. If policy makers opt not to concede, then MNCs will leave or not invest in the ﬁrst place. As a result, states will lose their access to corporate capital, technology, jobs, tax revenues, and so on. It is further argued that globalization indirectly strengthens the hand of MNCs by unleashing international capital, whose expansion and mobility globalization has only fueled. The logic here is that if states fail to adopt

502

JEAN-MARC F. BLANCHARD

business-friendly policies then not only will MNCs ﬂee but so will international investors. Thus, policymakers have yet more reasons to capitulate to the pressures of MNC managers. Beyond this, it is asserted that globalization dissolves national loyalties and creates global governance problems that states cannot solve on their own. Thus, it weakens the ability of states to stand up to MNCs. In the ﬁrst instance, it becomes much more difﬁcult for national decision makers to promote their MNCs, protect infant industries, or impose regulations restricting the availability of foreign products and services when their citizens have to have Coca-Cola, Starbucks, or ‘‘their MTV.’’ In the latter instance, the need for governance mechanisms is resulting in increases in the number and extent of private international regimes that give MNCs the ability to set standards, enforce contracts, and regulate other MNCs. The States in Command View Strongly disputing the MNCs in Command perspective are those who fall within what I term the States in Command camp. In the view of the States in Command school, the study of the international political economy largely should be the study of inter-state relations. Indeed, this has been the case since the conclusion of the Peace of Westphalia, which ensconced the notion of territorial sovereignty as the core principle of the modern international system. According to this school of thought, state power is built on ﬁve speciﬁc pillars. First, states control important political, economic, and coercive (military and police) resources, though different states obviously possess varying quantities and qualities of such endowments. These resources are important because states can use them inter alia to subsidize domestic companies, fund innovation, or build infrastructure that privileges local enterprises. Of course, at the extreme, they can use their resources, particularly coercive assets, to threaten MNC staff and management, seize MNC factories and equipment, or prevent MNCs from obtaining the raw materials and labor needed to operate their businesses.2 Second, states provide the public goods, domestic and global institutions, and domestic and international bureaucracies that enable multinational ﬁrms to operate and thrive. Third, states can exert control over the markets in which MNCs operate through producer cartels or other techniques. A powerful example of this is OPEC, which has great power, irrespective of the wishes of the major oil companies, to inﬂuence the price of all petroleumbased products by manipulating the supply of crude oil.

Multinational versus State Power: Microsoft in China

503

Beyond these noteworthy sources of power, states are sovereigns over their territory. Thus, they have the authority to tax MNCs. In addition, they have the authority to regulate their dividend and interest remittances, to force technology transfers, and to require joint ventures. They also have the right to deny MNCs access to the resources, labor, and markets that lie within their boundaries. Moreover, sovereignty gives them the right, within certain limits, to expropriate the assets of MNCs within their territorial conﬁnes. Thus, the principle of territorial sovereignty endows even the smallest states with a potent power resource (Krasner, 1985, p. 195). The last major source of state power over MNCs derives from the fact that international law and international government organizations (IGOs) largely recognize states as the subject of international law and the only rightful members of IGOs. Clearly, this is evolving, but for the most part it is states that bring cases before international judicial and arbitral bodies, set the agenda of IGOs, and vote on resolutions and matters that IGO deliberative bodies are considering. States in Command Perspectives on Globalization Globalization may be the buzzword of the day, but it is an irrelevant one as far as the States in Command school is concerned. While trade and capital ﬂows have grown dramatically, they are not that much larger as a percentage of GDP or other bases than what was witnessed during the period between 1870 and 1914. Further, prices are not equalized across interstate boundaries (Frankel, 2000). Even if globalization merits attention, those championing the States in Command viewpoint strongly question if it has undermined the state. In their view, states retain a considerable ability to intervene in their economies and set their tax, labor, and social policies (Mosley, 2000; Drezner, 2001). In any event, States in Command proponents note that globalization has not made states as footloose as the MNCs in Command camp would have us believe. Firms with large investments in plant and equipment remain relatively immobile. Furthermore, multinational executives believe it is critical that they operate in certain areas – for example, India – in order to proﬁt from these countries’ development, to take advantage of just in time delivery, and to generate the economies of scale needed to fund the costly branding, marketing, and R&D efforts associated with today’s products. Finally, there are so-called ‘‘clusters of excellence’’ in the world – for example, Silicon Valley in California (high-tech) – where MNCs feel an imperative to operate (Stopford, 1998, pp. 12–13; Bartlett & Seleny, 1998).

504

JEAN-MARC F. BLANCHARD

One last point that States in Command adherents make when commenting on globalization is that states still possess the loyalty of their citizens. There is no credible evidence that people are abandoning, in any meaningful way, the state as the locus of national identity, much less conceiving of themselves as citizens of the world or, at the extreme, identifying themselves with IGOs, nongovernmental organizations (NGOs), or MNCs (Norris, 2000). Limits to the Dominant Schools Evaluating the empirical validity of the MNCs in Command camp, one is struck by the number of high-proﬁle cases in which MNCs not only failed to command, but were commanded. In 2004, the French government successfully blocked the planned merger of the Swiss pharmaceutical ﬁrm Novartis with the Franco-German pharmaceutical ﬁrm Aventis, instead forcing Aventis to accept a hostile buyout by the French pharmaceutical ﬁrm Sanoﬁ-Synthelabo (Taipei Times, 2004). In April 2006, the government of Zimbabwe seized the sugar estates of the Anglo American Corporation while announcing that it also planned to take at least a 50% share in all existing and future foreign-owned mining corporations (Chiriga, 2006). Besides its inconsistent ﬁt with the facts, the MNCs in Command approach has three conceptual shortcomings of signiﬁcance. For instance, it problematically assumes that MNC power resources give MNCs power over outcomes. Additionally, it adopts an overly narrow analytical perspective, focusing on the isolated interactions of MNCs and the target government. Finally, it ignores the institutional environment. I discuss each of these issues in sequence. Those trumpeting the might of MNCs believe the mere possession of ﬁnancial, technological, or lobbying assets means power. This belief, though, neglects a basic insight of power analysis that the power afforded by the possession of power resources is a relational phenomenon, one that depends upon inter alia the other side’s power as well as the other sides’s need for those resources.3 A MNC with a market capitalization of $500 million is indeed a large company, but it still may not gain much leverage from this fact if it is bargaining with a country with a $500 billion GDP. Similarly, a multinational in the oil sector may be planning to invest $1 billion in oil ﬁeld development, but that may not provide power if the host country can borrow large sums of money cheaply from international markets (Fagre & Wells, 1982, pp. 15–17). When thinking about the power of multinational ﬁrms, it is important to keep in mind that states often have access to allies ranging from other countries to IGOs to NGOs. Thus, it is not correct in all cases to assume

Multinational versus State Power: Microsoft in China

505

that the interaction of countries and multinational enterprises is analogous to two lone boxers battling each other in a ring. Unfortunately, many MNCs in Command adherents take this position. Compounding this overly narrow perspective is the fact that MNCs in Command proponents tend to assume that while host countries lack allies, MNCs can always count on their home governments to assist them. The historical record, though, suggests such a presumption is unwarranted. During the Cold War, for example, the U.S. government proved unwilling to apply meaningful pressure on Japan to open its markets lest such pressure damage the two countries’ security partnership (Encarnation & Mason, 1990). Writers adopting the MNCs in Command perspective also tend to ignore the salience of the institutional environment or to presume it favors MNCs. The institutional environment is signiﬁcant because it includes various international structures – for example, multilateral treaties, principles, and normative understandings – that proscribe certain kinds of corporate behavior. Moreover, these agreements, principles, and norms limit the ability of MNCs to play states off against one another. They also increase the transparency and predictability of MNC operations, which facilitates state control and regulation. It is true, as those studying private authority have shown, that MNCs may capture international institutions. But it hardly follows that what is true in select cases is true in all areas all the time. Contrary to what MNCs in Command adherents assume, it is not obvious that globalization indeed will breed more muscular multinationals. One reason is that globalization has not necessarily made all MNCs more footloose, as discussed above in the review of the States in Command perspective of globalization. Moreover, MNC triumphalists fail to recognize that many aspects of globalization have the potential to undermine MNCs. After all, globalization diffuses knowledge, expands the pool of multinational ﬁrms, expands the number and size of investment capital pools, and improves the availability of managerial talent. It also implies reduced barriers to entry for foreign ﬁrms (Bergsten, 1974). The States in Command approach suffers from diverse ﬂaws. For one, there are many instances where states have failed in their efforts to control multinational enterprises. For instance, the U.S. government had mixed success in getting both American and foreign ﬁrms to comply with its sanction regimes against the Soviet Union and other countries (Rodman, 1995). The Indonesian government’s difﬁculties in regulating mining companies and its recent decision to defer to ExxonMobil with respect to the production of oil from the jointly owned Cepu oil ﬁeld in East and Central Java also illustrate how states do not always get what they want (McCarthy, 2006).

506

JEAN-MARC F. BLANCHARD

Ironically, the States in Command approach commits many of the same conceptual errors as its competitor, the MNCs in Command school. One is that it assumes that power resources translate into power over outcomes. In actuality, though, it depends. The Czech Republic may offer a market of 10 million potential consumers to Korean conglomerate Samsung, but this is unlikely to sway Samsung’s decision making if it can gain access to the German or Polish markets. Likewise, Kazakhstan’s oil reserves may not endow the Kazakh government with power over MNCs such as BP, ChevronTexaco, or Exxon if they feel they can more easily and proﬁtability exploit oil reserves in other locations such as Canada, West Africa, or the South China Sea. For the States in Command perspective, the balance of power between states and MNCs is also one largely between a state and a multinational. Oftentimes, though, MNCs seek succor from their home governments, other MNCs, and industry associations. Gary Gerefﬁ’s (1978) study of Mexico’s interactions with the American pharmaceutical industry over rights to natural steroids revealed that American government intervention played a decisive role in facilitating the unwanted takeover of the Mexican steroid industry by American pharmaceutical MNCs. Obviously, MNCs will not always ﬁnd allies, but the possibility of such support must always be considered. As States in Command proponents correctly note, states are sovereigns, have (in some cases) tremendous political, military, and coercive assets, and possess exclusive membership in the world’s leading IGOs. Yet states wield this authority and these assets within certain institutional parameters. For example, those states that have subscribed to the WTO General Agreement on Tariffs and Trade (GATT), General Agreement on Services (GATs), and Trade-Related Aspects of Intellectual Property Agreement (TRIPs) are restricted in the measures they can take to establish or nurture domestic banks, distributors, service providers, insurance ﬁrms, and wholesalers. Turning to the issue of globalization and its effects, it seems that the state triumphalists miss some important ramiﬁcations of globalization. First, although the volumes of trade and capital ﬂows do not seem to herald a qualitatively new era, the nature of these ﬂows has been changing in noteworthy ways. Trade in services has become more important; intraﬁrm trade has become more signiﬁcant; and short-term, portfolio capital ﬂows have become more prominent. Second, globalization has spurred the formation of large numbers of corporate partnerships – for example, strategic alliances, production networks, and licensing deals – that have no obvious national home and thus seem outside the traditional jurisdictional

Multinational versus State Power: Microsoft in China

507

boundaries of states. These developments make it hard for states to monitor and regulate multinationals, to block foreign ideas and practices, and to shun policies favorable to international capital (Stopford, 1998; Kobrin, 2002). Third, in theory, the ‘‘democratization’’ of international relations has given MNCs the space to inﬂuence the creation and operation of IGOs, international agreements, and global normative standards.

A MODIFIED BARGAINING POWER PERSPECTIVE The preceding review and critique clearly shows that archetypical versions of the MNCs in Command and States in Command schools are ﬂawed. A serious issue is that neither is congruent with the facts despite the former’s certainty that multinational ﬁrms always triumph and the latter’s conﬁdence that states always win. As shown, these schools also have some shared conceptual limitations. Aside from their generally static view of MNC–state interactions, they take the questionable stance that power resources translate into actual power. Another is that they typically ignore allies (adversaries) outside the immediate bilateral MNC–state relationship. Yet another is that they tend to neglect or downplay the salience of the institutional environment. The critique above reveals, too, that both camps do not do justice to the complexities of globalization. In light of the criticisms enumerated above, a useful theory of the power balance between states and MNCs must contain three broad elements. First, it must incorporate variables that allow analysts to assess if states can translate their power resources into actual power. Second, it must include variables that allow researchers and policy makers to weigh the import of actors outside the immediate bilateral relationship being studied. Third, it must incorporate variables to account for the potential impact of the institutional environment. Beyond this, it should not assume that globalization has any predetermined effect. In recognition of these points, I offer a theory of the power balance between states and multinationals that has three variables: (a) the balance of need; (b) allies; and (c) the institutional environment. In this section, I discuss each of them. The Balance of Need This variable recognizes that resources do not convey power unless there is a need for them. Early on, the distinguished international business scholar

508

JEAN-MARC F. BLANCHARD

Raymond Vernon (1971, pp. 697–698) directed our attention to the balance of need when he noted that countries that were fearful of losing ‘‘some useful outside resource – capital, technology, or markets’’ would tread carefully in imposing regulations on MNCs. As for MNCs, they, too, had to be cognizant of their desirability or the need of host countries for them. Per Vernon, ‘‘enterprises have no special advantages in more mature industries.’’ In a work published a decade later, Nathan Fagre and Louis T. Wells (1982, p. 9) explicitly state, ‘‘the deal eventually struck between foreign investors and host governments reﬂects the need forythe resources offered by the two parties.’’ Need is a function of many different factors. With respect to host countries, it is a function of their indigenous production capabilities, their R&D skills, their possession of or access to capital, their need for foreign exchange, and their ability to create jobs. Other factors that matter include the newness of the technology that the MNC is offering and the rapidity with which the technology becomes obsolescent. Also relevant are the potential of a host country to work with other MNCs and the extent to which a host country will interact in the future with the MNC with which it is bargaining in the present (Moran, 1978, pp. 82–83; Fagre & Wells, 1982, pp. 11–15, 18–19; Kobrin, 1987, pp. 619–622; Tarzi, 2000, pp. 158–160; Eden & Molot, 2002, pp. 364–367). Although it is not always the case, developing countries often ﬁnd themselves in an unfavorable position with respect to the balance of needs. Simply put, they tend not to have the advanced R&D, processing, manufacturing, advertising, and managerial capabilities that would allow them to efﬁciently develop, produce, or brand a wide variety of goods and services as MNCs can. In this vein, Toyota’s advanced automobile design and engineering skills, control of a global supply chain of automobile part production and distribution, ability to generate jobs, well respected brand, and investment capital give it considerable leverage against a country like Nicaragua. Nicaragua badly needs jobs, is in dire need of foreign capital, and lacks the indigenous capability to produce automobiles. Even if Nicaragua could produce cars, it would ﬁnd it highly difﬁcult to gain market share for them, given the oligopolistic nature of the automobile industry. As for MNCs, the need for a given host country depends upon the availability of alternative resource suppliers, their ability to generate revenues and proﬁts internally or through production or sales in countries other than the host country, and their desire to maintain a presence in the region in which the host country is located. Besides these factors, it relates to the need of a MNC to gain a presence in a country that belongs to a regional

Multinational versus State Power: Microsoft in China

509

economic organization in whose territory the MNC wishes to operate. Need may also link to the skill level of workers within the host country. It is important to think broadly when assessing a host country or MNC’s need for the other. For instance, a host country may want a particular MNC’s investment not because it needs the MNC’s investment per se but rather because it wants the ‘‘stamp of approval’’ that such an investment represents. This stamp of approval can signal to other investors that the host country is now a suitable place to invest in or that it might be an opportune time to invest additional sums. Furthermore, it is wise not to equate the ‘‘potential of a host country to work with other MNCs’’ solely to the number of MNCs that exist in an industry. Even in oligopolistic industries, the tendency of MNCs to react defensively to moves by their competitors can increase the options for countries to work with MNCs, though the pool of multinational ﬁrms is small (Bennett & Sharpe, 1979; Bartlett & Seleny, 1998). Allies (Adversaries) Although the tendency seems to be for multinational enterprises and states to try to resolve their issues on their own, there are numerous episodes where each of them has sought allies to strengthen its position against the other. Allies can magnify a MNC’s economic incentives and sanctions by supplementing them with additional economic and even political ones. These political incentives and sanctions might entail, respectively, support for a host country’s inclusion in an international agreement or a reduction in security guarantees. Furthermore, allies can indirectly – that is, through local politicians, bureaucrats, or special interest groups – cause a government to embrace policies that help foreign MNCs (Tarzi, 2000, pp. 162–163). As far as countries are concerned, allies can restrict the ability of multinationals to turn to other markets, natural resources, and other goods. Further, allies can increase the ability of a host country to bargain with, monitor, and impose sanctions on a MNC and to mold the institutional environment in which MNCs operate. A MNC’s obvious potential allies include its home government, other foreign MNCs, and industry associations. Less obviously, they can include host country parts suppliers, distributors, service providers, and workers whose proﬁts and livelihoods are linked to the activities of the MNC. Even less obviously, they can include consumers who may have a preference for the goods produced by a particular foreign multinational as opposed to a domestic company (Kobrin, 1987, pp. 613–614). All of these ‘‘allies’’ came into play in the Mexican government’s dealings with the automobile multinationals

510

JEAN-MARC F. BLANCHARD

in the 1960s and 1970s and severely constrained its ability to impose ownership, production (for example, limits on makes and models), and other restrictions on foreign multinational ﬁrms (Bennett & Sharpe, 1979).4 A host country’s potential allies include other countries, bureaucracies and parliamentarians in those countries, domestic ﬁrms, domestic special interest groups, the general public, and other foreign ﬁrms. These allies can increase a host country’s power vis-a`-vis an individual multinational or groups of multinationals when negotiating an international accord. To illustrate, during the negotiations over the GATS and the Trade-Related Investment Measures (TRIMs) agreements, developing countries successfully opposed MNC pressures for faster liberalization schedules and more liberal principles for foreign investment in general and the ﬁnancial services sector in particular because they were uniﬁed, found allies among European governments, and, in the case of TRIMs, obtained support from the U.S. government (Sell, 2000). IGOs can empower host governments in a number of ways. For instance, they can monitor, impose penalties on, and even force the breakup of MNCs. We see the potent role of IGOs in the case of the European Union’s initiatives against the Coca-Cola Corporation. In 2004, it forced the company to release some of its shelf space to rival products, to stop offering special incentives to retailers who ordered Coca-Cola products, and to end exclusivity deals in shops, bars, and restaurants (Buck, 2004; BBC, 2004). In the contemporary period, NGOs may function as a vital ally for states too. As various researchers have shown, NGOs signiﬁcantly enhanced the ability of states to demand lower-cost products as well as production (compulsory licensing) and import (parallel importation) rights from the leading pharmaceutical multinationals (Blanchard, 2004). Institutional Environment Analysts frequently neglect the institutional environment in their treatments of the power relationship between multinational companies and states.5 Institutions, however, are highly consequential and thus warrant signiﬁcant attention. Institutions are not ‘‘actors’’ and, thus, should not be considered as potential allies (or adversaries). More properly, one should think of institutions as structures of principles – for example, sovereign equality – and operating procedures – for example, one state, one vote. Additionally, institutions are structures of prescriptions and proscriptions. Beyond this, institutions are structures of norms – for example, statements of what states and multinationals should do as opposed to what they are obligated to do

Multinational versus State Power: Microsoft in China

511

and to not do. Lastly, institutions are structures of identity. This is to say that they inform their adherents how they should envision themselves – for example, as Europeans versus Germans. Voting rules can do much to empower states. The one state, one vote principle in many IGOs, for example, provides weaker states with a voice they would not have if they had to depend on their political, military, and economic endowments alone. Similarly, norms that favor state involvement in the economy as opposed to market determination of production and allocation can empower states. Norms that call upon the developed world to practice multilateralism as opposed to unilateralism or great power management speciﬁcally can aid developing countries because they put a moral obligation on developed states to take into account the wishes and capabilities of developing states.6 In a penetrating analysis of Eastern European countries’ interactions with leading automobile multinationals such as Ford, General Motors/Opel, Volkswagen/Audi, and Suzuki, David Bartlett and Anna Seleny (1998) demonstrate the power of institutions. They show that European Union treaty language and the Central European Free Trade Agreement prevented countries such as Hungary from providing special privileges to foreign automobile MNCs. As a consequence, these multinational ﬁrms were not able to leverage their capital, technology, and control over automobile markets to extract preferential treatment from Hungary.

WINDOWS ON MICROSOFT’S POTENTIAL POWER Basic Background Founded in 1975, the Microsoft Corporation has grown to become one of the world’s most prominent, proﬁtable, and, in the view of some, powerful high-tech companies. Microsoft’s chairman, Bill Gates, who also doubles as the world’s richest man, has near-celebrity status as do many other Microsoft executives such as Steven Ballmer, the company’s Chief Executive Ofﬁcer. These executives lunch with prime ministers and Communist Party general secretaries, are feted at the world’s most prominent international conferences, and are widely quoted in the leading business media such as the Financial Times, Fortune, and the Wall Street Journal. Thomas L. Friedman reports that prior to the end of 1995, Chinese President Jiang Zemin had met more with Bill Gates than he had with American President Bill Clinton (Friedman, 2000).

512

JEAN-MARC F. BLANCHARD

Headquartered in Redmond, Washington, Microsoft has more than 100 subsidiaries in Africa, Asia, Europe, the Middle East, North America, and Latin America. The company also operates research centers in China, Ireland, Denmark, India, Israel, and the United Kingdom. Microsoft only manufactures products, though, in two countries: Ireland and Puerto Rico. Similarly, one can ﬁnd operations centers, which undertake licensing, logistics, and operations, only in Ireland and Singapore (Microsoft, n.d.). Microsoft is the largest computer software company in the U.S., more than three times larger, in terms of revenues, than the next largest ﬁrm, Oracle (Fortune, 2006c). According to reports, Microsoft operating system software runs on almost 90% of personal computers (PCs) in use. In addition, the company has a dominant position in PC application software (for example, word processing, spreadsheet, and email) markets as well as Internet browser software (85–90% of the market). Furthermore, it has a large share of the business server operating system (50% of the market) and business applications software markets, an increasing share of the gaming market (its ﬂagship product is the Xbox), and a growing share of networking software. Beyond this, Microsoft’s Hotmail is the world’s largest email program (Thurrott, 2003; Claburn, 2005). Microsoft’s Financial Might What has traditionally ﬁred the imagination of competitors, investors, policy makers, regulators, and Wall Street ﬁrms both within the U.S. and elsewhere, however, is Microsoft’s tremendous growth in revenues and proﬁts, though revenue growth has slowed signiﬁcantly in recent years. Microsoft’s ability to increase revenues as well as proﬁts (even in times of slower revenue growth) has made it the 48th largest American company by revenues, the seventh largest by proﬁts, and the third largest by market capitalization. It is also the eighth most proﬁtable American company in terms of return on revenues and one of the most proﬁtable software companies (Fortune, 2006a, 2006b). Table 1 shows Microsoft’s total revenues and net income before taxes from 1998 to 2005. Microsoft’s Standard Setting and R&D Activities Few see Microsoft as a leader in developing cutting edge technologies. Nevertheless, because of its size and marketing acumen, it is a standard setter in many areas. Most are familiar with the standards it has set in the area of PC operating systems and application software. Fewer, however,

Multinational versus State Power: Microsoft in China

Table 1.

513

Total Revenues and Net Income before Taxes (in Millions of Dollars) (Reuters, 2006, p. 9).

Year

Total Revenues

Net Income

1998 1999 2000 2001 2002 2003 2004 2005

15,262 19,747 22,956 25,296 28,365 32,187 36,835 39,788

7,117 11,891 14,275 11,525 7,875 11,054 12,196 16,628

Table 2.

Research and Development Spending (in Millions of Dollars) (Reuter, 2006, p. 9).

Year

Amount

1998 1999 2000 2001 2002 2003 2004 2005

2,601 2,970 3,772 4,379 6,299 6,595 6,469 6,184

realize that Microsoft is a standard setter with respect to business server software and is becoming a standard setter in the PDA and mobile phone software markets. Beyond this, Microsoft has a multibillion dollar annual R&D budget, which supports research activities pertaining to search, workﬂow management, antispyware, storage, and work collaboration software (Microsoft, 2005). Table 2 tracks Microsoft’s R&D spending from 1998 to 2005. In light of this data, the Multinationals in Command approach would lead us to believe that Microsoft should get its way with host countries. Its money, its technology, its status, its operational and managerial expertise, and its $40 billion hoard of cash and short-term investments (coupled with zero debt) should enable the company to surmount the opposition of even the most reticent host country. Indeed, according to some, the lure of Microsoft is

514

JEAN-MARC F. BLANCHARD

quite powerful, driving otherwise reluctant countries to adopt inter alia highly demanding intellectual property laws (Friedman, 2000, p. 178).

CHINA’S POTENTIAL POWER OVER ITS GATES7 As is well known, China has the world’s largest population. Geographically, it is the world’s fourth largest country. Using the purchasing power parity method, China’s economy was the second biggest in the world in 2005 ($8.18 trillion). Using traditional GDP methods, China’s economy came in as the sixth largest in the world in 2005 ($2.2 trillion). Not only is China’s economic size noteworthy, but the Chinese economy has also been rapidly growing since 1978, 9.3% in 2005 alone. Admittedly, this growth is not evenly distributed, but such rates are still remarkable after all these years. China has also grown as a participant in the global economy. Last year, it exported nearly $800 billion. In terms of other economic/ﬁnancial resources, China holds foreign currency reserves of more than $800 billion. It also has a government budget of $392.1 billion. This budget helps pay the salaries of a large number of government workers. Beyond this, the Chinese government has access to resources because of its control over state-owned enterprises (SOEs), of which there are nearly 200,000. Moreover, some of these SOEs are quite large. Despite their size, many of these SOEs are not doing well. Yet they still provide the Chinese government with resources that they can use to inﬂuence other actors. The aforementioned government budget, coupled with other sources, also funds a 2.3 million strong military. The Chinese military currently possesses or is acquiring nuclear weapons, advanced intermediate-range missiles, extremely capable guided missile destroyers and diesel attack submarines, high-quality ﬁghters, and so on. China is a member of all the world’s leading international organizations, ranging from the WTO to the International Monetary Fund to the World Bank. Importantly, it is a permanent member of the U.N. Security Council and thus has veto rights. China is a participant in leading regional and subregional dialogues in East, Southeast, Central, and South Asia as well. Lastly, it should not be forgotten that China is a sovereign state and thus can grant access to its territory, labor pool, and markets. Given all these resources, the States in Command perspective would predict victory by China. It simply would not make sense to expect multinational ﬁrms to triumph in the face of China’s large and growing economy, its immense market, its large pool of workers, its massive military, and its membership in the world’s most important IGOs.

Multinational versus State Power: Microsoft in China

515

WINDOWS ON THE ACTUAL POWER OF MICROSOFT AND CHINA Microsoft’s Need for China Despite the impressive growth of Microsoft’s revenues and proﬁts, and a promising product lineup, Microsoft long had a need to tap into new sources of income and proﬁts (read ‘‘new markets’’). This is the case for three reasons. First, its sales growth over the past quarter, year, and ﬁve years has been less than what its competitors in its industry have achieved (Reuters, 2006, p. 8). A fundamental cause of this is that Microsoft’s main product market in the U.S. – that is, for operating system and business applications software – is saturated. There are just not that many more growth opportunities left for Microsoft’s operating system and applications software products in the U.S. Second, the company has been facing strong challenges from established competitors like Apple, IBM, Nokia, Oracle, and Sony. Third, Microsoft has been facing intense challenges from ‘‘upstarts’’ on a number of fronts. This includes competing server operating system software (Linux), Internet browsers (Firefox), and search providers (Google) (Microsoft, 2004, 2005). China represented an attractive opportunity to a company like Microsoft that was in search of new markets. Not only was China’s economy growing, but it also offered increasing opportunities for companies to make signiﬁcant proﬁts, despite some rough spots along the way (Asia Times, 2005a, 2005b). Moreover, the market for high-tech goods in China seemed extraordinarily promising. For instance, it was estimated by China Computer World Research that information technology (IT) hardware, software, and service markets would all experience double-digit growth rates between 2003 and 2004, with 18.5% annual compound growth between 2005 and 2009 (United States Commercial Service – American Embassy, Beijing, 2004). Similarly, CCID Consulting estimated that the Chinese PC market would grow 9.9% a year between 2004 and 2009, with PC sales reaching nearly $22 billion (United States Commercial Service – American Embassy, Beijing, 2005). It was also not possible to overlook the fact that China had a growing number of Internet users, which, by 2005, had reached a ﬁgure of 100 million, the second largest population of the so-called netizens in the world (People’s Daily, 2005 g). Aside from looking for new market opportunities, Microsoft had a need for foreign talent to support its product development activities. Microsoft needed to exploit foreign engineers and scientists because it would otherwise

516

JEAN-MARC F. BLANCHARD

miss an opportunity to proﬁt from pockets of R&D excellence around the world that could assist it in creating the innovative products that would help it maintain its market leadership. Microsoft also needed foreign talent because foreign talent was often sensitive to local market conditions in a way that workers in Redmond were not. With regard to China, Richard F. Rashid, Microsoft’s Senior Vice President of Research, stated, ‘‘We felt there was a tremendously deep pool of talent there.’’ He added, ‘‘The more smart people, the more innovation, and the more beneﬁts for companies like Microsoft’’ (Buckley, 2004). Of course, over the longer haul, supporting R&D in China would curry favor with Chinese authorities, something of which Microsoft was undoubtedly aware. On balance, then, Microsoft had a strong ‘‘need’’ for China. It wanted access to China’s market. Moreover, it wanted to tap into the talents of China’s scientists and engineers. This need, in turn, put Microsoft in a somewhat unfavorable position vis-a`-vis the Chinese government because it meant that Microsoft could not ignore, incompletely respond to, or place excessive pressure on the Chinese government. Microsoft’s Allies Yet Microsoft had an extremely powerful ally, the U.S. government, which counteracted the ill effects of Microsoft’s need for China. On behalf of Microsoft, other companies (both within and without the software sector), industry associations, and, in defense of treaties to which China had committed itself, the U.S. government routinely pressured the Chinese government to comply with its obligations. It did this by shaming China (that is, publishing dozens of reports showing China’s noncompliance with its obligations), by collaborating with other countries experiencing similar difﬁculties with China, by exploiting institutionalized dialogues such as the Joint Commission on Commerce and Trade (JCCT) to negotiate with China, and by pressuring China at meetings such as leader summits. As we shall see, this did not guarantee results, but it enabled Microsoft to achieve progress in areas in which it might otherwise have achieved no progress or experienced backsliding.

CHINA’S NEED FOR MICROSOFT Since 1978, but especially after China joined the WTO in 2001, foreign investors – Americans, Europeans, Japanese, overseas Chinese – ﬂocked to

Multinational versus State Power: Microsoft in China

517

China. Initially, they were attracted to China’s large potential market. Later, though, they shifted their attention to operating export-oriented factories that would allow them to proﬁt from China’s low labor and other costs. Subsequently, as China boomed, many returned to their focus on the China market, opting, in line with this, to pursue the development of marketing, service, and distribution capabilities in order to enhance their ability to sell to the China market. Others invested because they saw China as the hub of their regional operations (People’s Daily, 2005b, 2005d). Regardless of their objectives, they kept investing in China. From 2001 through 2005, for instance, China received almost $275 billion in foreign investment in sectors ranging from automobiles to services to warehousing and distribution operations (People’s Daily, 2006d). Generally speaking, China was very receptive to this investment. Chinese policy makers, after all, believed foreign investment would upgrade the country’s capital stock, create jobs, and bring advanced technology and production processes. As foreign capital kept coming, however, they began to raise questions about the degree to which foreigners were assuming control of the economy, the extent to which foreign investment was upgrading China’s capital stock beyond making it the toy factory of the world, and so on. In some measure, the willingness to question the value of foreign capital is attributable to the fact that China had so much of it. Indeed, Chinese leaders over the past three years have given far less time to CEOs (McGregor, 2005). In any event, the watchword for Chinese decision makers became ‘‘quality’’ versus ‘‘quantity’’ foreign capital. Quality foreign capital generally meant capital invested in R&D centers, communications, computer and electronics manufacturing, and the production of transportation equipment (People’s Daily, 2006c). The belief was that such investment versus investments in assembly operations would allow China to climb the value-added ladder, to avoid the trade frictions associated with low value added Chinese exports such as textiles, and to bolster an area that was becoming an increasingly promising source of growth and China’s industrialization (American Chamber of Commerce, 2005; People’s Daily, 2006e). Ironically, then, as surging volumes of foreign investment reduced China’s need for foreign investors like Microsoft, China’s desire for quality foreign investment, particularly in high-tech sectors, was growing. In fact, Chinese Commerce Minister Bo Xilai explicitly highlighted cooperation with foreign ﬁrms and foreign investment in R&D centers as one of six measures that should be taken to promote China’s software industry (People’s Daily, 2005f)!

518

JEAN-MARC F. BLANCHARD

Of course, China’s need for Microsoft would decline if domestic alternatives existed. Viable alternatives, though, have not been available and are not currently available.8 While China touts the purchase of Chinese software by consumers such as the education ministry of Benin and the widespread use of Chinese software by Chinese local governments, it also admits that its education system is not yielding ‘‘qualiﬁed software personnel’’ (People’s Daily, 2005c, 2005e). Furthermore, in remarks on China’s high-tech development, Chinese President Hu Jintao acknowledged that China lacked its own technology and did not have sufﬁciently productive research programs (People’s Daily, 2006a). Moreover, Ministry of Commerce statistics reveal that around 90% of China’s high-tech exports are foreign brands (People’s Daily, 2005a). Outside observers commenting on China’s technological development report that China’s electronics only further evidence the importance of foreign inﬂuences on China’s high-tech production (Mackey, 2005). In 2000, China took steps to create an alternative to Microsoft, one that it continues to pursue today. China speciﬁcally supported the development of Linux, an open-source operating system, which is free, whose code is readily available for examination, and which allows individuals, companies, and bureaucrats of any country to shape it. To aid in the spread and adoption of Linux, China held seminars on Linux, denigrated Microsoft Windows while touting Linux, set up an international Linux standards testing lab, facilitated cooperation with Japan and South Korean to support it, and encouraged the spread of Linux through the use of preferential government procurement contracts with Chinese ﬁrms offering Chinese-language Linux (Goad & Holland, 2000; Bolande, 2000; Dickie & Nakamoto, 2004; People’s Daily, 2006b). As of 2004, Linux had hardly displaced Microsoft, acquiring a 10.5% share of the market, though it presented an increasingly viable alternative (Einhorn & Greene, 2004).9 Interestingly, a group of Chinese software ﬁrms seems to have lobbied the Chinese government to back off from Linux, taking the position that China’s favoritism towards Linux has stunted the development of China’s software sector because it has taken proﬁts out of the market (Asia Times, 2005c). Overall, then, it is safe to say that China had a need for Microsoft that was not diminished by the availability of alternatives. China’s Allies China did indirectly have some allies in its dealings with Microsoft. First, between 2000 and the present, the European Union and subsequently South Korea investigated Microsoft, imposed ﬁnes on it, required it to unbundle some of its software, and, in the former case, demanded that the company

Multinational versus State Power: Microsoft in China

519

share some of its source code with competitors (Fiﬁeld, 2005). Second, countries such as Russia, Japan, and others pressured Microsoft to reveal its source code so that it was clear there were no backdoors that would present national security risks. Third, Brazil, India, the Philippines, and others pressured Microsoft to lower the price of its software or, at a minimum, design more affordable versions of its software (Waters, 2004). China’s Institutional Environment While allies helped China’s case in its interactions with Microsoft, the institutional environment undermined China’s position. Speciﬁcally, China opted as part of its WTO commitments to accept, upon accession, full compliance with TRIPS. This meant it was legally obligated to protect the patents, copyrights, and trademarks of foreign companies like Microsoft and, if it did not, to accept possible WTO-authorized retaliatory sanctions. It is hard to believe that this obligation did not weaken China in its dealings with Microsoft. Summary To summarize the preceding discussion of my three variables as they apply to Microsoft and China, I ﬁnd, as far as Microsoft is concerned, that Microsoft had a need for China that gave China leverage against it. On the other hand, Microsoft had the U.S. government as an ally, which compensated for its unfavorable balance of needs. With respect to China, we observe that it too had a need for Microsoft, which diluted its bargaining power against Microsoft. Even so, the presence of allies helped moderate the weakness resulting from China’s need for Microsoft. Unfortunately for China, the leverage gained from the presence of allies was undermined by the institutional environment. Given the mixed picture that emerges from this analysis, my modiﬁed bargaining power model leads us to expect mixed outcomes rather than a full Microsoft victory (MNCs in Command) or a complete victory by China (States in Command). I now turn to an analysis of the speciﬁcs of the case.

MICROSOFT IN CHINA The Early Days In 1987, Microsoft established a presence in the Greater China region – consisting of Hong Kong, Taiwan, and the People’s Republic of China – in

520

JEAN-MARC F. BLANCHARD

order to take advantage of the region’s surging production of PCs and expanding PC markets. Two years later, Microsoft began distributing its software into China through various intermediaries, and within the next two years Microsoft had stationed its ﬁrst employee in China. Initially, Microsoft was wary about plunging into the China market, given China’s well-documented reputation as a haven for software piracy. Indeed, around this time, Microsoft estimated that nearly 100% of Chinese software was pirated. It took the signing of a special Sino-American agreement on intellectual property rights (IPR) in 1992 to assuage Microsoft’s piracy concerns. Building on this accord, Microsoft concluded a licensing agreement with leading Chinese PC producers such as the Great Wall Computer Group that ensured a position of dominance for Microsoft software on Chinese-manufactured PCs. Soon thereafter, Microsoft opened a Beijing sales ofﬁce. In addition to bolstering sales by strengthening software distribution channels, Microsoft planned to maximize its China sales by establishing authorized Microsoft training centers in various places, by creating more simpliﬁed Chinese-character programs, and by educating Chinese government ofﬁcials at the central and provincial level about IPR rules. In the early 1990s, Microsoft staff indicated a willingness to devote signiﬁcant resources to these endeavors, projecting that the China market would ‘‘become one of the most important software markets in Asia within the next decade’’ and that those who had ‘‘made an early commitmentywill be well positioned for success’’ (Lee, 1993). Unfortunately for Microsoft, its staff seemed unaware of the delicate nature of the Taiwan issue. Thus, they assumed they could use their Taiwan operations to supply a simpliﬁed-character, Chineselanguage version of Windows. But this proved unacceptable to China’s Electronic Industry Ministry. It not only wanted the system to be designed in China but ‘‘also insisted on deﬁning the coding and standards for Chinese character fonts – something Microsoft had done independently everywhere else in the world.’’ Microsoft Changes Course When Gates arrived in China in March 1994 to launch Chinese Windows, he got a frosty reception. Meeting with ofﬁcials, he argued that the marketplace, not governments should set standards. But China threatened to ban Chinese Windows and President Jiang Zemin personally admonished Gates. Gates sacked his China management team and promised to cooperate with Beijing (Engardio, 1996).

Multinational versus State Power: Microsoft in China

521

Two years later, China welcomed Microsoft with open arms. At one gathering at a restaurant in Guangzhou, nearly a thousand local admirers and dignitaries turned out to give a warm welcome to Bill Gates. Despite their concerns about Microsoft’s potential dominance of China’s software markets, Chinese leaders recognized their country’s continuing need to embrace foreign investment. As for Microsoft, it was clear that it had its eye on the huge Chinese market. Although Microsoft was doing well in Japan, South Korea, Hong Kong, and so on, it still had a comparatively low market share in Asia, relative to its position elsewhere (Thornton, 1996). In the background, Microsoft had an ally in its quest to grow market share in China. Speciﬁcally, the U.S. and China were ﬁghting over IPR and China’s modest efforts to protect them. Happy to let the U.S. government carry the heavy burden of transforming China’s IPR protection practices, Microsoft publicly praised the United States Trade Representative (USTR) for obtaining new commitments from the Chinese government to enforce the IPR agreement signed by the U.S. and China in February 1995 (Microsoft, 1996). From Microsoft’s perspective, any efforts to advance the cause of IPR in China would beneﬁt it. However, it did not wait for the Chinese government to fulﬁll its part of the bargain. Speciﬁcally, it elected to spend millions a year to train Chinese technicians and programmers at various universities, academies, and centers in Microsoft software. It also worked with government ministries, local computer markets, and universities to create Windows applications in Chinese, going so far as to share code. Microsoft’s tactic seemed to be one of using both the carrot and the stick. However, it was hard for it to wield the stick well, given that it saw China as ‘‘the world’s biggest emerging software market.’’ In fact, a Hong Kong based director of Microsoft gushed, ‘‘we’re looking at 100% growth every year as far as we can see.’’ Microsoft Intensifies its Courtship of China In November 1998, Microsoft and Chinese decision makers and researchers jointly announced the opening of Microsoft’s second international research lab (Microsoft Research China or MSR China), the company’s ﬁrst research facility in Asia. According to a company press release, Chinese ofﬁcials saw the $80 million basic research lab as helping to ‘‘accelerate China’s IT development pace’’ by supporting research and Chinese ‘‘efforts to educate and maintain talent’’ (Microsoft, 1998a). For Microsoft, the lab would help identify hardware and software solutions for integrating the Chinese language into the PC. Facilitating the integration of Chinese into computing

522

JEAN-MARC F. BLANCHARD

was vital if it was going to exploit the rapidly growing Chinese market to its fullest. Furthermore, the lab would give Microsoft an opportunity to tap into local talent and to ‘‘demonstrate its ongoing commitment to fundamental research in China and to the host country itself’’ (Gelb, 2000). Later, Microsoft again signaled its commitment to China by establishing a major support center in Shanghai (Microsoft, 1998b). Microsoft’s establishment of MSR China and the Shanghai support center undoubtedly had something to do with Microsoft’s hopes for the China market. This is strongly implied by a Microsoft press release that gushed that the Chinese market had been growing at the rate of 30% per year and was expected to continue to rise as the per capita income increased and ownership of the PC became popular. Michael Rawding, Microsoft’s Great China Regional Director, observed, ‘‘China is the most populous country in the world, and it’s becoming an ever more important location for information technologyyIt’s very important for Microsoft to understand and really be in the forefront of what’s happening there’’ (Microsoft, 1998c). In March 1999, Microsoft Chairman and CEO Bill Gates signed a slew of agreements designed to bolster computing markets and Internet usage in China. Most interestingly, he signed agreements with various Chinese ministries such as the People’s Bank of China and subnational units such as Shenzhen. These agreements called for Microsoft to inter alia provide Y2 K support and to aid the Chinese State Economic and Trade Commission create an information center. The agreements called upon the Chinese signatories to protect IPR and to promote legal software. Juliet (Shihong) Wu, the general manager of Microsoft China, touted Gates’s visit as evidence of Microsoft’s commitment ‘‘to working closely with the Chinese government and industry to provide China with the latest and most advanced technology and products to contribute to China’s drive in building up a knowledge economy for the next century’’ (Microsoft, 1999; Lombardo, 1999). Microsoft Faces a Fickle China and Offers More Goods The year 2000 started on a bad note when a news story circulated that China was going to ban Windows 2000, a report that Microsoft and Chinese government ofﬁcials quickly rebutted (BBC, 2000). Regardless, relations between China and Microsoft warmed, perhaps because Microsoft actively lobbied for the approval of Permanent Normal Trade Relations (PNTR) between China and the U.S. as well as China’s admission to the WTO. Microsoft did not lend its support for altruistic reasons. It felt WTO

Multinational versus State Power: Microsoft in China

523

membership for China would increase market access and protect IPR (Microsoft, 2000a). In March, the China–Microsoft relations blossomed somewhat more when Microsoft launched Windows 2000 with simpliﬁed characters and accompanied it with a big ceremony attended by Chinese and American ofﬁcials, the glitterati of the local IT industry, and Chinese computer industry elites (Microsoft, 2000b). Partly as a result of Chinese concerns that Microsoft’s software had secret backdoors that allowed inter alia inﬁltration by American intelligence agencies, the glow of Microsoft’s Windows 2000 Chinese edition quickly wore off (Chowdhury, 2000). Rather than reaching out to the world’s leader in operating system software, the Chinese government turned to undermining it by actively championing Linux. Chen Chong, a Deputy Minister of Information Industries who oversaw the computer industry in China at the time, stated that China’s move would ‘‘break the monopoly of the Windows operating system in the Chinese market.’’ Linux’s miniscule market share in 2000 meant Microsoft could dismiss this challenge. Still, the defection of Liu Bo, Microsoft’s deputy manager for China and Hong Kong, to Linux Red Flag Software and the subsequent resignation of Juliet Wu, who wrote a book attacking her former employer, created some angst for the company. In any event, China did not press too hard about the Windows’s ‘‘back doors’’ issue. It recognized that China was not ready to go at it on its own. Indeed, China still needed foreign technology, capital, managerial expertise, and talent if it was to create a world-class software industry (Associated Press, 2000; Smith, 2000). Acknowledging Chinese software security and other concerns, Microsoft CEO Steve Ballmer met with Chinese Premier Zhu Rongji in September 2000 to discuss Microsoft’s investment and business in China. Zhu reemphasized the importance of the software industry in China’s national economic development and welcomed Microsoft’s plans for further investment in and deeper cooperation with industry in China. Ballmer stated, ‘‘We would like to cooperate with local industry, partners, and government organizations to build up a win–win relationship and to help China embrace the knowledge economy.’’ After the meeting, Microsoft announced a plan to expand its regional engineering center in Shanghai. It also agreed to explore the possibility of establishing a joint venture software company with Chinese partners (Microsoft, 2000c). Although China had made numerous promises to improve its protection of the IPR of foreign companies, it accomplished little. On the contrary, software piracy continued to be a major problem for Microsoft in China. Evidencing this, pirated versions of Microsoft XP in both English and

524

JEAN-MARC F. BLANCHARD

Chinese became available in November 2001 even before Microsoft had ofﬁcially launched the product in China (BBC, 2001a). To deal with the problem, Microsoft signed agreements with four leading Chinese computer makers – Legend, TCL, Tsinghua Tongfeng, and Great Wall, representing perhaps 60% of the market – by which they agreed to preinstall Windows XP on their machines. Microsoft certainly could not retreat from the China market to deal with its piracy problems. At the time, China not only had the world’s third largest personal computer market but was also experiencing quite rapid growth (BBC, 2001b). In October 2001, Microsoft took additional concrete steps to improve its position in China. First, it signed a memorandum of understanding (MOU) with the Shanghai Municipal government, pursuant to which it agreed to promote the development of Shanghai’s software industry by establishing a joint Sino-foreign technical support services company. Second, it decided to expand its Shanghai center into a Global Technical Support Center. Previously, Microsoft had agreed to expand it to cover all of Asia. Third, it committed to train 2,000 software architects for Chinese companies in 10 cities. Not coincidentally, perhaps, the deal came at a time when Microsoft saw a substantial drop in its sales in Asia (Tsang, 2001). In January 2002, Microsoft entered into its ﬁrst joint venture, called Zhongguancun Software, with two Chinese software ﬁrms, Centergate Technology (Centek) and Stone Group. The plan was to have the venture develop enterprise and government applications for mainland and overseas Chinese markets. The deal was only $12 million in total (of which Microsoft took a 30% stake) but ‘‘represented a shift in strategy for Microsoft, which traditionally shuns joint ventures.’’ One beneﬁt of it for Microsoft was that it offered it a way of penetrating the local software market, ‘‘especially while competing in tenders against local companies’’ (BBC, 2002). Zhang Qi, Director General of the Ministry of Information Industry, remarked, ‘‘we appreciate Microsoft’s good will in supporting China’s software companies, and its efforts to help us develop software with its own intellectual property rights.’’ Craig Mundie, Microsoft Senior Vice President and Chief Technology Ofﬁcer, replied, ‘‘Microsoft has long been, and continues to be, committed to partnering and growing with the local IT industry in China’’ (Microsoft, 2002a). In April, Microsoft partnered with Shanghai Alliance Investment Co. to create a $4 million joint venture called Shanghai Wicresoft that would develop and market software for the China market and also handle some outsourcing work (Foo, 2002). Microsoft’s support of China’s domestic software industry likely had something to do with the fact that China was still seen, in 2002, as one of the

Multinational versus State Power: Microsoft in China

525

world’s few growth markets for PCs, with shipments of desktops, notebooks, and servers expected to increase by double-digit rates (Kanellos, 2002a). Moreover, it may have had something to do with the fact the Beijing Municipal Government shunned Microsoft in six software contracts that it awarded in December 2001, including one for 2,000 desktop OS sets (Chai, 2003). On a macro level, the President of Microsoft China, Jun Tang, voiced concern about Microsoft’s revenue growth in the China market. Regardless of its motivations, Microsoft’s gestures appear to have enabled it to give the Chinese government feedback about piracy issues. Furthermore, they laid the groundwork for Microsoft to win some contracts from the Shanghai Municipal Government (Kanellos, 2002b). Interestingly, Microsoft also found it possible, in June 2002, to join the Chinese Software Industry Association, the ﬁrst overseas-funded business to obtain such a status (People’s Daily, 2002). The same month, Microsoft signed an MOU with the State Development & Planning Commission (SDPC), pursuant to which it would invest $750 million over the next three years in education and training, academic and research cooperation, hardware manufacturing outsourcing, software outsourcing, and local software companies. The SDPC and Microsoft further agreed to establish a cooperation committee and a subsidiary working group to discuss software industry issues and implement the MOU. Microsoft also committed to license some MSR China technical development accomplishments to local Chinese enterprises (Microsoft, 2002b). In January 2003, Microsoft offered a signiﬁcant concession to the central Chinese government. It speciﬁcally allowed the Government to see the source code for the Windows operating system, agreeing a year later to let the Chinese see the source code for all Ofﬁce 2003 products (Gao, 2003; Nuttall, 2004).10 The next month, Microsoft and Beijing’s Commission on Science & Technology signed a MOU whereby Microsoft would make a $2.2 million investment in Beijing and both parties would carry out cooperation in e-government, software outsourcing, and training of software professionals. As part of the MOU, Microsoft also announced plans to invest in a ‘‘PC Innovations Lab’’ to support the innovation and business needs (for example, compliance testing) of China’s PC and device manufacturers. Liu Qi, the Party Secretary of Beijing, stated that ‘‘we attach high attention to the cooperation with Microsoft, and we think this kind of cooperation will help enable us to further develop the construction of Beijing’s IT infrastructure and the growth of local software industry’’ (Microsoft, 2003; People’s Daily, 2003). Despite Microsoft’s efforts to build good relations with the Chinese government at the central government and local levels, Microsoft repeatedly

526

JEAN-MARC F. BLANCHARD

found itself the victim of discriminatory treatment in 2003. For instance, in response to the discovery that Shanghai public schools were using pirated Microsoft software, it was decided to replace all the software with a Chinese company’s Kingsoft WPS Ofﬁce 2003 product rather than to obtain legitimate product licenses from Microsoft (CNET Asia Staff, 2003a). Not long after the Shanghai school software episode, it was reported that the Chinese government would require all ministries to purchase Linux-based software. Still Microsoft persisted with initiatives designed to improve its standing in China. For example, it signed a product co-development deal with the China National Computer Software and Technology Service, China’s largest domestic software development and systems integration ﬁrm, and agreed to support the training of developers and IT architects. It also agreed to contribute $10 million to bring IT services to primary schools. Microsoft’s agreements followed an agreement between Sun Microsystems and some Chinese ﬁrms for the distribution of desktop software that would compete with Microsoft’s (CNET Asia Staff, 2003b). On the bright side, Microsoft won, in late 2004, a three-year, $3.6 million contract with the Beijing Municipal government. This deal provoked an outcry from those who felt the Beijing Municipal government should purchase Chinese software as opposed to foreign software and initially led the Beijing Municipal government to cancel the deal (Hallett, 2004; Dickie, 2004; People’s Daily, 2004). Although the Beijing Municipal Government did not revoke its cancellation of the deal, it later quietly purchased substantial quantities of software from Microsoft. One reason was to meet central government requirements that government agencies stop using pirated software (Dickie, 2005a). However, this small victory for Microsoft did not obviate the fact that the Chinese government continued to try to promote Linux and also that it sought to enact legislation that would discriminate against software that was not developed in China (Dickie, 2005b).

CONCLUSION The relationship between states and MNCs in an era of globalization is a crucial research question. While it may not be prima facie clear what the impact of MNCs on states is, it is clear that an era of globalization will entail more MNCs, bigger multinational ﬁrms, and involvement by multinational enterprises in a much wider realm of economic, political, and social activity. Thus, the landscape of the international political economy will be different even if we do not know the exact impact of changes in the terrain.

Multinational versus State Power: Microsoft in China

527

Yet the MNC in Command and States in Command perspectives both believe they know what the new world of states and MNCs will resemble. In the former perspective, multinationals will shape the world the way they want it to be. After all, the multiple endowments of multinationals allow their preferences to triumph over states. In the latter case, states will reshape the world the way they want it to be. After all, the multiple endowments of states allow their preferences to triumph over MNCs. As I have shown, though, both theory and fact suggest that we should not put much faith in these perspectives. Instead, we should embrace a multivariate perspective that focuses on the balance of needs, alliance patterns, and the institutional environment. Phrased succinctly, we should adopt a modiﬁed bargaining power perspective on the MNC–state relationship. To demonstrate the value of my argument, I applied my model to a study of the interactions of Microsoft and China. My expectation was that neither Microsoft nor China would always win. To a great extent my expectations were validated. On the one hand, Microsoft did not succeed in getting the Chinese government to make a meaningful dent in the piracy of its goods. In addition, Microsoft did not get, despite considerable gestures on its part, the Chinese government to level the playing ﬁeld for it vis-a`-vis Chinese software ﬁrms. Finally, Microsoft did not get the Chinese government to back away from its pursuit of Linux and other alternatives to the Windows operating system. As for China, it could not get Microsoft to make any major joint venture investments. Furthermore, it did not succeed in getting Microsoft to offer Windows at a substantially discounted price. Lastly, it did not succeed in creating any meaningful local competitors. This said, it does seem that Microsoft got the worse end of the deal. This suggests that the rough balance between needs, allies, and institutions that I initially posited was incorrect. It is conceivable that the American government, Microsoft’s main ally (and a very powerful one at that), did not press China as hard as it could have. Alternatively, perhaps, Microsoft’s need for China was more signiﬁcant than China’s need for Microsoft. These observations hardly suggest that my model has no explanatory value, but they do suggest that it needs reﬁnement. Going forward, it will be important to subject my model to additional tests – for example, by looking at other MNCs and countries. It also will be necessary to consider the possibility of ﬂuctuations in my variables over the span of time covered in my cases. Lastly, it will be important to marry my qualitative study with quantitative analyses as only the latter can help to illuminate the relative weight, if any, that should be assigned to the three variables in my model.

528

JEAN-MARC F. BLANCHARD

NOTES 1. A standard textbook deﬁnition of a MNC is that it is a corporation that owns and manages operations in two or more countries (Cohn, 2006, p. 315) 2. Bolivia recently sent in the army to take control of all the country’s energy installations – gas and oil ﬁelds, pipelines, and reﬁneries – from the foreign energy ﬁrms operating them (BBC, 2006). 3. The effect of ‘‘need’’ on bargaining power was recognized at least three decades ago by economists and international business scholars. Writing in 1974, Fred Bergsten (1974, p. 138) noted that the bargaining power of host countries vis-a`-vis multinational ﬁrms was bolstered because they had ‘‘a far wider array of options’’ (meaning less need for a given multinational). 4. It is important to recognize that domestic competitors have their own ‘‘needs’’ that may make them an ally of foreign MNCs. For instance, Japanese ﬁrms pressured, in several instances, their government to open to foreign MNCs because they wanted what foreign multinationals had to offer or were worried that foreign MNCs would cause their home governments to close off their markets to Japanese ﬁrms if the Japanese government did not open its markets (Encarnation & Mason, 1990, pp. 34–38, 44–48). 5. Institutions consist of regulations, procedures, codes of conduct, and normative standards. Institutions can apply to a bilateral relationship, a region (for example, the region consisting of ASEAN member states), or the globe. 6. For a fascinating treatment of the role of institutions in shaping the ability of developing countries to achieve their goals with respect to multinationals as well as the civil aviation and maritime sectors, see Krasner (1985). 7. The information below comes from the United States Central Intelligence Agency (n.d.); People’s Republic of China (2005); and United States Department of Defense (2006). 8. Obviously, this can change as a result of actions by the Chinese government and the Chinese private sector. The Chinese government, for instance, is trying to bolster China’s indigenous high-tech capabilities by signiﬁcantly increasing R&D spending (People’s Daily, 2005 h). 9. A more recent article suggests that Linux is making greater inroads, but no statistics were offered (McGregor & Dickie, 2004). 10. While signiﬁcant, this concession should not be exaggerated. One reason is that starting in 2001, Microsoft had already allowed its major clients to see Windows source code. Furthermore, in January 2003, Microsoft started allowing governments to see its source code (Parloff, 2004).

ACKNOWLEDGMENTS I would like to thank the editors, Rodney Bruce Hall, Virginia Hauﬂer, Subin Im, Jing Men, and Dale D. Murphy for their feedback on various drafts of this paper.

Multinational versus State Power: Microsoft in China

529

REFERENCES American Chamber of Commerce. (2005). National Development and Reform Commission Vice Minister Zhang Xiaoqiang’s speech to American Chamber of Commerce Members in Shanghai. AmCham China Brief, 13(4), 21–24. Asia Times Online. (2005a, September 3). AmCham bullish on China. Asia Times Online. Retrieved from http://www.atimes.com/atimes/China/GI03Ad03.html Asia Times Online. (2005b, September 24). China now a key proﬁt source for MNCs. Asia Times Online. Retrieved from http://www.atimes.com/atimes/China_Business/ GI24Cb01.html Asia Times Online. (2005c, October 21). Embattled Linux ﬁghts back. Asia Times Online. Retrieved from http://www.atimes.com/atimes/China_Business/GJ21Cb02.html Associated Press. (2000, May 9). Microsoft faces China challenge. CNET News.com. Retrieved from http://asia.cnet.com/newstech/sg/0,39001173,10030456,00.htm Barber, B. (2001). Jihad Versus McWorld. New York: Ballentine Books. Bartlett, D., & Seleny, A. (1998). The political enforcement of liberalism: Bargaining, institutions, and auto multinationals in Hungary. International Studies Quarterly, 42(2), 319–338. BBC. (2000, January 7). China denies Microsoft ‘‘ban.’’ BBC News. Retrieved from http:// news.bbc.co.uk/1/hi/business/594518.stm BBC. (2001a, November 4). China ‘‘ﬂooded’’ with pirate Windows XP. BBC News. Retrieved from http://news.bbc.co.uk/2/hi/business/1637685.stm BBC. (2001b, December 6). Chinese computers to take Microsoft software. BBC News. Retrieved from http://news.bbc.co.uk/2/hi/asia-paciﬁc/1695085.stm BBC. (2002, January 17). Microsoft signs Chinese JV. BBC News. Retrieved from http:// news.bbc.co.uk/2/hi/business/1765637.stm BBC. (2004, October 19). Coke opens fridge door to rivals. BBC News. Retrieved from http:// news.bbc.co.uk/go/pr/fr/-/2/hi/business/3756984.stm BBC. (2006, May 2). Bolivia gas under state control. BBC News. Retrieved from http:// news.bbc.co.uk/go/pr/fr/-/2/hi/americas/4963348.stm Bennett, D. C., & Sharpe, K. E. (1979). Agenda setting and bargaining power: The Mexican state versus transnational automobile corporations. World Politics, 32(1), 57–89. Bergsten, F. (1974). Coming investment wars? Foreign Affairs, 53(1), 135–152. Biersteker, T. J. (1980). The illusion of state power: Transnational corporations and the neutralization of host-country legislation. Journal of Peace Research, 17(3), 207–221. Blanchard, J. M. F. (2004). Corporate hegemony in remission: The pharmaceutical industry and the HIV/AIDS crisis. Paper presented at the 2004 annual meeting of the International Studies Association (March 17–20). Canada: Montreal. Bolande, H. A. (2000, April 25). China looks to Linux as a way not to get locked into Windows. Wall Street Journal, A23. Buck, T. (2004, October 20). Coca-Cola ends antitrust battle with EU pledge. Financial Times, 15. Buckley, C. (2004, September 13). Let a thousand ideas ﬂower: China is a new hotbed of research. New York Times, http://www.nytimes.com/2004/09/13/technology/ 13china.html Catan, T., & Webb-Vidal, A. (2006, April 11). Venezuela warns international oil companies of more tax increases. Financial Times, 8.

530

JEAN-MARC F. BLANCHARD

Chai, W. (2003, May 12). Asian govts draft Linux into national service. ZDNet Asia. Retrieved from http://www.zdnetasia.com/news/hardware/0,39042972,39129443,00.htm Chiriga, E. (2006, April 28). Mkwasine takeover to scare away investors. Zimbabwe Independent, http://allafrica.com/stories/200604280634.html Chowdhury, N. (2000, April 17). Gates & Co attack Asia. Fortune. 197–202. Claburn, T. (2005, March 18). Firefox eats more Microsoft market share. Information Week. Retrieved from http://www.informationweek.com/story/showArticle.jhtml?articleID=159902316 CNET Asia Staff. (2003a, August 26). Shanghai: School’s out for Microsoft Ofﬁce. CNET News.com. Retrieved from http://news.com.com/2100-1012_3-5068050.html CNET Asia Staff. (2003b, November 21). Microsoft signs. Net deal in China. Zdnet.Com. Retrieved from http://news.zdnet.com/2100-3513_22-5110415.html Cohn, T. (2006). Global political economy: Theory and practice (3rd ed.). New York: Pearson Longman. Cutler, A. C. (2002). Private international regimes and interﬁrm cooperation. In: R. B. Hall & T. J. Biersteker (Eds), The emergence of private authority in global governance (pp. 23–40). Cambridge, UK: Cambridge University Press. Dickie, M. (2004, November 30). Beijing cancels purchase from Microsoft after local protests. Financial Times, 17. Dickie, M. (2005a, February 23). Beijing City buys Microsoft products in spite of widespread local opposition. Financial Times 6. Dickie, M. (2005b, April 25). Beijing takes on Microsoft to win control of desktop. Financial Times, 4. Dickie, M., & Nakamoto, M. (2004, April 3–4). Asian nations to promote Linux as alternative to Microsoft. Financial Times, 4. Drezner, D. (1998). Globalizers of the world, unite. Washington Quarterly, 21(1), 209–225. Drezner, D. (2001). Globalization and policy convergence. International Studies Review, 3(1), 53–78. The Economist. (2002, February 16). The businessman as villain. The Economist, 32. Eden, L., & Molot, M. A. (2002). Insiders, outsiders, and host country bargains. Journal of International Management, 8(4), 359–388. Einhorn, B., & Greene, J. (2004, January 19). Asia is falling in love with Linux. Business Week, 42. Encarnation, D. J., & Mason, M. (1990). Neither MITI nor America: The political economy of capital liberalization in Japan. International Organization, 44(1), 25–54. Engardio, P. (1996, June 24). Microsoft’s long march. Business Week, 52–55. Fagre, N., & Wells, L. T. (1982). Bargaining power of multinationals and host governments. Journal of International Business Studies, 13(2), 9–23. Fiﬁeld, A. (2005, December 8). Microsoft hit with $ 31 m ﬁne by Korean watchdog. Financial Times, 19. Foo, F. (2002, April 11). MS unveils software venture in China. Zdnet News. Retrieved from http://news.zdnet.com/2100-3513_22-880709.html Fortune. (2006a, April 17). The 500 largest U.S. corporations. Fortune, F-1–F-2. Fortune. (2006b, April 17). Biggest moneymakers: How the companies stack up. Fortune, F-30. Fortune. (2006c, April 17). The Fortune 1,000: Ranked within industries. Fortune, F-47. Frankel, J. (2000). Globalization of the economy. In: J. S. Nye & J. D. Donahue (Eds), Governance in a globalizing world (pp. 45–71). Washington, DC: Brookings Institution. Friedman, T. L. (2000). The Lexus and the olive tree. New York: Farrar Straus Giroux.

Multinational versus State Power: Microsoft in China

531

Gabel, M., & Bruner, H. (2003). Global Inc.: An atlas of the multinational corporation. New York: The New Press. Gao, K. (2003, February 28). Gates reveals Windows code to China. CNETAsia. Retrieved from http://zdnet.com.com/2100-1104-990537.html Gelb, C. (2000). Software research in – and for – China. China Business Review, 27(4), 40–43. Gerefﬁ, G. (1978). Drug ﬁrms and dependence in Mexico: The case of the steroid hormone industry. International Organization, 32(1), 237–286. Goad, G. P., & Holland, L. (2000). China joins Linux bandwagon. Far Eastern Economic Review, 163(8), 8–12. Hall, R. B., & Biersteker, T. J. (Eds) (2002a). The emergence of private authority in global governance. Cambridge, UK: Cambridge University Press. Hall, R. B., & Biersteker, T. J. (2002b). The emergence of private authority in the international system. In: R. B. Hall & T. J. Biersteker (Eds), The emergence of private authority in global governance (pp. 3–22). Cambridge, UK: Cambridge University Press. Hallett, T. (2004). Microsoft’s Beijing win raises concerns in China. Zdnet.Com (November 29). Retrieved from http://news.zdnet.com/2100-3513_22-5470208.html Kanellos, M. (2002a, May 31). Connections key for Western ﬁrms in China. CNET News.com. Retrieved from http://news.com.com/2100-1001_3-929671.html Kanellos, M. (2002b, June 5). Microsoft gets diplomatic in China. CNET News.com. Retrieved from http://news.com.com/2100-1001_3-932927.html Kobrin, S. J. (1987). Testing the bargaining hypothesis in the manufacturing sector in developing countries. International Organization, 31(4), 609–638. Kobrin, S. J. (2002). Economic governance in an electronically networked global economy. In: R. B. Hall & T. J. Biersteker (Eds), The emergence of private authority in global governance (pp. 43–75). Cambridge, UK: Cambridge University Press. Krasner, S. D. (1985). Structural conflict: The Third World against global liberalism. Berkeley: University of California. Lee, D. (1993). Project Network: Windows to China. The China Business Review, 20(5), 27–29. Lombardo, H. (1999, March 11). Gates gains inﬂuence by backing China’s govt web project. Internet News. Retrieved from http://www.internetnews.com/bus-news/article.php/79231 Mackey, M. (2005, July 7). Local innovations modest at electronics fair. Asia Times Online. Retrieved from http://atimes.com/atimes/China/GG07Ad02.html Mathews, J. T. (1997). Power shift. Foreign Affairs, 76(1), 50–66. McCarthy, S. (2006, April 17). Exxon plays hardball. Globe and Mail. Retrieved from http:// www.theglobeandmail.com/servlet/story/RTGAM.20060417.wxrexxon17/BNStory/ Business/home McGregor, R. (2005, November 8). Chinese leaders give less time to CEOs. Financial Times, 8. McGregor, R., & Dickie, M. (2004, July 19). Oracle says China opts for Linux. Financial Times, 16. Microsoft Corporation. (n.d.). Fast facts. Retrieved from http://www.microsoft.com/presspass/ insidefacts_ms.mspx#EPCAC Microsoft Corporation. (1996, June 17). Microsoft praises efforts of the USTR in China negotiations. Microsoft PressPass. Retrieved from http://www.microsoft.com/presspass/ press/1996/jun96/china.asp Microsoft Corporation. (1998a, November 4). The Company’s second international research lab outside the United States Highlights Microsoft’s vision and commitment to computer science research efforts in China. Microsoft PressPass. Retrieved from http:// www.microsoft.com/presspass/press/1998/nov98/chinapr.mspx

532

JEAN-MARC F. BLANCHARD

Microsoft Corporation. (1998b). Microsoft Research, China to conduct advanced, long-term research to simplify computer use. Microsoft PressPass (November 5). Retrieved from http://www.microsoft.com/presspass/features/1998/11-5mschina.asp Microsoft Corporation. (1998c, November 5). Microsoft’s investment in China spans the last six years. Microsoft PressPass. Retrieved from http://www.microsoft.com/presspass/ features/1998/11-5mschinalab.asp Microsoft Corporation. (1999, March 10). Bill Gates meets with industry representatives, government institutions on trip to China. Microsoft PressPass. Retrieved from http:// www.microsoft.com/presspass/press/1999/mar99/shenzhenpr.mspx Microsoft Corporation. (2000a, March 6). China, trade, and technology. Retrieved from http:// www.microsoft.com/issues/essays/03-06china.asp Microsoft Corporation. (2000b). Microsoft launches Windows 2000 in China. Microsoft PressPass. Retrieved from http://www.microsoft.com/presspass/features/2000/03-20w2kchina.mspx Microsoft Corporation. (2000c, September 18). Visiting Microsoft CEO Steve Ballmer meets Chinese premier Zhu Rongji & announces further investment in China. Microsoft PressPass. Retrieved from http://www.microsoft.com/presspass/press/2000/sept00/ chinaapr.asp Microsoft Corporation. (2002a, January 16). Microsoft announces its ﬁrst software joint venture in China. Microsoft PressPass. Retrieved from http://www.microsoft.com/presspass/press/2002/jan02/01-16chinapr.asp Microsoft Corporation. (2002b, June 27). The State Development & Planning Commission of the People’s Republic of China & Microsoft Corp. sign a memorandum of understanding to begin the largest Sino-foreign software industry cooperation. Microsoft PressPass. Retrieved from http://www.microsoft.com/presspass/press/2002/jun02/06-27chinapr.mspx Microsoft Corporation. (2003, February 27). Beijing municipal government and Microsoft sign memorandum of understanding. Microsoft PressPass. Retrieved from http:// www.microsoft.com/presspass/press/2003/feb03/02-26beijingmoupr.asp Microsoft Corporation. (2004). Annual report (hard copy and online). http:/www. Microsoft.com/msft/reports/default.mspx Microsoft Corporation. (2005). Annual report (hard copy and online). http:/www. Microsoft.com/msft/reports/default.mspx Moran, T. H. (1978). Multinational corporations and dependency: A dialogue for dependentistas and non-dependentistas. International Organization, 32(1), 79–100. Mosley, L. (2000). Room to move: International ﬁnancial markets and national welfare states. International Organization, 54(4), 737–773. Norris, P. (2000). Global governance and cosmopolitan citizens. In: J. S. Nye & J. D. Donahue (Eds), Governance in a globalizing world (pp. 155–177). Washington, DC: Brookings Institution. Nuttall, C. (2004, September 20). Microsoft opens Ofﬁce code to governments. Financial Times, 17. Parloff, R. (2004, October 18). Redmond’s open-door policy. Fortune, 43–44. Pauly, L. W. (2002). Global ﬁnance, political authority, and the problem of legitimation. In: R. D. Hall & T. J. Biersteker (Eds), The emergence of private authority in global governance (pp. 76–90). Cambridge, UK: Cambridge University Press. People’s Daily. (2002, June 13). Microsoft eyes China’s government procurement deals. People’s Daily. Retrieved from http://english.peopledaily.com.cn/200206/13/eng20020613_97735.shtml

Multinational versus State Power: Microsoft in China

533

People’s Daily. (2003, March 3). Microsoft to further expand investment in Beijing. People’s Daily. Retrieved from http://english.peopledaily.com.cn/200303/03/eng20030303_112598.shtml People’s Daily. (2004, November 26). Microsoft procurement deals arouse concern. People’s Daily Online. Retrieved from http://english.people.com.cn/200411/26/eng20041126_165266.html People’s Daily. (2005a, January 18). China’s foreign trade volume of high-tech products tops US$300 billion. People’s Daily Online. Retrieved from http://english.people.com.cn/ 200501/18/eng20050118_171025.html People’s Daily. (2005b, March 24). New changes of multinationals in China. People’s Daily Online. Retrieved from http://english.people.com.cn/200503/24/eng20050324_178093.html People’s Daily. (2005c, March 28). Chinese software purchased by foreign government. People’s Daily Online. Retrieved from http://english.people.com.cn/200503/28/eng20050328_178416.html People’s Daily. (2005d, May 17). Multinationals shifting China strategies. People’s Daily Online. Retrieved from http://english.people.com.cn/200505/17/eng20050517_185445.html People’s Daily. (2005e, June 15). Domestic software widely used in governmental institutions. People’s Daily Online. Retrieved from http://english.people.com.cn/200506/15/ eng20050615_190291.html People’s Daily. (2005f, June 23). Bo Xilai: China to promote software industry with 6 measures. People’s Daily Online. Retrieved from http://english.people.com.cn//200506/23/ eng20050623_191963.html People’s Daily. (2005 g, July 22). Internet users top 100 mln in China, 2nd largest in world. People’s Daily. Retrieved from http://english.people.com.cn/200507/22/eng20050722_197759.html People’s Daily. (2005 h, December 28). China’s spending on scientiﬁc R & D hits record 24.58 billion US dollars: NBS report. People’s Daily Online. Retrieved from http:// english.people.com.cn/200512/28/eng20051228_231339.html People’s Daily. (2006a, January 9). China’s scientiﬁc level still behind world advanced, Hu. People’s Daily Online. Retrieved from http://english.people.com.cn/200601/09/ eng20060109_233918.html People’s Daily. (2006b January 14). World’s second Linux int’l standard testing lab settles in China. People’s Daily Online. Retrieved from http://english.people.com.cn//200601/14/ eng20060114_235397.html People’s Daily. (2006c, January 16). China uses less, but better foreign capital in 2005. People’s Daily Online. Retrieved from http://english.people.com.cn/200601/16/eng20060116_235862.html People’s Daily. (2006d, January 29). China uses FDI of 274 billion dollars for ﬁve years. People’s Daily Online. Retrieved from http://english.people.com.cn/200601/29/ eng20060129_239186.html People’s Daily. (2006e, February 10). Foreign investors select China as research and development base. People’s Daily Online. Retrieved from http://english.people.com.cn/200602/ 10/eng20060210_241457.html People’s Republic of China. (2005, December 20). China ranks 6th in world economy. Retrieved from http://www.china-embassy.org/eng/gyzg/t227445.htm Reuters. (2006, April 28). Microsoft Corporation. Roach, B. (2005). A primer on multinational corporations. In: A. Chandler & B. Mazlish (Eds), Leviathans: Multinational corporations and the new global history (pp. 19–44). Cambridge, UK: Cambridge University Press. Rodman, K. A. (1995). Sanctions at bay? Hegemonic decline, multinational corporations, and U. S. economic sanctions since the pipeline case. International Organization, 49(1), 105–137.

534

JEAN-MARC F. BLANCHARD

Sassen, S. (2002). The state and globalization. In: R. D. Hall & T. J. Biersteker (Eds), The emergence of private authority in global governance (pp. 91–112). Cambridge, UK: Cambridge University Press. Sell, S. K. (2000). Big business and the new trade agreements: The future of the WTO? In: R. Stubb & G. Underhill (Eds), Political economy and the changing global order (pp. 174–183). New York: Oxford University Press. Sklair, L. (2002). Democracy and the transnational capitalist class. Annals of the American Academy of Political and Social Science, 581(May), 144–157. Smith, C. S. (2000, July 7). Fearing control by Microsoft, China backs the Linux system. New York Times. AI. Stopford, J. (1998). Multinational corporations. Foreign Policy, 113, 12–24. Strange, S. (1996). The retreat of the state: The diffusion of power in the world economy. Cambridge, UK: Cambridge University Press. Taipei Times. (2004, April 27). Aventis and Sanoﬁ end hostilities, agree to merger. Taipei Times. Retrieved from http://www.taipeitimes.com/News/worldbiz/archives/2004/04/27/ 2003138323 Tarzi, S. M. (2000). Third World governments and multinational corporations: Dynamics of host’s bargaining power. In: J. A. Frieden, & D. A. Lake (Eds), International political economy. (4th ed., pp. 156–166). Belmont, CA: Thomson Wadsworth. Thornton, E. (1996). At China’s gates: Microsoft boss conquers a key Asian market. Far Eastern Economic Review, 159(1), 54–55. Thurrott, P. (2003, October 9). OS market share: Microsoft stomps the competition. WindowsITPro. Retrieved from http://www.windowsitpro.com/Articles/Index.cfm?ArticleID= 40481&DisplayTab=Article Tsang, S. (2001, October 19). Microsoft unveils new China initiatives. CNET.com. Retrieved from http://www.zdnetasia.com/news/hardware/0,39042972,38026371,00.htm United States Central Intelligence Agency. (n.d.). CIA world factbook. Retrieved from http:// www.cia.gov/cia/publications/factbook/geos/ch.html United States Commercial Service – American Embassy, Beijing. (2004, May 28). China Commercial Brief, 2, 159. United States Commercial Service – American Embassy, Beijing. (2005, March 25). China Commercial Brief, 2, 173. United States Department of Defense. (2006). The military power of the People’s Republic of China 2005. Washington, DC: Ofﬁce of the Secretary of Defense. Vernon, R. (1971). Multinational business and national economic goals. International Organization, 25(3), 693–705. Waltz, K. N. (1979). Theory of international politics. New York: McGraw-Hill. Waters, R. (2004, September 30). Cut-price Windows trial widened to India. Financial Times, 17. Webb-Vidal, A. (2006, April 4). Venezuela takes over oilﬁelds from Eni and Total. Financial Times, 7. Weiss, L. (1998). The myth of the powerless state. Ithaca, NY: Cornell University Press. Weitzman, H. (2006, March 31). Petrobras halts Bolivia plans after breakdown in gas talks. Financial Times, 4.

INTERNATIONALIZATION AND VALUE CREATION IN THE GLOBAL TEXTILES AND APPAREL INDUSTRY: A COMPARATIVE ANALYSIS OF LITHUANIA AND MOLDOVA Sanford L. Moskowitz ABSTRACT This study examines the internationalization process within the textiles and apparel industry in two countries: Lithuania and Moldova. In particular, this study shows how the evolution by an industry toward greater internationalization is intricately linked to its ability to move up its specific value chain. This analysis compares and contrasts the ability of this industry in a Western European (Lithuania) and a nonaccession Eastern European country (Moldova) to move up the textiles and apparel value chain and so achieve higher levels of internationalization. In examining and relating the relevant factors, this analysis provides insights into – and suggests important modifications to – important concepts and themes such as the stage theory of internationalization, the role of ‘‘inward-outward’’ linkages in the value creation process, the mechanism of the Value Creation in Multinational Enterprise International Finance Review, Volume 7, 535–564 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07022-1

535

536

SANFORD L. MOSKOWITZ

internationalization of small and medium-sized enterprises, and the part played by the European Union in the internationalization (and thus globalization) process.

1. INTRODUCTION Recent studies have shown the close relationship between the process of internationalization and the rate and direction of value creation. More particularly, small and medium-sized enterprises (SMEs), and their role in the international value creation process, are gaining increasing attention in both the academic and the business communities. The decline in trade barriers over the last few decades, the increasing importance for the world’s economies of innovation and new technology, and the general spread of globalization have meant a growing role for SMEs in international markets (Oviatt & McDougall, 1994, 1997). As a part of these studies, investigations examine the rise of SMEs in, and the internationalization of SMEs from, a growing range of countries. As the European Union (EU) has expanded and become more integrated, Central and Eastern European SMEs ﬁnd themselves having greater access to international markets (Roolaht, 2002). Thus, attention increasingly has been shifting to how these so-called transition economies – Hungary, Poland, the Czech Republic, the Baltic States, etc. – can effectively compete internationally with Western Europe and the USA. The rise of entrepreneurial SMEs in Eastern Europe and their ability to generate value added by operating in the global arena is seen as key to the eventual success of these countries. Certain recent theories linking SME development, value creation, and the internationalization process look at the reasons why SMEs have increased the breadth and depth of their exporting activities (Knight, 2000). When barriers to internationalization arise, these investigations tend to focus on internal problems in a country, such as political instability, decaying infrastructure, and so forth. But beyond this more empirical focus are broader theories that advance either incremental (staged) or ‘‘rapid leap’’ models of international expansion and value creation. In the former case, rapid SME internationalization is assumed unlikely because of ﬁnancial constraints, the lack of international market experience or information, cultural friction, and other factors. Thus, the incrementalists hold that SMEs will internationalize and create value for a country in gradual steps, ﬁrst undertaking the least complex forms of entry (e.g., exporting) and then evolving into

International Value Creation in the Textiles and Apparel Industry

537

more complicated types of market entry strategies—from the rather passive forms of contracting work (e.g., becoming part of an outsourcing network) to the more active entry modes of licensing, joint venturing, and merging and acquiring. This study examines the link between value creation and the internationalization of SMEs with respect to two countries: Moldova and Lithuania. This paper then compares and contrasts the internationalization process within one country (Moldova) whose SMEs are ‘‘stuck’’ in the more primitive international forms with another (Lithuania) whose ﬁrms are clearly more successfully breaking free of older, traditional forms of international activity and expanding outward and becoming an integral part of the growing, more integrated European market. These discussions show that the former case reﬂects stagnating or declining value-added capability, while the latter suggests rapid growth in value-added capabilities. In this study, Moldova represents the outlier, nonaccession country, one of the Eastern European countries that has not been – or in the foreseeable future is expected to be – integrated into the EU. In contrast, Lithuania, and the Baltic region in general, has recently become part of the Union. Lithuania in particular is one of the fastest growing countries economically in the northwestern part of Europe. This analysis then allows us to examine the different forces acting on countries that gain value-added beneﬁts from being in the EU vis-a-vis the nonaccession areas of Eastern Europe and assess the different factors that come into play in determining the rate and direction of the internationalization process in 21st century Europe. This analysis centers on a particular industry – the textile and apparel industry – in both countries. This industry is one of the most important for both countries in terms of employment and the percentage of gross domestic product (GDP) and of exports generated. In addition, SMEs generally dominate these industries in each of these countries. Because of low investment cost per operator, the garment industry is often one of the ﬁrst to emerge as a center of SME activity in the transition and increasingly privatized economies. It is in this industry that international activity has become very active since the early 1990s. Indeed, for the textile and apparel industry in both countries, entering into and thriving in international markets are critical requirements for future economic growth. Thus, a study of this industry allows a close examination of the internationalization process within a common SME sector – textiles and apparel – operating in different and highly contrasting countries. This study examines in some detail the dynamics of the internationalization process for the SMEs. In doing so, this paper accepts that the

538

SANFORD L. MOSKOWITZ

‘‘gradual staged’’ paradigm of internationalization pertains to both Moldova and Lithuania. The evidence collected thus far appears to conform to this model. As the evidence shows, textile and apparel ﬁrms in these countries, for various reasons, are at different stages of evolution in the internationalization process. In this study, we explore why these industries and their SME populations are in the stages they are and what factors are inﬂuencing the rapidity and extent of their internationalization. This investigation then will help increase understanding of the spurs and hindrances that exist in such countries when it comes to the internationalization of SMEs and the relationship of this process to value creation. An important aspect of this study is that it adds to the recent area of scholarship that examines the ‘‘inward’’ and ‘‘outward’’ relationships that exist in the internationalization process (Welch & Luostarinen, 1993). In particular, this study suggests that in certain contexts, such linkages can hinder as well as propel outward international expansion. In doing this, this study will broaden our understanding of how the EU can signiﬁcantly inﬂuence – both positively and negatively – the rate and direction of SME internationalization and, in turn, the ability of an industry to create and capture value added, which is essential for future economic growth.

2. METHODOLOGY AND STRUCTURE The research for this paper draws broadly on empirical data from a range of sources. This is fundamentally a qualitative, rather than an econometric, study. Such an approach has been taken as most appropriate for an analysis that is essentially ‘‘comparison and contrast.’’ Such a study requires rich contextual analysis obtainable from a pool of relevant studies and assessments undertaken by the international community, the textile and apparel industry itself, as well as agencies and departments of the individual country governments involved. Recently, international organizations – including the World Bank, the IMF, and the UN – have become interested in analyzing the current and future prospects of Moldova for a number of reasons. It is a small, nonaccession country, and therefore it might possibly be impacted by the recent (2004) addition to the EU of 10 Eastern European countries. These government studies are designed to better understand just where this country stands with respect to the integration of Europe. Moldova also sits in a fairly pivotal position between the West and the East and therefore serves as a unique model of internationalization in general and SME

International Value Creation in the Textiles and Apparel Industry

539

internationalization in particular. The EU itself – through the council – has examined Moldova and its economy in some detail. Similar interest exists for the Baltic countries, especially Lithuania. This is because Lithuania is a fast-growing country, the de facto ﬂagship country of the Baltic region. Then too, over the last few years, the Baltic countries have been preparing to become part of the EU, thus generating much interest – and studies – by the EU and international organizations. Indeed, given that the Baltic countries entered the EU at about the same time as Eastern Europe, comparative analysis of the two regions seems natural. The textile and apparel industry in these countries has received considerable attention, in part because of the industry’s economic importance to the Baltic and Eastern European regions. In addition, the EU’s recent reduction in tariffs on clothing and textiles within the Union as a result of WTO agreements has led to signiﬁcant changes in the global trading environment and so spurred relevant analyses on the role of this industry in the internationalization process. The analysis undertaken in this study involves an examination of the industry and its trends within each country as well as an exploration of selected case study examples. This article ﬁrst looks at the value chain concept in general and as it applies to the textile and apparel industry in general. It then analyzes the industry within Lithuania and then Moldova. In this analysis, Lithuania serves as the paradigm for the ‘‘value creation’’ model. In effect, Lithuania serves as the standard of value creation on which to compare and contrast the very different case of Moldova. The study then analyzes the similarities and differences observed in the two countries and concludes by placing the analysis in the broader perspective of value creation and the internationalization process in general.

3. THE VALUE CHAIN IN THE GLOBAL TEXTILE AND APPAREL INDUSTRY Before proceeding to the analysis of the individual countries, it is essential that we understand the nature of the value chain that exists within the textile and apparel industry in general. The concept of the value chain and its role in competitive strategies is a central concern in attempting to understand the structure and dynamics of Moldova’s clothing and apparel industry (Porter, 1980, 1985). The value chain of the textile and apparel industry is made up of a number of steps, each one creating more value than the step that precedes it. At the lowest end is the so-called ‘‘cut and manufacture’’ (CM). This lowest

540

SANFORD L. MOSKOWITZ

part of the chain is a purely outsourcing activity. In this stage, the customer (the EU company) typically provides design speciﬁcations, furnishes all the materials, handles compliance with norms and standards, and markets and distributes the ﬁnished goods to the end consumer. For example, Steilmann, one of the largest apparel groups in Germany, has a joint venture factory in Moldova producing ladies coats, jackets, and blazers. Production, marketing, and distribution are planned and controlled by Steilmann’s corporate ofﬁces in Germany. The only skills and technology required of the producer – generally an SME – who operates in this early stage are manufacturing and production machinery. The most important comparative advantage a country in this stage can have is cheap labor. A ﬁrm ‘‘stuck’’ in the ﬁrst stage can, on average, claim no more than 15–25% of value input per garment. A contracted ﬁrm might do a little better in terms of value added in the next stage, known as ‘‘cut, manufacture, and trim’’ (CMT). Everything that applies to the CM stage applies for CMT as well, except that the contracted ﬁrm has some responsibility in the purchase of secondary material (e.g., trim). But for both CM and CMT, the contracted ﬁrm is mostly cut off from the ﬁnal market – located in the EU, for example – and at the mercy of the middleman customer that is looking for the lowest price. The CM and CMT phases in the value chain do not require high management skills, and ﬁrms stuck in this lowest portion of the value chain establish few contacts with the West. In general, for these ﬁrms, the only means of competition is the price and speed to market if geographically well-placed. CM and CMT create the lowest level of loyalty on the customer side (larger EU ﬁrm), as it is relatively easy to move production to another location if more favorable terms come up. At the same time, the SME ﬁrms operating at these levels can stay aﬂoat as subcontractors and thus help create jobs. In addition, these ﬁrms may beneﬁt from a certain amount of foreign direct investment (FDI) from the EU as customers attempt to upgrade subcontractor facilities in order to help them maintain quality, costs, and production schedules. A subcontracting ﬁrm can capture a signiﬁcantly greater percentage of value added if it can move from the CMT to the ‘‘full price, full package’’ (FOB) level. To operate at the FOB level, the supplier must at least have the know-how, skills, and ﬁnancial means for sourcing materials internationally. For this, direct contacts with the weaving mills and continuous visits to fabric fairs are required to be up-to-date in fashion information. The FOB producer must be able to ﬁnance material purchases. Because FOB suppliers offer more value to their customers, customers become more dependent on them. Further, FOBs can approach the retail customers directly and achieve higher margins by bypassing wholesalers. Firms that reach this level can

International Value Creation in the Textiles and Apparel Industry

541

capture up to 50% (or more) of total value added. In addition to production machinery, they require Internet access in order to contact and communicate with suppliers. In some cases, design and CAD capability is also required, especially for the larger and more diversiﬁed FOB producers. With ‘‘private label’’ (PL) capability, the producer looks for and secures materials, designs collections, and presents customers with product development ideas. However, the producer who has reached this level still remains a de facto subcontractor, albeit on almost an equal level with the customer. This is so because the ﬁnal decisions on design and products are still made by the customer, and products are supplied under the customer’s label (or brand). This level is the highest level of value adding before ‘‘own label’’ (OL) sales, in which the producer designs collections and sells them under its own labels and brands. It is only at these upper stages of the value chain that the producer can hope to create and retain signiﬁcant proﬁts. In these segments, brand equity value accounts for an enormous portion of value added. As the producer moves up the chain into these realms, it retains an increasing percentage of this value. Table 1 shows the stages toward increased value-added production and the percentages of total value added captured in each stage. As suggested by the above descriptions of the various levels, as a producer ascends to higher levels along the ‘‘value-added stairway,’’ it takes greater control of the production and distribution functions. To do so, cheap labor becomes less of a comparative advantage. Rather, it must expand the breadth and depth of its skill base and technological capability. It must also incorporate greater strategic management, logistical (e.g., just-in-time), design, and marketing/ distribution capabilities. Then, EU producers play a less dominant role in the contractor–subcontractor (usually joint venture) relationships. This is so because as a ﬁrm takes on a greater share of the production, marketing, and distribution functions, it relies less on its European partner to handle these matters and becomes more dependent on its own resources and those offered within its country. In this case, the costs of transportation, logistics, and related matters increase as the ﬁrm takes on more of the responsibilities of the production cycle (World Bank, 2004).

4. LITHUANIA: THE TRADITIONAL MODEL OF INTERNATIONALIZATION AND VALUE CREATION The Baltic States are an example of countries whose clothing and apparel industries have been gradually moving up the value chain. As countries with

542

Table 1. Stage

I

Value-Added Stairway of the International Garment Trade. Meaning

CM

Cut and make

Skills Needed

Manufacturing

Technology Needed

Control over Process (% value added) 15–20

Production machinery, Internet

20–25

Manufacturer sells cutting and manufacturing services only; temporary imports of all materials, which are owned by the customer; manufacturer buys no materials (which are all supplied by customer) II

CMT

Cut, make, and trim For CMT, same as for CM, except that manufacturer buys some of the accessories and trimmings, such as buttons, threads, etc.

SANFORD L. MOSKOWITZ

Production machinery

FOB

Full price, full package Manufacturer buys all materials according to the customers speciﬁcations and styling and at delivery, invoices the full value of the product

Manufacturing, materials sourcing, pattern design

Production machinery, pattern design CAD, Internet

25–50

IV

PL

Private label Manufacturer designs collections independently or jointly with the customer; the full value products are delivered under customer’s trademark

Manufacturing, materials sourcing, pattern design, garment design

Production machinery, pattern design CAD, Internet, garment design CAD

50–75

V

OL

Own label

Manufacturing, materials sourcing, pattern design, garment design, marketing

Production machinery, pattern design CAD, Internet, garment design CAD, marketing

75–100

Source: Gerefﬁ and Memedovic (2003), Nordas (2004).

International Value Creation in the Textiles and Apparel Industry

III

543

544

SANFORD L. MOSKOWITZ

growing economies, their SMEs are, on average, somewhat larger and therefore possess more resources than their counterparts in a number of Eastern European countries. This is true in particular of those small and medium-sized ﬁrms within the textile and clothing industry. Lithuania is located on the eastern coast of the Baltic Sea. The textile and apparel industry traditionally has been a critical sector of the country. It is the second biggest external supplier of textiles (after Estonia) among the Baltic States. However, Lithuania is the largest supplier of clothing by far and thus the biggest producer of clothing and textiles combined within the Baltic region. As in many Eastern European countries, Lithuania’s local market is far too small to be able to absorb the production of its textile and apparel industry. Accordingly, access to export markets, especially within the EU, is critical for the success of this industry. As was true for a number of European countries, Lithuania’s textiles and clothing industry suffered from a fairly rapid decline in the early 1990s. This decline was due to the economic-ﬁnancial crisis in the Russian market, a once important market for Lithuania’s industries. This crisis made it clear that Lithuania had to radically reorient its trade, especially in textiles and clothing, toward Western markets. Lithuania’s proximity to the EU stood it in good stead to establish trade links with a number of EU countries, especially Germany, the UK, France, and the Scandinavian countries. Since 1997, we have seen steady growth in the industry, with respect to both its domestic and its foreign markets. In 2003, Lithuania ranked second only to Portugal among the EU countries in terms of the number of employees in the industry per 10,000 inhabitants (for Lithuania, this number is 141 compared with 251 for Portugal). Between 1997 and 2003, the domestic market for textiles and clothing increased nearly 50%, representing an average annual growth of 7.2%. Production of textiles and sewn garments represents 14% of Lithuania’s GDP. Further, textiles and apparel are Lithuania’s leading export products, with the industry accounting for 15% of total exports, with 85% of textiles, 92% of apparel, and 72% of leather products exported, mostly to the EU. (EKT Group, 2004.) The question remains as to the nature of that growth and trade. A large number of Lithuanian companies work on a CM and CMT basis. This means that many Lithuanian textile and apparel companies are little more than junior members of joint venture arrangements with European companies. The latter supply all (or most) of the materials that are imported into Lithuania for processing. These European ﬁrms moreover control marketing, distribution, and branding. These smaller Lithuanian textile and

International Value Creation in the Textiles and Apparel Industry

545

apparel ﬁrms perform essentially tolling functions. In these cases, value added in the process is lost to the Lithuanian industry as it is generated by the larger Western apparel ﬁrms. For these companies, especially the smaller ones, proﬁt margins remain small and little money is created that can be rolled back into the company to modernize equipment and plant and, in turn, allow the company to be competitive in international – and especially Western – markets. Nevertheless, changes have occurred in Lithuania over the last few years that appear to break this pattern. An increasing number of textile and apparel SMEs have been growing in size (in terms of number of employees, for example) to the point where they can no longer be considered within the SME category. In some cases, this growth has been dramatic. Even more, these ﬁrms have become increasingly competitive internationally, in large part through the re-equipping of their factories. These ﬁrms are characterized by their evolution up the process value chain. We note then that a growing number of Lithuania’s textile and apparel ﬁrms are rising above the lower CM and CMT stages of production. Indeed, we see that several ﬁrms have begun to employ designers and start their own collections, as well as offer FOB and PL services to foreign customers. This country are host to a swelling number of clothing and apparel joint ventures and fully foreignowned ﬁrms. These, in turn, bring in further know-how and trigger new and well-organized spin-off companies, leading to an expanding capability in design, marketing, ﬁnance, and merchandizing. Many large companies are integrating forward, with retail shops in large cities. Several companies operate their own chains of clothing stores in Lithuania and neighboring countries. The company Utenos Trikotazas offers one case example. In 2005, this company stood as Lithuania’s second largest textile and clothing manufacturer and its leading knitwear producer. Privatized in 1993, the ﬁrm has sustained growth over the last decade. The company, which now operates independently of any joint venture arrangement, employs over 1,000 persons. Its products are sold to both the retail and the wholesale markets (Anson, 2002). The company is highly integrated. Its production processes consist of knitting, ﬁnishing, dying, cutting, printing, sewing, embroidery, quality control, and packaging. All processes take place under one roof. Utenos uses state-of-the-art machinery and processes. In 2001, the company became the ﬁrst Lithuanian textile company to obtain an ISO 14001 certiﬁcate for environmental management, which raised its competitive position in Western European markets. Approximately 90% of its output is exported, mostly to EU countries such as Sweden, Spain, Germany,

546

SANFORD L. MOSKOWITZ

Denmark, Finland, Switzerland, and the UK. Utenos also manages a network of 20 shops throughout Lithuania, Latvia, and the Ukraine and has retail and wholesale subsidiaries. Utenos has begun to brand its own line of clothes and will be setting up retail shops in Western European countries within the next few years (EKT Group, 2004). Lithuanian internal conditions and policies clearly played a role in the rise of Lithuania’s textile and apparel industry up the value chain, as illustrated by ﬁrms such as Utenos: cheap, yet educated, labor; improved transportation and logistics support; a favorable tax structure for the industry; an enlightened privatization policy; internal sources of investment; low labor and energy costs, and so forth. But these internal conditions existed in Lithuania prior to the rise of the industry as an expansive force in the 1990s. (They also existed in other countries in Eastern Europe that did not witness a similar evolution in their textile and apparel industry.) What did change was that through the 1990s, the EU expanded and became more integrated (such as through adoption of the euro in 1999). Then too, during this period, Lithuania – along with its Baltic neighbors – began preparing for its eventual entry into the Union (in 2004). This meant that even before its actual entry into the EU, Lithuania had become gradually integrated into the Union on a number of fronts, including trade. This integration was crucial for Lithuanian ﬁrms in expanding prior relationships with EU countries as junior members of CM- and CMT-based joint ventures into more equal partnerships by which they moved gradually, but decidedly, up the value creation chain. Thus, Lithuania shows some of the ways in which integration into the EU apparel and textile production expanded into wider production networks and, therefore, the implications of these trajectories for our understanding of industrial upgrading (Anson, 2002; EKT Group, 2004). For example, by 1998, six years before actual entry into the Union, the EU liberalized its trade with Lithuania, providing the country free access to EU’s preference system and therefore competitive advantage in European markets over non-EU accession countries. And with the liberalization of free trade came increased volumes of FDI from EU countries. No doubt, internal conditions in Lithuania made such investments a good bet, but the trigger for this investment was Lithuania’s (and the other Baltic States’) anticipated entry into a growing and more integrated EU, and the advantages it could be expected to obtain therefrom. For example, investors were attracted to Lithuania because they saw Lithuanian ﬁrms – now gaining in stature from their position as near colleagues within the EU – as powerful allies (and not just junior partners) in their attempts to launch activities

International Value Creation in the Textiles and Apparel Industry

547

throughout a region. Important EU clothing and textile groups such as Marzotto from Italy, Tuch Fabrik Wilhelm Becker from Germany, Chargeurs from France, and Tolaram from Singapore joined forces with the larger and growing Lithuanian textile and apparel ﬁrms to explore new market arenas in Eastern Europe. The bulk of the investments and subsidies that helped modernize and refurbish Lithuania’s better-run textile and apparel plants came externally from these EU countries. In 2003, while Lithuania’s domestic capital sources invested only 30 million euros into the industry, foreign sources provided four times this amount – 120 million euros. As of January 2003, More than 900 foreign companies have invested in joint ventures in Lithuania’s textile industry. The major investing countries were Germany, the UK, and Denmark. Subsidies and investments became increasingly linked with integration into the EU system of product standards, thus furthering Lithuania’s attractiveness as a growing center of textiles and high-end clothing. Located in an accession country preparing for entry into the EU, companies like Uteanoa Trikotazas in the late 1990s could work with the European Bank for Reconstruction and Development for funding and ﬁnancial help and guidance in meeting European-wide and ISO standards. Being in accord with the production and market standards of other European companies put Lithuanian’s ﬁrms on an equal footing competitively with their European counterparts and served as an incentive for the formation of integrated, full-service companies competing within both Western and Eastern European markets.1 Then too, the close collaboration and interaction with EU ﬁrms and organizations has resulted in knowledge transfer, thus allowing Lithuanian companies to expand their capabilities in production, marketing, and distribution. The larger Lithuanian companies have established relations with the best Italian and Swiss equipment manufacturers, which further enhances the companies’ ability to meet EU environmental and quality standards. On the marketing side, Lithuanian companies increasingly collaborate with designers from Italy and France and have the opportunity to exhibit their products at European fabric fairs in Brussels, Germany, and other countries on the fair circuit. In general, Lithuanian ﬁrms are increasingly using the experience and knowledge of style, patterns, fabric selection, and production methods developed from export contacts to increase their in-house capacity to develop own-design manufacturing (ODM) and, less commonly, ownbrand manufacturing. This is resulting in close association with EU retailers and catalog ﬁrms and the establishment of in-house design capability and

548

SANFORD L. MOSKOWITZ

the selling of products with their own brand name – Audimas, Utenos Trikotazas (UT), Lelija, 3S, Nijole – in EU fashion houses and outlets (EKT Group, 2004).

5. MOLDOVA: THE NONTRADITIONAL MODEL – BARRIERS TO INTERNATIONALIZATION AND VALUE CREATION As with Lithuania, the textile and apparel industry is an important source of economic activity in the Eastern European country of Moldova. Apparel and clothing account for 10% of all employment and 15% of all of Moldova’s exports. As with Lithuania, the bulk of these exports go the countries in the EU. Nevertheless, the case of Moldova presents a very different situation from what we see in Lithuania and the Baltic countries in general. In contrast to Lithuania, Moldavian textile and apparel ﬁrms – for the most part SMEs – have not made the move up the value chain. Moldova is part of a group of countries that are considered as making slow progress in their development of an entrepreneur-based economy (IMF, 2005). This is true of the textile and apparel industry itself. Under the Soviet system, Moldova’s garment factories had annual production quotas to fulﬁll. Materials such as yarn and fabric were assigned to them automatically according to plan. The factories would take the materials, produce garments, and then ship the ﬁnal product back to Russia. Prior to 1989, Moldova’s textile and clothing industry traded with the Soviet Union and its satellites. The Soviet central government essentially subcontracted CM and CMT work to Moldavian ﬁrms. After 1989, with the collapse of the Russian economy, the industry attempted to turn westward to the EU. This redirection westward succeeded, but only up to a point. European clothing ﬁrms increasingly turned to Moldova as a subcontractor. Moldova’s cheap labor supply was only one of the reasons for this realignment of trade eastward. EU regulations themselves were instrumental in inducing the creation of this system. Indeed, it is believed that this West to East outsourcing or ‘‘outward processing trafﬁc’’ (OPT) is the main driving force for the formation of the subcontractor relationships that developed between the EU and Moldova since the 1990s. OPT is a special EU regulation allowing preferential customs treatment to EU companies that outsource apparel production through the providing of materials of EU origin. In this case, import duty by EU companies is paid only on the value added at reentry to the EU. In this system, EU customers supply materials – yarn

International Value Creation in the Textiles and Apparel Industry

549

and fabrics – to Moldova-based companies, which then process the materials in the form of sewing and packing. In 2001, more than 80% of textile materials were imported from EU customers for processing by Moldova, with 75% of apparel exports heading back to the EU (Italy, Germany, the Netherlands, and Belgium). The period from 1998 to 2003 saw 25% annual growth in the export of apparel to the EU. This CM activity on the part of Moldavian ﬁrms is essentially a tolling function that creates little value added in Moldova itself. Such tolling activity generates little income for the ﬁrms and helps exacerbate Moldova’s trade deﬁcit situation. Regional studies of the Balti, Soroca, and Chisnau regions of Moldova show that the value added that is captured in these areas varies from 2% to 15%. Though the incentive to retain greater value added, and therefore proﬁts, exists in Moldova, as it does in Lithuania, there is little evidence that this is happening. Some of the larger apparel, carpet, and leather ﬁrms, such as Zorile, Ionel, Floare, and Coavare, have attempted to integrate forward by setting up retail stores at their factory sites as well as ofﬁces in, for example, downtown Chisneau to sell apparel, shoes, and carpets that come directly from the factory. But these products are of poor design, of low quality, and nonbranded, and have few buyers. Virtually no links have been made between these ﬁrms and shops and the clothing and fashion industry in the EU countries. Overall, Moldova’s apparel and clothing ﬁrms sell 93% of their capacity on a CM basis and less than 2% on full price subcontracting or under their own label. None of Moldova’s apparel companies sell PL services. With such low value added permeating the industry, very little income is generated within Moldavian apparel ﬁrms. Consequently, these SME ﬁrms cannot by themselves generate sufﬁcient cash to modernize plants and equipment, such as advanced machinery and computerized pattern design. For Moldova’s clothing and apparel industry, the next important step in its internationalization is to move up the value chain in the production process – beyond the simple and passive CM operation – so that the industry creates and captures a greater proportion of the total value added in the making and selling of clothing products. In attempting to locate possible causes for the restraints on Moldova’s apparel and clothing industry, attention is often drawn to internal problems, including weaknesses in – and the high costs associated with – transportation and logistics networks, customs and bureaucratic inefﬁciencies and costs, underdeveloped ﬁnancial institutions, and destructive taxation policies (in particular, value-added tax on cutting wastes). In addition, factories and equipment and machinery are old and in poor condition. There are few modern ground ﬂoor facilities. Rather, buildings are of the older multiﬂoor

550

SANFORD L. MOSKOWITZ

type, therefore rendering internal transport of materials difﬁcult and timeconsuming. There are only a few computerized product design and cutting facilities. Certainly, these problems compromise Moldova’s capability to compete internationally in the apparel and clothing market. Moldova’s advantages – low-cost labor and geographical closeness – become compromised due to numerous trade barriers created by the Moldavian government. For example, customers cannot exploit the same speed to market advantages as in EU and other CIS countries. This is important for fashion products, where speed to market is critical.2 However, just as it is not clear that internal conditions are fundamental to the evolution of Lithuania’s textile and apparel ﬁrms, so the assumption that Moldova’s internal problems is the decisive factor in the observed stagnation of that country’s clothing sector remains problematical. We should note that all the internal problems discussed above apply whether Moldova deals with the EU in the West or other transition economies and Russia in the East. Yet, Moldova’s agricultural and wine sectors retain a robust export trade with these countries, despite political risks, costly transportation, and a still-developing logistical system. Then too, SMEs of other transition economies – such as those of Poland, Hungary, and the Czech Republic – also suffer from internal restraints yet have been internationalizing westward, not only in terms of breadth but in terms of depth as well. The ﬁrms in these country’s industries – clothing and otherwise – are playing a more active role in the EU economy and bolstering their nation’s economies despite facing a set of internal problems similar to those of Moldova (World Bank, 2004). For these countries, as for Lithuania, outside sources of FDI offer the key to the modernization of infrastructure and of factories and equipment and, in turn, the evolution of ﬁrms further up the value chain. Why then has Moldova’s textile and apparel industry failed to beneﬁt in any substantial way from outside investment funding to the same extent as its counterpart in Lithuania? Moving away from internal restraints and shifting to external issues forces the question of to what extent the continuing growth and integration in the EU are detrimental to Moldova and the internationalization of its clothing and apparel industry. Does this offer possibilities that can help Moldova overcome its internal barriers to internationalization, or does it exacerbate the already difﬁcult situation in Moldova? And why, if the EU actively engages Moldova’s industries (such as apparel and clothing) through such inward activity as subcontracting, has this not led to a more outward international push on the part of Moldova? Indeed, it might be expected that

International Value Creation in the Textiles and Apparel Industry

551

the presence of a growing, more coherent EU market, lower tariffs, lessened restrictions on FDI, and the fact that Moldova’s Eastern European neighbors can serve as entry points into the EU as a whole should spur Moldova’s SMEs and their internationalization activities. Certainly, we found this to be the case for Lithuania, but why has this same process not occurred for the Moldavian clothing and apparel industry? The role of the EU in Moldova’s internationalization efforts offers a compelling possible explanation. If the EU proved a spur to value creation in Lithuania, it acts more as a brake to Moldova’s garment and clothing SMEs in their attempts to evolve along the industry’s value chain. There are certainly indirect constraints that arise from the fact that the EU is expanding and integrating, that Moldova is not considered a likely candidate for accession, at least in the foreseeable future, and that the industry is composed mostly of small, low net proﬁt ﬁrms. There is the issue of customs requirements, for example. The more a company attempts to move up the value chain, the more expensive and time-consuming become customs clearance, import duty procedures, and related EU requirements. Especially onerous are the costs and time required to satisfy EU customs authorities on the ‘‘rules of origin’’ requirements in order to obtain preferential treatment. In order to access the EU on a preferential basis, Moldova producers would have to prove that they imported their fabric from the EU rather than from low-cost areas, such as India and China. Currently, the costs and responsibilities for doing this rest with the dominant, European ‘‘joint’’ partners. But if companies in Moldova were to attempt to move up the value chain, they would need to take over these responsibilities for themselves. In doing so, they would be placed in a potentially untenable economic position: to employ cheap labor countries and forego critical trade preferences in entering EU markets or to enter the Union under preferential treatment but at considerable expense and time proving ‘‘rule of origin’’ requirements and prohibited from taking advantage of important economies by sourcing from non-EU cheap labor countries. There is the issue of international standards. Just as the rise of the global supermarket within the EU constructed barriers to entry into Europe by Moldova’s agricultural and wine producing sectors, so the recent induction of international standards in clothing and apparel is a serious hindrance for Moldova’s ﬁrms to enter into higher stages of the value chain. In particular, the industry standards developed by the International Organization for Standardization (ISO 9000) are a product of, and closely embraced by, the EU community. In part, this is because, in contrast to the US legal system, the EU legal regime is based largely on a code system that is less reliant on

552

SANFORD L. MOSKOWITZ

private (civil) litigation than in the USA. European consumers who purchase products that fail cannot readily seek redress in the courts. As a consequence, Europe relies more heavily on certiﬁcation requirements in general and ISO quality management standards (ISO 9000) in particular across the EU. As the EU has grown, meeting ISO quality standards has become increasingly critical for any company to do business in the region. But meeting ISO standards in European markets is difﬁcult and costly, involving the performing of multiple tests and the obtaining of different certiﬁcates of conformity. The small ﬁrm in Moldova does not generally have the resources to obtain the needed ISO 9000 certiﬁcation. Currently, Moldova has very few companies with ISO 9000 certiﬁcation. In attempting to move into higher stages of the value chain, a company becomes progressively more responsible for quality control. Without the ﬁnancial means – or the government’s help – for adopting ISO quality standards, the company cannot attract FDI from Europe or effectively compete against other ISO-compliant ﬁrms in the European market. Beyond these issues, and in contrast to Lithuania, there is simply no incentive for the EU to subsidize or invest in Moldova’s textile and clothing industry. To a greater extent than Lithuania, Moldova offers very cheap labor. This comparative advantage separates Moldova from other countries in Eastern Europe. Due to such a low labor cost level, attempts by EU countries to help Moldova develop sophisticated production systems that are associated with higher value chain activity – such as computerized cutters, CAD systems, and computer controlled transport networks – would not be feasible investments. Furthermore, with the evolution of the European single market, customers throughout the EU expect to be provided with full service by clothing and apparel companies, including garment and pattern design, material sourcing, and manufacturing. This provides an added incentive for European companies to retain as much control over the entire process as possible. If they form joint ventures with companies in cheap labor countries such as Moldova, companies such as Steilmann’s want to be sure to retain the dominant role. In any case, and in contrast to the situation in Lithuania, it is difﬁcult for Moldavian ﬁrms to piggy back off the reputation and contacts of the EU venture partners. The venture partner simply does not allow it. This is so because the EU contractor does not see Moldavian clothing ﬁrms as being on any higher level than simple subcontracted production. There are a number of reasons for this. Moldova is not part of the EU and so not bound by the standards of the Union. There is therefore a persistent concern in just securing proper production quality. Then too,

International Value Creation in the Textiles and Apparel Industry

553

because Moldova is not an accession country, it is more difﬁcult than it is for Lithuania for EU companies to use Moldova as a base from which to launch marketing activities throughout a region. The EU companies are, therefore, less than anxious to help Moldova out of its Stage 1 position by providing the investment funds and technology needed for high-valued activity. Moreover, any foreign investment money that might ﬁlter from the West to the East is more likely to be diverted to Eastern European countries that are already part of the EU or are expected to become a member shortly. For example, Romania is far more attractive for Western FDI than Moldova since it is an accession country, while Moldova is not. This means that the EU will support Moldavian clothing ﬁrms, but not to the point that they can begin moving up the value chain. EU suppliers will sell materials in bulk at discounts in order to keep these ﬁrms aﬂoat and thus ensure a market and processing of their materials using cheap labor. They will, however, not invest money that could help these ﬁrms advance signiﬁcantly. For instance, Moldova’s SMEs have limited access to ready-to-use service spaces and must themselves make investments in water, heating, and power networks since foreign ﬁrms do not pay for such infrastructural expenditures. This serves to further reduce proﬁt margins and limit future growth. Even larger, growing companies have trouble enticing investment from European partners. Such was the case with the company Zorile, which recently has been looking for a buyer/investor to assist the company’s management to increase production, explore new markets for exports, upgrade existing equipment, and increase working capital. The company’s German partner demonstrated little interest, resulting in the company becoming employee owned. The negotiation process itself that occurs between Moldavian clothing ﬁrms and EU companies is designed to maintain the latter as the dominant partner. A study of the industry in the Soroca region of Moldovia shows that the clothing ﬁrms there suffer from a lack of price transparency in negotiation practice due to currency differences, lack of international experience, and, most critically, the EU’s attempts to keep hidden from nonEU members the standard negotiating practice within the Union. Interviews with entrepreneurs indicate that their ignorance of prices at the regional (EU) level makes them work way under regional price. This lack of knowledge and ‘‘savvy’’ is then exploited by EU ﬁrms to disadvantage when attempting to negotiate higher value. Overall, ‘‘the impression remains that since the early 1990s, neither EU companies of government or EU Commission have really pushed to develop synergies with the textile and clothing sectors in [certain Eastern European]

554

SANFORD L. MOSKOWITZ

countries y As a result, these countries have lost key parts of their value chains’’ (Coster, 2004).

6. DISCUSSION This study compared and contrasted the internationalization process within the textile and apparel industries of two countries, one (Lithuania) that is an integral part of the EU and the other (Moldova) that has remained outside of the Union and thus has been marginalized and isolated within Eastern Europe. Through an examination of Moldova in particular, especially in contrast with the paradigm of Lithuania, the study offers us an interesting case of the extent and limits of value creation within the SME population in one of the ‘‘slow growth’’ transition economies. 6.1. Stage Theory and Value Creation Our study shows that both Lithuania and Moldova, at least with regard to our case industry, appear to follow the stage theory of internationalization, that is, one of step-wise increasing commitment to foreign markets (Johanson & Vahlne, 1977). At the same time, even within a country, not all industries are in the same stage of international development. Within Moldova, the agriculture, wine producing, and clothing and apparel industries are at different stages of internationalization. We found that this is certainly true also for the same industry in two countries. Thus, for Lithuania and Moldova, the textile and clothing industry is not following the same step-wise trajectory: for Lithuania, it is moving fairly steadily and rapidly up the stagewise value creation ladder, while in Moldova it remains stagnating in the very lowest (and least value creating) levels. Our investigation also uncovered a complicating factor of the stage-wise theory of internationalization and value creation. In the ﬁrst place, it is incorrect to make too great a distinction between the different stages of internationalization. In particular, our study shows that rather than exporting and joint venturing being separate stages in the value chain, with exporting being situated lower down the chain, we have seen a converging or even merging of these stages. For the textile and clothing industry in Lithuania and Moldova, we have seen that the two ‘‘stages’’ actually operate in tandem in a reinforcing manner: ﬁrms in these countries form joint venture arrangements with EU (and sometimes Eastern) ﬁrms, with the former exporting ﬁnished (or semiﬁnished) products back to the EU partners.

International Value Creation in the Textiles and Apparel Industry

555

Yet the textile and clothing industry in Lithuania is very different from that in Moldova, and understanding this difference partly requires that we reject the common view that each of the various components of internationalization –exporting, licensing, joint venturing – is homogenous and indivisible. In reality, as our study shows, what are traditionally taken to be discrete stages of international value creation in reality can be further divided into distinct substages with different capability and potential for value creation. For instance, we ﬁnd that there are gradations of internationalization and value creation within the category of exporting activity itself. If exporting is considered one of the early stages in SME internationalization, there are higher and lower substages – within the exporting category. Within Moldova, for example, economic growth in Eastern Europe depends on redirecting international allegiances, at least economically, from Russia and other CIS countries to the expanding EU. These older, traditional Eastern markets remain highly risky for a country too dependent on them and can offer but limited value creation potential in the long run. Rather, tapping into the ever-expanding and increasingly inﬂuential, wealthier, and more sophisticated EU market is the sine qua non of future market growth for the Eastern European countries. Even considering its exporting activities alone, Lithuania has more closely integrated with the EU market, which has indeed led it to further value creation initiatives. For its part, Moldova has only been partly successful in shifting its exporting activity to the West. It has not yet reached the level of integration into the EU. Thus, even this early stage of internationalization has not been fully stabilized and made coherent enough for Moldova’s textile and apparel industry to serve as a solid foundation on which to base further upward movement along the value creation ladder. In a similar vein, even if we accept that the joint venture is a more advanced stage of international development and value creation, it is also true that there are different gradations and degrees of joint venture arrangements. Increasingly, Lithuania’s textile and clothing industry has entered into what we can call ‘‘equal partner’’ joint venture arrangements with EU countries. This, in turn, has led to increased investment into Lithuania, modernization of machinery and plants, and growing participation by Lithuanian ﬁrms in such high-value-added activities as vertical integration of production and international retailing and branding. In stark contrast, we saw how Moldova’s textile and clothing industry stagnated within passive joint venture arrangements. As the minor partner in these arrangements, the industry captured only a small percentage of the total value added, thus retarding its economic growth. The next developmental stage for this

556

SANFORD L. MOSKOWITZ

industry then would be to evolve along the same lines as Lithuania, from passive contract partner to a more dominant player in joint ventures that accommodated the industry’s increasing control over the production, distribution, and marketing process and, in turn, over the creation and capturing of the potential total value added. 6.2. SMEs and Value Creation Our investigation of the textile and clothing industries of Lithuania and Moldova also helps us understand the extent and limits of small and medium-sized ﬁrms in the internationalization and value creation process. In the ﬁrst place, the internationalization of SMEs appears to be quite important for value creation and economic growth in a European transition economy. This is because, in the wake of the privatization movement, a multitude of SMEs were carved out of the large, uneconomic formerly staterun companies. As a consequence, by the turn of the 21st century, the health of these countries’ economies – their ability to create jobs and growth – depended on the viability and creativity of the SMEs. For industries such as textiles and clothing, the vast majority – between 60% and 95% – can be categorized as SMEs. SMEs are generally considered to be ﬂexible and innovative. For example, within the USA and the EU, the bulk of the companies considered as leaders in 21st century technology (such as nanotechnology) are micro to small and medium-sized ﬁrms. This trend has been contrasted with the norm for innovative technology in the post–World War II period, when the large integrated ﬁrms were the innovative engines of creative destruction. Nevertheless, our study forces us to reconsider some of these assumptions about the role of the SME in 21st century value creation and economic growth. For instance, the case of Moldova suggests that the size of a ﬁrm can hinder internationalization and value creation if it becomes too small. In this case, it falls below some minimum size so that various economies are lost. In Moldova’s textile and clothing industry, ﬁrms could not proceed beyond a certain evolutionary point because they simply never grew large enough to have the capability, capital pool, and economic (and political) inﬂuence to be able to control the negotiation of terms with EU companies forming joint venture agreements with them. Indeed, their lack of size and, in turn, reach left them unaware of the pricing policies of EU companies and thus put them at a severe disadvantage when negotiating terms. In these cases, small ﬁrms shift down the value as they ‘‘are on the weak side of the contract relationship.’’ Such ﬁrms are in a precarious position in the value chain.

International Value Creation in the Textiles and Apparel Industry

557

The case of Lithuania is quite different. As our study indicates, textile and apparel ﬁrms could grow out of their SME status. As they did so, they became more integrated – in both forward and backward directions – and eventually achieved an equal (or close to equal) status with their EU partners. These ﬁrms are able to build on subcontracting to invest to capture higher value production, such as adding new sewing processes, including buttonholing and embroidery, or ﬁnishing, such as labeling, packing, and bar coding, as well as distribution, marketing, and advertising. In this case, size and reach – and geographical proximity – meant that the ﬁrms captured more leverage with EU companies in their negotiations. These cases appear to suggest that SMEs can be at a distinct disadvantage in international value creation processes. They can indeed wield economic leverage only if they have a clear core competency with which to leverage, such as new technology and patent protection. Such was not the case with either of the industries we examined in Lithuania and Moldova. The difference between the countries is that the SMEs in Lithuania appear to be evolving past the SME stage, while those in Moldova are not. 6.3. The EU and Inward–Outward Linkages How does one account for the differences that we have observed in the rate and direction of value creation in the textile and clothing industry between Lithuania and Moldova? On one level, we can turn to internal policies and regulations. In the case of Lithuania, internal conditions have proven beneﬁcial to value creation. In contrast, Moldova illustrates how internal problems such as political instability, transportation and logistical costs, lack of cheap capital, and unfavorable taxation fail to offer critical support to – and indeed put serious barriers in the way of – the country’s SME ﬁrms in their efforts to create value by competing in international markets. This study then at a minimum suggests the need for certain transition economies to look inward to revise economic, ﬁscal, and political policy in order to better support their SMEs as they attempt the critical process of internationalizing. But this study also indicates that such efforts, while important, are not by themselves sufﬁcient. As observed, other industries in Moldova and other Eastern European countries show economic vitality and an active trade stance that creates value, even in the face of these internal problems. Moreover, attention by the West in the form of FDI would alleviate a number of constraints to internationalization in Moldova, as was seen to be true in Lithuania. This then leaves the question of why FDI is forthcoming in certain countries for certain industries and not others.

558

SANFORD L. MOSKOWITZ

The progressive expansion and further integration of the EU is generally believed an essential agent for growth for the developing and transition economies of Eastern Europe. This study suggests that this may be true enough for those countries that have either just entered the EU family or are on track to do so over the next few years. After all, these countries – such as Lithuania – secure essential beneﬁts that they did not previously have (despite the problems that often arise for EU countries due to nonoptimal economic area issues). These beneﬁts include free access to an ever-growing market, economic support in the form of subsidies, greater monetary stability, regulatory protection, and so forth. But what of those countries that are not so favored? These nonaccession countries ﬁnd themselves increasingly distanced from this market and therefore marginalized. Whereas in the past these countries’ SMEs could at least trade with individual Eastern European neighbors, even this option becomes closed to them as the EU has expanded from west to east and absorbed these neighbors into the EU system. Thus, the nonaccession countries ﬁnd themselves facing a growing monolithic entity that provides competitive advantages for its own members and thus relatively less opportunities for outlier countries to compete in the vital Western market. We have found then that, as in the case of Lithuania, the EU can nourish and support value creation, but it does not necessarily do so. In the case of Moldova, it acts – both directly and indirectly – as a barrier to value creation. Thus, the Lithuanian textile and clothing industry, increasingly supported by EU policies, regulations, and resources, has been able to move up the value chain; in contrast, in the case of Moldova, the industry – made up of small and undercapitalized SMEs – remains imprisoned and stagnating within the lower (and unproﬁtable) reaches of the value chain. This issue centers on the question of inward–outward linkages in the internationalization process. Certainly, we ﬁnd such strong links in the case of Lithuania, where earlier inward trade and investment on the part of the EU did indeed lead through a sort of multiplier effect to even greater valuecreating external activity by the Lithuanian ﬁrms as they grew and extended their reach – geographically and functionally – across Europe. But the case with Moldova is very different. The EU’s internal activity there – in the form of contracting SMEs to undertake tolling work – has not evolved in any regular way. Indeed, because such activity suits the EU countries quite well – especially considering Moldova’s very low labor costs – and because as a nonmember Moldova cannot readily integrate with the Union’s regulatory and quality control requirements, the EU is not particularly interested in helping push Moldova’s textile and clothing industry up the value chain. In this sense, we have a case that seems to contradict the current theory

International Value Creation in the Textiles and Apparel Industry

559

of inward–outward linkages, i.e., the very inward activity conducted by the EU – because it is so comfortable to maintain – actually acts as a deterrent to Moldova moving up the value chain. The EU, with its attention – and investments – focused on other, closer, and more integrated countries, has scarce resources for upgrading Moldova’s industries. And in any case, Moldova provides the EU with what the Union wants and needs. It is not necessarily to the EU’s advantage to have Moldova expand internationally beyond its current status. This ‘‘don’t rock the boat’’ mentality on the part of the EU means that the observed inward–outward linkages, so useful to Lithuania, act as a break to further international exposure for Moldova.

7. CONCLUSION: INTERNATIONALIZATION, VALUE CREATION, AND THE ‘‘SEAMLESS’’ INDUSTRIAL NETWORK This study comparing Lithuania with Moldova suggests that regionalization – such as that typiﬁed by the expanding EU – can directly and indirectly signiﬁcantly hinder the extent and intensity of internationalization activities – and thus value creation – of other countries that are outlier nations (such as Moldova) in the sense that they are outside of the regional grouping’s orbit. It makes sense that Western European countries would favor EU members. For example, it is easier and cheaper to trade with other EU members. Since 2004, this applies to the 10 Eastern European accession countries. In this study, the question arose why Western Europe goes further than this and actively attempts – often with success – to invest and upgrade the value creation capabilities of industries in accession countries – such as Lithuania – so that they can more readily move up the value chain, while holding back outlier nations such as Moldova. We found, certainly, that Moldova’s SMEs were on the whole below the minimum economic size, that its internal infrastructure remains troubling, and that the country’s low wages provide little inducement for Western European companies to help upgrade machinery and plant. However, as was pointed out, other accession countries – including Lithuania itself – were in a similar situation in the 1990s. What can be suggested here is that an important advantage that the accession countries can claim is that they, because of common processing standards, currencies of exchange, and product quality requirements, more easily ﬁt into the fabric of the Pan-European industrial network. This is important because, with the rise of Internet-based global production and the demand for international just-in-time logistics, these stages in the value

560

SANFORD L. MOSKOWITZ

chain, such as in the apparel industry, are becoming increasingly interdependent. The supply chain from sourcing of raw materials via design and production to distribution and marketing is being organized as an integrated production network where the production is strategically located where it can contribute the most value of the end product, including providing the value-added capability of just-in-time delivery. The critical point is to make goods, information, and payments ﬂow smoothly at each link in the production chain of an industry. To do this, a number of logistics and business services are needed. Integration of information and ﬂow of goods and payments are possible only if all the links in the chain use compatible standards. For example, subcontractors in the textile and apparel industry are increasingly required to add bar codes that comply with EU industry standards to garments before they are shipped. Therefore, as countries join the EU, they also become part of the everexpanding and increasingly integrated production network (such as the Europe-wide textile and apparel industry network) that deﬁnes their industry. The growth of an industry’s network – such as the textile and apparel network – occurs as it becomes part of the overall production structure. The accession countries that join the network certainly achieve signiﬁcant economies of scale over a number of dimensions that they did not have access to before. But the EU and the major companies working within it also increasingly have a stake in ensuring that these new additions to the network grow and prosper as they integrate more or less seamlessly into the PanEuropean industrial structure. This is so because successful additions to the network beneﬁt all members through ever-growing economies. Moreover, as the network expands and becomes increasingly interconnected, bottlenecks that occur anywhere in an industry’s network cause problems throughout the system and that decreases the value in the activities of all members. In the apparel industry, for example, such problems can lead to delays in delivery, thus reducing the ﬁnal value of fashion-centered clothing. The EU and the major companies in its various industries understand that it is then critical that the new companies be supported as they attempt to move up the value chain of their industry. By doing so, all members within that system themselves achieve greater value added. It is for this reason that inward–outward linkages can appear so strong in Lithuania’s textile and apparel industry. In contrast, a nonaccession country like Moldova simply does not ﬁgure in this equation. While the fact that Moldova’s textile and apparel industry is more rooted to the EU than is its agricultural sector indicates that it is more ‘‘integrable’’ with Western Europe’s industrial network, this integration only goes so far. As we have seen, Moldova’s industry has not had an

International Value Creation in the Textiles and Apparel Industry

561

easy time more fully incorporating itself into Europe’s production system. Quality, standards, currency, and even cultural issues have prevented this. As a result, Moldova’s textile and apparel ﬁrms have not been able to become a seamless part of the Pan-European clothing industry. Thus, investing in this country does not add much value for existing EU members, nor would problems that arise in this industry impact greatly the existing European textile and apparel industrial network. This is not to say that Western European companies do not value Moldova’s comparative advantage in cheap labor. But this is as far as it goes. Indeed, the very fact of this advantage – and that the EU has little incentive to help the country’s clothing industry move beyond it – helps to maintain Moldova’s textile and apparel industry in its low position in the value chain. This is why, for example, we found a weaker inward–outward internationalization link here than we did in Lithuania. 7.1. Implications These conclusions point to a number of implications. In the ﬁrst place, as indicated previously, they force a reconsideration of certain current theories and beliefs regarding the internationalization process in general. For example, we see that the role of SMEs in economic growth is far from certain. The creation of SMEs through privatization (especially in Eastern Europe) can result in ﬁrms too small to exert a critical inﬂuence and to achieve required economies. Further, the results underscore factors that may retard the internationalization process of SMEs (and other ﬁrms) through the various stages of internationalization. Thus, while not contradicting the stage theory, the results show that movement up through the various stages is far from automatic but may be thwarted for various internal and external reasons. Then too, the study reconsiders the concept of inward–outward linkages in the internationalization process. SMEs in particular are assumed to beneﬁt from initial inward activity by larger companies. In fact, we ﬁnd that this is not necessarily the case. Indeed, in the case of an outlier country like Moldova, inward activities of EU companies serve as barriers to Moldavian SMEs in their attempt to enter into more sophisticated and valuecreating international arrangements. Beyond these theoretical considerations are potential geoeconomic consequences that stem from the rigidities inherent in the EU’s growing and increasingly interdependent industrial network. These considerations question the relationship that the EU will continue to have with those Eastern European countries that are not part of the Union. The case of Moldova

562

SANFORD L. MOSKOWITZ

begs the question of whether wealth – in the form of value added – continues to be transferred from East to West, and what this implies with respect to the economic, social, and political stability of Eastern Europe. Indeed, given the fact that a number of factors continue to prevent even the Eastern European accession countries from seamlessly integrating into the EU’s industrial network – currency issues, the subsidy question, labor inelasticities, standard differentials, and so forth – this may remain a problem even for current Eastern European members. Will then Eastern Europe’s inability to meld into Western Europe’s economic network exacerbate tensions between the West and the East with the heightening of problems and conﬂict between these regions? As one recent study indicates, it is important to ‘‘examine some of the ways in which the integration of East European [industrial] production into wider production networks is occurring and consider the implications of these trajectories for our understanding of industrial upgrading’’ (Pickles & Smith, 2005). There is also the question of the EU’s ability to absorb innovation into its structure. If new technology threatens the EU’s integrated and ﬁne-tuned network – for example, by requiring different production standards, setting new levels of quality, and so forth – it may be rejected by the EU as a disruption to the system as a whole. There is indeed evidence that this is happening and that, just as it has been growing and integrating, the EU has been falling behind the USA and Asia in new technology and productivity (Moskowitz, 2006). All this may have profound implications for the continued stability and competitiveness of the world’s largest regional grouping and therefore is of enormous importance for the future course of the globalization movement in Europe and beyond. Finally, this research suggests potential future research activity in the ﬁeld of internationalization and value creation within ﬁrms, countries, and regions. Our work here examined the important industries in but one outlier country. Additional studies of other industries, countries, and even regions are strongly suggested to further our understanding of wide-ranging industrial networking, the internationalization process, and value creation in an increasingly global and interconnected economy.

NOTES 1. The EU has also worked with Lithuania’s industry to help develop central institutes for the advancement of the textiles and apparel industry and to bring the industry in closer accord with European standards. The Lithuanian Textile Institute

International Value Creation in the Textiles and Apparel Industry

563

(LTI) creates, designs, and manufactures textile fabrics and protective clothes. The product range is the result of research and technological development in close cooperation with Western European leaders in the ﬁeld. A modern testing laboratory provides analysis of textile garments according to EN and ISO standards. In the near future, the LTI will issue the European Certificate for Textile Fabrics. In 2002, the last LTI accreditation procedures were completed, which are valid in all European countries. The Lithuanian Apparel and Textile Industry Association – representing 80% of the local output – has been a member of the European Apparel and Textile Organization since late 2001 (Lasiauskas, 2004; Swedish Trade Council, 2003). 2. Long lead times compromise a company’s ability to capture optimal value from its production. For example, Zara, the leading fashion chain in Europe, say that due to their short lead times, they sell 80–85% of their products at full price compared with traditional retailing, which barely reaches 60–70% (Pickles & Smith, 2005).

REFERENCES Anson, R. (September, 2002). Profile of the textile and clothing industry in Lithuania (pp. 5–38). Wilmslow, UK: Textiles Intelligence Ltd. Coster, J. (2004). The apparel industry in the ten EU accession states. Worcestershire, UK: Aroq. EKT Group. (2004). The textile and apparel industry in Lithuania: A study of the industry’s prospects. Vilnius, Lithuania: Lithuanian Development Agency. Gerefﬁ, G., & Memedovic, O. (2003). The global apparel value chain: What prospects for upgrading by developing countries (Sectoral Studies Series). Vienna: United Nations Industrial Development Organization. International Monetary Fund. (2005). Republic of Moldova: Selected issues. IMF country report. Washington, DC. Johanson, J., & Vahlne, J. (1977). The internationalization process of the ﬁrm: A model of knowledge development and increasing market commitments. Journal of International Business Studies, 8(1), 23–32. Knight, G. A. (2000). Entrepreneurship and marketing strategy: The SME under globalization. Journal of International Marketing, 8(2), 12–32. Lasiauskas, L. (2004). Textile and apparel industry development trends in Lithuanian and other Central and Eastern European countries. Fashion.Net international conference. Vilnius, Lithuania. Moskowitz, S. (2006). Globalization, technology, and national competitive advantage: The case of the US vs. the EU. Proceedings of the Academy of International Business, Oklahoma City, OK. Nordas, H. (2004). The global textile and clothing industry post the agreement on textiles and clothing (Discussion Paper No. 5). Geneva, Switzerland: World Trade Organization. Oviatt, B. M., & McDougall, P. P. (1994). Toward a theory of international new ventures. Journal of International Business Studies, 25(1), 45–61. Oviatt, B. M., & McDougall, P. P. (1997). Challenges for Internationalization Process Theory: The case of international new ventures. Management International Review, 73(2), 85–99. Pickles, J., & Smith, A. (2005). Liberalization, integration, and convergence: Pan European production networks, post-socialist apparel, and the embedded geographies of competition. Berlin: VII ICCEES World Congress.

564

SANFORD L. MOSKOWITZ

Porter, M. E. (1980). Competitive strategy: Techniques for analyzing industries and competitors. New York: Free Press. Porter, M. E. (1985). Competitive advantage, creating and sustaining superior performance. New York: Free Press. Roolaht, T. (2002). The internationalization of Estonian companies: An exploratory study of relationship aspects. Dissertationes Rerum Oeconomicarum, Universitatis Tartuensis, Tartu, Estonia. Swedish Trade Council. (2003). Textile industry: Business opportunities in Lithuania. Vilnius, Lithuania. Welch, L. S., & Luostarinen, R. K. (1993). Inward and outward connections in internationalization. Journal of International Marketing, 1(1), 46–58. World Bank. (2004). The Republic of Moldova: Trade diagnostics study No. 30998-MD. World Bank Report.

STRATEGY DURING AN INDUSTRY CRISIS: THE POST-QUOTA EXPERIENCE OF TURKISH APPAREL MANUFACTURERS Liesl Riddle ABSTRACT Discussions about the elimination of apparel quotas have focused on countries that obviously benefit or are harmed by their demise. Little attention has been paid to countries for which the post-quota environment is uncertain – and vital. As quotas were lifted in January 2005, uncertainty loomed particularly large for Turkey, the world’s fourth largest apparel exporting nation. This paper utilizes secondary data and a survey to chronicle Turkish apparel exporters’ strategic expectations, preparations, and responses to the post-quota environment. The case details the unexpected consequences of quota elimination for the industry, including how the new competitive environment catalyzed many manufacturers to locate production in foreign lands. Apparel manufacturing is one of the most globalized industries in the world (Rossen, 2004). Today, world production, valued at US$ 226 billion Value Creation in Multinational Enterprise International Finance Review, Volume 7, 565–587 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3767/doi:10.1016/S1569-3767(06)07023-3

565

566

LIESL RIDDLE

(3.1 percent of world merchandise trade), is fragmented and dispersed around the globe. The top 15 textile exporting countries – which originate from every continent except Australia and Antarctica – account for more than three fourths (79 percent) of world apparel exports (see Table 1). In stark contrast, the importation of textiles is narrowly concentrated: approximately three fourths (76 percent) of world apparel exports are imported into just two countries, the European Union (EU) and the United States (US) (http://www.wto.org/english/res_e/statis_e/its2004_e/its04_bysector_e.htm). Since 1974, textile and apparel imports to the EU and the US have been dominated by a system of quotas, mostly established through the Multiﬁber Arrangement.1 This quota system has limited the amount of textile and apparel exports that quota-subjected countries could send into European and American markets.2 But quotas also have reduced uncertainty for exporters: they set known, quantiﬁable limits on the amount of goods that competitor countries could send into quota-limited markets. The quota system encouraged the proliferation of textile and apparel manufacturing in many countries (Anderson, 2000). Quota limits often forced buyers to ‘‘quota hop,’’ to source products from alternative countries when their supplying countries approached their quota limits (de Jonquieres, 2004). Quotas often helped create and sustain textile and apparel markets in many developing countries (Appelbaum, 2004). Table 1. Top 15 Exporting Countries by Export Value and Share in World Exports. Exporting Country European Union Extra-EU (15) China Hong Kong Domestic Turkey Mexico India United States Bangladesh Indonesia Romania Thailand Korea, Republic of Vietnam Morocco Pakistan

Export Value (billions of US dollars)

Percent of World Exports

59.95 19.04 52.06 8.20 9.94 7.34 6.46 5.54 4.36 4.11 4.07 3.62 3.61 3.56 2.83 2.71

26.5 8.4 23 3.6 4.4 3.2 2.9 2.5 1.9 1.8 1.8 1.6 1.6 1.6 1.3 1.2

Strategy During an Industry Crisis

567

On January 1, 2005, as per the WTO’s Agreement in Textiles and Clothing (ATC), the quota system was fully abolished.3 Women’s Wear Daily, a leading apparel industry journal, predicted that quota elimination would ‘‘begin a seismic shift’’ in worldwide apparel sourcing (Ellis, 2003a). A Financial Times article warned of a ‘‘global shake-out as quotas end’’ (de Jonquieres, 2004). Leading up to 2005, there was a ﬂurry of tradejournal articles and analyst reports speculating about the quota-free competitive environment. Although there was a general consensus that China and India would be the big winners and AGOA4 member countries would be the big losers of quota elimination, there was disagreement about how most other countries would fare after January 1, 2005. As the quota-elimination date approached, the uncertainty loomed particularly large for Turkey. In 2004, Turkey was the world’s fourth largest clothing exporting nation. Turkish clothing exports were valued at around US$ 9.94 billion and comprised 4.4 percent of worldwide clothing exports (http://www.wto.org/english/res_e/statis_e/its2004_e/its04_bysector_e.htm). The textile and apparel export industry was vital to the Turkish economy; it accounted for approximately 10 percent of GDP, 18 percent of industrial production, 20 percent of the manufacturing labor force, and generated 32 percent of Turkey’s export earnings (http://www.itkib.org.tr/). Turkey, like many other developing countries, initially fought for the removal of these quotas when the ATC was signed in 1994. They anticipated that quota elimination would bring about increased market share, particularly in the US where they faced stringent quotas in their major export categories (Wilson, 1999). However, China’s entry into WTO membership affected the market-share prospects for many developing countries, including Turkey. Leading up to 2004, analyst and trade-journal reports were mixed in terms of their expectations for Turkish apparel sales in a post-quota environment. Some argued that Turkey would maintain or even gain market share. Others warned that the Turkish apparel industry would be devastated; several predicted that small and medium-sized ﬁrms – which comprise the majority of the industry – would be especially hard-hit. As the quotaelimination date approached, Turkish business associations were at the forefront of an international movement organized to persuade the WTO to postpone quota elimination (Zarocostas, 2004). This paper details Turkey’s quota-elimination experience. I begin with a summary of reports about post-quota elimination national winners and losers. Next, I turn to the Turkish case. After providing an overview of the Turkish clothing industry, I draw upon a survey conducted among Turkish

568

LIESL RIDDLE

apparel exporters during the ﬁrst weeks of 2005 to assess their expectations of and strategic preparations for the quota-free competitive environment. I conclude with a discussion of what happened to Turkish apparel exports in 2005.

EXPECTATIONS ABOUT POST-QUOTA ELIMINATION WINNERS AND LOSERS In 1994, most developing countries hailed the quota-elimination plan as a great victory (Smeets, 1995). But this enthusiasm waned once China became a WTO member on December 11, 2001. Alarm among China’s competitors began in earnest when the results of 2002 ﬁrst-quarter imports to the US market were revealed. In just a matter of months, China’s share of the 29 categories that had been removed from quota that year had more than tripled, rising from 9 percent in 2001 to 45 percent by the ﬁrst quarter of 2003 (American Textile Manufacturers Institute, 2004). By 2004, China’s share of quota-released categories soared to 65 percent (National Council of Textile Associations, 2004). China achieved similar dramatic success in the EU and Canadian markets as well. Industry reports pointed to remarkable price declines after quotas were eliminated in Chinese apparel imports: the average price in quota-free Chinese apparel categories decreased from US$ 6.23 in 2001 to US$ 3.12 in 2004 (National Council of Textile Organizations, 2004, p. 3). In response to these events, between 2002 and 2004, several reports were written by industry associations, policymakers, and journalists, claiming that China would be the primary winner of the ‘‘big bang’’ of quota elimination (Gresser, 2004). Even under quota limits in the major textile and apparel-importing countries, China was the world’s leading textile exporter and second-leading apparel exporter in 2004. Textile and apparel manufacturing accounted for 10 percent of the country’s manufacturing output and 20 percent of Chinese exports (Buckman, 2004). Although ofﬁcial statistics indicated that the combined textile and apparel industry comprised 21,000 enterprises and 7.9 million workers (14.5 percent of China’s industrial employment), the China National Technical Importer/Exporter Corporation Group estimated that employment – once informal labor was taken into account – was closer to 15 million workers (ITC, 2004). The sector was dominated by several large state-owned companies; the remainder comprised small ﬁrms (ITC, 2004). China’s relatively low hourly wages in the textile and apparel sector (see Fig. 1) were a major factor in the price-competitiveness of Chinese apparel

Strategy During an Industry Crisis

$12.00

$11.16 $10.20 $10.03

average hourly rates

$10.00

$8.00

$6.00

$5.13 $5.11

$4.00

$2.13 $2.00

$1.75 $1.57 $1.08 $0.94 $0.86 $0.71 Pakistan

Indonesia

India

China

Mauritius

El Salvador

South Africa

Mexico

Turkey

Korea

Hong Kong

Germany

UK

Hourly Labor Costs in the Textile and Apparel Industry. Source: ILO Labor Statistics http://laborsta.ilo.org/

569

Fig. 1.

USA

$0.24 $0.23 $0.00

570

LIESL RIDDLE

products. In anticipation of WTO membership and quota elimination, China embarked on what the government referred to as a ‘‘reform equals rescue’’ plan in 1998, eliminating 1.5 million jobs in the sector and upgrading technologies to enhance efﬁciency. International Textile Manufacturers Federation statistics illustrate China’s commitment to technological improvements: China purchased three fourths of the shuttleless weaving machinery that was sold in world markets between 2002 and 2004, two thirds of all texturing sales in the world between 2002 and 2003, and 60 percent of global wool- and cotton-ring spinning frames between 2002 and 2003 (National Council of Textile Associations, 2004, p. 4). In 2004, additional reports were released that made predictions about post-quota winners and losers. There continued to be a general consensus that China would be the major winner of a post-quota world market, but there was disagreement about the magnitude of China’s expected marketshare gains (Malone, 2004). A WTO report estimated that Chinese apparel would increase 29 percent in the EU and 50 percent in the US once quotas were lifted. A McKinsey study examining the impact of quota elimination on total world exports posited that China’s share in worldwide apparel would increase from 22 percent to 31–50 percent, depending on how long it would take the EU and the US to impose safeguard measures (Padhi, Pauwels, & Taylor, 2004, p. 10). Fashiondex Inc., a textile and apparel consulting company, argued for a more conservative estimate of China’s market share prospects (market share of 27 percent for both the EU and US), noting that more than half of China’s current garment exporters were not regulated by quotas (Malone, 2004). India – the only country to maintain market share in the EU and US markets in quota-free apparel categories after 2001 – was also often mentioned as a post-quota ‘‘winner’’ (National Council of Textile Organizations, 2004). Countries whose industry mostly comprised foreign direct investors from China seeking to circumvent quotas on Chinese goods, were often mentioned as expected losers in a post-quota world since much of that investment was expected to be retracted once quotas were lifted (Appelbaum, 2004). But for most of the rest of the world’s apparel-producing nations, the future was questionable. Appelbaum’s (2004) review of over 60 studies and reports on the impact of quota elimination observes a substantial disagreement in the literature regarding who would be helped and hurt by quota removal – and by how much. Turkey was among the list of nations for which the post-quota-elimination future was uncertain.

Strategy During an Industry Crisis

571

TURKISH APPAREL BEFORE QUOTA ELIMINATION The Turkish textile and apparel industry has been central to the historic economic development of the country. The textile and apparel sector was granted the ﬁrst priority for industrial investment and development in the newly independent republic’s 1925 First Development Plan (Owen & Pamuk, 1999), and the sector grew exponentially throughout the end of the 20th century. A 1999 survey of Istanbul textile and apparel ﬁrms suggests that 92 percent were founded after 1980 (Riddle, 2001). As the industry grew, many manufacturers began to focus their exports on apparel. In 1980, clothing exports contributed only 14 percent to total Turkish textile exports. But, by 2003, this percentage had climbed to 76 percent (http://www. itkib.org.tr/). Only a small percentage of apparel exports comprises high value-added products; most Turkish apparel exports are in lower value-added categories, such as T-shirts (Murphy, 2002). Cotton- and wool-knitted apparel and accessories account for a little more than half (51 percent) of Turkish apparel exports, while cotton-woven apparel and accessories (35 percent) and made-up textile articles (14 percent) comprise the remainder. Turkish apparel exports are primarily sent to the EU (destination for 72 percent of Turkish apparel exports) and the US (16 percent of Turkish apparel exports), the two largest apparel markets in the world (http:// www.itkib.org.tr/).5 Prior to 2005, Turkish apparel exports were subject to several quotas in the US market (33 categories).6 Over the course of the Multiﬁber Arrangement, quota limits were raised several times for Turkey, including following the 1999 earthquake (Ostroff, 1999) and during the USled invasion of Iraq in 1991.7 At the end of 2004, Turkey’s average quota-ﬁll rate was 38.6 percent. Yet, several products – particularly in men’s and boys’ apparel categories – regularly reached 80 percent or greater ﬁll rates,8 often within a couple of months (see Table 2). Turkey was not subject to quotas in the EU, where they enjoyed Customs Union status since joining the union in 1995. Although in 2004 Turkey maintained an overall trade surplus in textiles and apparel (US$ 8.4 billion), it continued to import substantial apparel and apparel-accessories goods. EU countries provided the bulk of these imports (59 percent), while the US supplied 17 percent, China 12 percent, and other countries 12 percent. In late December 2004, Turkey was the ﬁrst country (followed by Argentina and the US) to impose safeguard measures against 42 categories of Chinese apparel imports (Zarocostas, 2005).

572

LIESL RIDDLE

Table 2. Category

219 313 314 315 317 326 617 625/626/627/628/ 629 625 626 627 628 629 200 300/301 335 336/636 338/339/638/639

338/339/638/639 S 340/640 340 Y/640 Y 341/641 341 Y/641 Y 342/642 347/348 347 T/348 T 351/651 352/652 361 369 S 410/624 410 448 604 611

Quota Fill Rates by Category for Turkey. Description

Percent Utilization as of 12/31/2003

Duck Sheeting Poplin & Broadcloth Printcloth Twills Sateens Twills & Sateens, Staple Descriptions Below

84.5 10.99 1.45 0.06 11.55 53.99 3.73 20.52

Poplin & Broadcloth, Staple/Filament Printcloth, Staple/Filament Sheeting, Staple/Filament Twills & Sateens, Staple/Filament Other, Staple/Filament Sewing Thread; Yarns for Retail Sale Carded (uncombed), Combed Coats, Women’s & Girls’ Dresses (Cotton & man-made Fiber) Knit Shirts, Men’s & Boys,’ Knit Shirts & Blouses, Women’s & Girls,’ Knit Shirts, Men & Boys,’ Knit Shirts & Blouses, Women’s & Girls Same as above Woven Shirts, Men’s & Boy’s Same as above Woven Shirts, Men’s & Boys,’ Woven Shirts & Blouses, Women’s & Girls’ Same as above Skirts (Cotton & man-made Fiber) Trousers, Shorts, etc., Men’s & Boys, Women’s & Girls’ Same as above Nightwear & Pajamas (Cotton & man-made Fiber) Underwear (Cotton & man-made Fiber) Sheets Other Manufacturers n.e.s. Woven, >36% by weight wool, Woven, >15% &o36% wool Woven, >36% by weight wool Trousers, Shorts, etc., Women’s & Girls’ Synthetic Staple Artiﬁcial Staple>85% by weight

18.95 0.43 0 2.18 29.75 60.12 23.37 47.72 24.08 95.33

90.85 17.17 12.54 20.84 7.39 31.97 52.62 96.82 96.73 96.67 100 43.28 23.78 33.97 95.96 94.34 1.71

Strategy During an Industry Crisis

573

The Turkish textile and apparel industry comprises over 50,000 ﬁrms, most of which are very small, and 83 percent of Turkish textile and apparel ﬁrms employ less than 10 people. The 41 largest ﬁrms, however, account for nearly 55 percent of all production capacity, and these companies rank among the 500 largest textile and apparel ﬁrms in the world (http://www.itkib.org.tr/). Most apparel ﬁrms are located in Istanbul (Riddle & Gillespie, 2003). Ofﬁcial statistics state that 500,000 are employed in the textile and apparel industry, but government ofﬁcials claim that informal employment comprises 2 million persons in the sector (Smid & Taskesen, 2002). The majority of Turkish apparel manufacturers’ costs are material costs; labor and high electricity costs are also considerable (Ghemawat & Baird, 1998). Small or medium-sized enterprises (SMEs)9 often struggle to compete with large ﬁrms in the sector. Larger ﬁrms have better access to market research and capital (Riddle, 2001), and smaller ﬁrms ﬁnd it challenging to comply with export formalities, such as collecting the 15 percent valueadded tax (Ghemawat & Baird, 1998). Several of Turkey’s largest apparel ﬁrms are fully integrated ﬁrms; their production includes some ﬁber processing, spinning, weaving, dyeing, printing, and ﬁnishing operations (Neidik, 2004). Many of these companies have developed well-known local brands, but only a few (10 percent of total exports) have begun to enter into foreign markets with branded products (Murphy, 2001). For example, Vakko and Mavi Jeans have opened their own branded retail stores in the German and US markets, respectively. On average, Turkish exporters’ delivery time (including manufacturing and transportation) to the EU market is eight weeks compared to 12–13 weeks for Chinese and other Asian exporters (Ghemawat & Baird, 1998). The Winners & Losers 2005 report, published by Fashiondex Inc., claimed that Turkey’s close proximity to the EU market would make it a post-quota winner. In addition, this report argued that Turkey would succeed in the US market after quota elimination because the industry was diversiﬁed and had substantially improved quality (Malone, 2004). Analysts’ assessments of the potential impact of quota-elimination on Turkey were mixed. Several reports conducted by American, European, and Turkish analysts argued that much lower labor costs in competing countries (including China) would render Turkish export goods uncompetitive in international markets (e.g., Nordas, 2004; American Textile Manufacturers Institute, 2004). A WTO study warned that China’s labor costs were low enough to sustain its price advantage even in the EU, a market in which Turkey was privy to duty-free access (Nordas, 2004). A survey of US buyers conducted by the ITC noted that buyer’s perceptions of higher Turkish

574

LIESL RIDDLE

apparel prices and lower relative product quality would result in ‘‘a shift in supply patterns from Turkey to China and Hong Kong’’ (ITC, 2004, L-42). The negative impact was expected to be particularly strong for SMEs, which typically would not have the resources or capability to absorb these losses. Other analysts’ reports were more upbeat, arguing that textile and apparel buying decisions were no longer made on price alone (e.g., Abernathy, Volpe, & Weil, 2005; Appelbaum, 2004; Ikenson, 2003). They pointed to the fact that retail consolidation in the EU and US and inventory and supplychain information systems had changed buyers’ decision-making criteria in the apparel industry. Retailers’ larger product-line offerings and lean inventories resulted in buyers placing greater importance on a supplier’s speed of delivery. Some analysts argued that Turkey’s advantages of geographic proximity to Europe and lower supply-chain throughput time would give it a distinct advantage over China and other competitors – particularly in the EU market (e.g., ITC, 2004; Munir 2004). But business association leaders in Turkey were convinced that quota elimination would be devastating to the industry. In July 2004, the Istanbul Textile and Apparel Exporters’ Union (ITKIB) decided to take action. Along with a US association, ITKIB organized a conference of over 80 textile and apparel associations representing 46 countries. The result of this conference has become known as the ‘‘Istanbul Declaration,’’ a plea to the WTO for a postponement of quota elimination. The declaration predicted that ‘‘trade-distorting practices,’’ such as ‘‘deliberate currency undervaluation, state subsidies, and a proliferation of non-performing loans and rebate schemes’’ by some countries would enable these countries to dominate global textile trade in a quota-free world, resulting in ‘‘massive job disruption and business bankruptcies in dozens of countries dependent on textiles and clothing’’ (http://www.itkibusa.org/). Despite these last-ditch efforts, on January 1, 2005, remaining quotas were abolished. As Women’s Wear Daily magazine reported, ‘‘A new world has dawned in textile and apparel trade. Now the fallout begins’’ (Ellis, 2005, p. 1).

TURKISH APPAREL EXPORTERS’ EXPECTATIONS AND PREPARATIONS FOR QUOTA ELIMINATION But what the fallout would be for Turkish apparel exporters – and how prepared they were for it – was unknown. To what extent were Turkish apparel exporters aware that these quota changes had occurred? What were their market-share expectations for Turkey and their ﬁrm in the formerly

Strategy During an Industry Crisis

575

quota-controlled EU and US markets? What strategic preparations – if any – had they made in anticipation of the post-2005 competitive environment? How might awareness, expectations, and strategic preparation vary among SMEs and larger ﬁrms? Methodology These research questions were explored via a survey of 100 export managers of Turkish apparel manufacturers. Only ﬁrms exporting some proportion of their sales to the EU and/or US markets were considered eligible for participation in the survey. The survey was backtranslated from English to Turkish. A sampling frame of Turkish clothing exporters was acquired from The Istanbul Textile and Clothing Exporters’ Union (ITKIB). ITKIB’s list is deemed a representative enumeration of the population of Turkish clothing exporters since every Turkish clothing exporter is required by law to register with ITKIB. A sample of 100 export ﬁrm owners was selected from the sampling frame. The enumeration was arranged from largest to smallest amount of 2004 export sales. In order to insure that the sample would include large, medium, and small exporters, the enumeration was divided into ﬁve equal strata, and 20 ﬁrms were chosen within each stratum by random selection. The sample was refreshed to achieve 20 completed interviews with ﬁrm owners of companies within each stratum. Data collection took place between January 2 and January 15, 2005. The overall response rate was 63 percent. A ﬁrmographic proﬁle of the sample is presented in Table 3. Respondents were asked about their awareness of quota elimination in the American and European markets, what their market share expectations were for 2005 (for Turkey in general and their ﬁrm speciﬁcally) in the American and European markets, and what strategic preparations they had made – if any – to prepare for competition in a quota-free world. Awareness To test exporters’ level of awareness of quota elimination, respondents were asked ‘‘Are you aware of any quota changes that have occurred recently in the apparel industry?’’ One quarter of the respondents answered ‘‘no’’ to this question . All but one of those responding negatively to this pseudoaided awareness question were export managers at SMEs – comprising almost one third (32 percent) of all SME respondents.

576

LIESL RIDDLE

Table 3.

Sample Proﬁle. Total Sample n ¼ 100

Date Founded Pre-1980 1980–1989 1990–1999 2000+ Average Firm Age in Years. (Std. Dev)

5% 31% 32% 32% 12.2 (9.4)

Number of Employees Less than 50 50–99 100–499 500+ Average No. of Employees (Std. Dev.)

54% 20% 20% 6% 141.1 (324.8)

Total 2004 Exports US$ Less than $500,000 $500,000-$999,999 $1 million-$4,999,999 $5 million or more Average Total Exports (Std. Dev.) Export Intensity Less than 25% 25%–49% 50%–99% 100% Average Export Intensity (Std. Dev.) Average no. Export Destination (Std. Dev.) Export Product Categories Cotton Apparel (SIC 300) Wool Apparel (SIC 400) Man-Made Fiber Apparel (SIC 600) Silk Apparel (SIC 700) Silk Blend/Non-Cotton Veg. Fiber Apparel (SIC 800) Export Quota-Controlled Products Experienced Export Growth: last 3 yrs. Experienced Export Decline: last 3 yrs. Average Export Sales Change: last 3 yrs.

33% 12% 33% 22% $4.8 million (1.1 million)

13% 9% 42% 36% 73.7% (30.6) 6.9 (7.0)

88% 40% 55% 26% 15% 69% 67% 26% 7.9% (32.3)

Comparison to Industry Averagea

4.1 (1.7)

Export Profitability: last 1 yr.b

3.9 (1.6)

pr0.10; pr0.05; pr0.01. a

Respondents were asked, ‘‘How does your average annual export sales growth/decline compared to the industry average?’’ The response was given using a 7-point scale, where 1 ¼ Poor and 7 ¼ Oustanding. b Respondents were asked, ‘‘Overall, how proﬁtable has exporting been over the last ﬁnancial year?’’ The response was given using a 7-point scale, where 1 ¼ Poor and 7 ¼ Outstanding.

Strategy During an Industry Crisis

577

Market Share Expectations for Turkish Apparel in the US Market Overall Overall, respondents were optimistic about Turkey’s prospects in a quotafree American market (see Fig. 2). Over one third (36 percent) expected an increase in Turkey’s market share in the US, and almost half (49 percent) expected Turkey to maintain its share. Those indicating an expected market-share increase for Turkey were queried as to why they expected this outcome. Among the 36 respondents expecting a market-share increase for Turkish products in the US market, 19 reported that they predicted that Turkey’s market share would increase because Turkish ﬁrms would be able to further penetrate categories that were previously limited. Seven respondents indicated that superior Turkish product quality would fuel a market share increase in the US. An additional seven argued that Turkey’s fast turnaround time would convince buyers to purchase Turkish products. Three respondents mentioned Turkey’s labor costs as the reason the country should expect a market-share increase. Among the 15 respondents who predicted that Turkey would lose market share in the US, four cited a general increase in competition from other countries. Two named the ‘‘Far Eastern countries’’ and eight cited China as a speciﬁc threat. Market Share Expectations for Respondents’ Specific Firm in the US Market Fewer respondents (17 percent) expected market-share gains for their ﬁrm in the 2005 US market. The majority of respondents (78 percent) believed their ﬁrm would maintain their current market share. Only 5 percent expected a loss of market share for their ﬁrm in the US market. Among the 17 respondents who expected a market-share gain for their ﬁrm in the US market, 15 cited the opportunity to penetrate previously limited categories as the reason for their positive expectation. Four mentioned that their prices would be lower, since they would not have to buy quotas from other companies or pay intermediaries. Three indicated product quality as the reason for their optimistic outlook. One respondent argued that, ‘‘with the elimination of quotas, there will be more ﬂights to Turkey, which will carry our products. Now it is difﬁcult to ﬁnd available transport.’’ All respondents predicting a market share loss in the US cited the fear that their prices would be too expensive relative to the competition.

578

90%

EXPECTATIONS FOR FIRM Increase Stay Same Decrease

80%

81%

78%

69%

70%

EXPECTATIONS FOR TURKEY 60%

46% 50%

40%

49%

49%

36% x= 33.5 (18.0)

x= 34.2 (21.4)

46%

32% x= 33.0 (16.4)

23% 19%

30%

x= 41.5 (21.5)***

15% 20%

x= 40.3 (20.2)

10%

17%

15%

x= 27.1 (12.0)

x= 29.1 (13.6)

8%

5%

x= 32.5 (3.5)

x= 40.0 (34.5)

x= 23.3 (8.2)

8% 4% x= 55.0 (39.1)***

x= 17.5 (3.5)

Overall

Fig. 2.

SME

Large

Overall

SME

Large

Market Share Expectations for Turkish Apparel and Respondent Firm in the US Market.

LIESL RIDDLE

0%

Strategy During an Industry Crisis

579

Market Share Expectations for Turkish Apparel in the EU Market Overall Respondents also registered positive expectations for Turkey’s market share in the EU (see Fig. 3). In fact, the reported percentages for market-share increase, staying the same, or decrease are virtually identical when compared to the expectations for the US market. More than one half (54 percent) expect Turkey’s market share to stay the same, over one third (34 percent) expect a market-share increase, while 12 percent expect a market-share decline. Among the 39 respondents predicting a market-share increase for Turkey in the EU market, 19 reported that they felt that Turkey’s market share would increase because of a ‘‘decrease in bureaucracy or procedures’’ associated with quota elimination. Five respondents mentioned Turkey’s product quality and geographic proximity, respectively. Four respondents maintained Turkey’s labor quality would enhance its share in a post-quota EU market, and two respondents mentioned Turkey’s fast turnaround time as a reason for a share increase. Two respondents offered that Turkey’s political importance for Europe would facilitate a market-share increase. Country-of-origin issues were addressed by two respondents. One observed, ‘‘products bought from Far East countries are less proﬁtable to Europeans,’’ while another commented, ‘‘Turkey is working on becoming a brand name.’’ Among the 12 respondents predicting a market-share loss for Turkey in the EU market, four mentioned that competition in general would increase. Half of the respondents speciﬁcally mentioned China as the reason for Turkey’s market-share loss, and two mentioned ‘‘Far East nations’’ as a more general culprit. Market Share Expectations for Respondents’ Specific Firm in the EU Market When asked about their expectations for their speciﬁc ﬁrm’s market share in the EU after quotas are removed, respondents were more cautious. Two thirds expected market share to stay the same, and a little over one fourth (27 percent) expected market share to increase. Only 7 percent expected a market-share loss. Although there were no observed differences in the proportions of expected increase or decrease among SME and larger ﬁrms, SMEs expected much larger impacts – positive and negative – on market share after quota elimination. Among the 30 percent of SMEs expecting a market-share gain, the average projected change was a 33 percent increase. Among the 8 percent predicting a market-share loss, the average expectation was a 32 percent decrease.

580

90%

EXPECTATIONS FOR FIRM

Increase Stay Same Decrease

80%

77%

70%

66% EXPECTATIONS FOR TURKEY

62%

60%

54% 49%

50%

40%

54% 38%

34%

35%

x= 33.8 (17.7)

x= 38.1*** (18.4)

x= 21.1 (4.9)

30%

27%

x= 33.4*** (18.5)

x= 30.4 (18.0)

30%

19% x= 17.0 (4.5)

16% x= 28.5 (11.3)

12%

20%

x= 37.0 (10.6)

8% x= 22.5 (3.5)

x= 29.3 (12.4)

8% x= 31.6 (11.7)

4% Single score= 15

0% Overall

Fig. 3.

SME

Large

Overall

SME

Large

Market Share Expectations for Turkish Apparel and Respondent Firm in the EU Market.

LIESL RIDDLE

10%

7%

Strategy During an Industry Crisis

581

Among the 27 expecting that their ﬁrm will increase their sales to the EU market, 11 mentioned that the gain would be due to the elimination of ‘‘limitations’’ in the market. Fewer ﬁrms mentioned ﬁrm-speciﬁc reasons as to why they expected a market-share increase. Eight ﬁrms mentioned their product quality would underpin a market share gain. There were single mentions of other ﬁrm-speciﬁc attributes, such as branding, product variety, innovation, and production capacity. All the ﬁrms expecting a market-share loss in the EU cited China as the reason. Strategic Preparations Respondents were queried whether their ﬁrm had made any in a list of nine particular changes or investments in the last three years. These changes or investments included production-oriented changes, such as increasing production capacity (domestic and foreign production), integrating production capacities (e.g., enhancing facilities so that cutting, sewing, ﬁnishing, packaging, etc. are all done in one place), decreasing turnaround time (e.g., decreasing the amount of time from buyer order to receipt of goods), and upgrading production technology. Marketing-oriented changes or investments were also included, such as expanding product line, increasing sales and marketing staff, developing a branded line of apparel for export, and becoming a new licensee for a foreign company. The most common reported changes or investments are production-oriented actions (see Table 4). Almost all (93 percent) report taking steps to decrease turnaround time. Eighty-two percent claim that their ﬁrm has increased production technology and increased production capacity. All respondents whose ﬁrms maintain production facilities outside Turkey (35 percent) report that they increased their production capacity in those facilities. Respondents also reported taking marketing-oriented actions. Three fourths of respondent ﬁrms increased the number of sales and marketing staff. Eighty-one percent of respondents reported that their ﬁrms had expanded their production line to better compete in a post-quota environment. But only 23 percent reported developing a new line of branded clothing for export markets. Fifty-one percent claimed to have become a new licensee of a foreign brand. For each action, respondents were asked in which year the change or investment was initiated. In every action category, at least one third of respondents reported that the change or investment was begun in 2004 (see Table 5). SMEs were by far the majority of those undertaking these initiatives in 2004. SMEs comprised over 80 percent of those ﬁrms reporting that their

582

LIESL RIDDLE

Table 4.

Strategic Actions Taken, 2001–2004.

Action

Percent of All Firms Taking Action

Percent of SMEs Taking Action

Percent of Larger Firms Taking Action

82% 35% 68% 93% 82% 81% 75% 23%

77% 26% 67% 89% 80% 82% 74% 8%

96% 58% 65% 96% 81% 73% 73% 65%

51%

50%

50%

Increased production capacity–domestic Increased production capacity–foreign Integrated production facilities Decreased turnaround time Upgraded production technology Expanded product line Increased sales and marketing staff Developed branded line of apparel for export Became new licensee for foreign company pr0.10; pr0.05; pr0.01.

Table 5. 2004 Action-Takers. Action

Increased production capacity–domestic Increased production capacity–foreign Integrated production facilities Decreased turnaround time Upgraded production technology Expanded product line Increased sales and marketing staff Developed branded line of apparel for export Became new licensee for foreign company

Percent of All Firms Taking Action that Started Change in 2004

Percent of Firms Taking Action in 2004 that are SMEs

43% 37% 35% 48% 43% 42% 43% 35% 43%

69% 77% 83% 86% 83% 84% 84% 77% 77%

ﬁrm began to integrate production facilities, decrease turnaround time, upgrade production technology, expand their product line, or increase sales and marketing staff in 2004. Over two thirds of those ﬁrms reporting that their ﬁrm began to increase production capacity, developed a branded line for export, or were working as a new licensee for a foreign company, in 2004, were SMEs. Thus, at the beginning of 2005, Turkish apparel exporters, in general, expected market shares for their companies and Turkey as a whole to stay

Strategy During an Industry Crisis

583

the same or improve once quotas were lifted. Many ﬁrms had made strategic preparations for the post-quota competitive environment, but most had made these changes relatively recently.

TURKISH APPAREL EXPORT SALES IN 2005 At the end of the ﬁrst quarter of 2005, it appeared that Turkey’s apparel manufacturers’ expectations were correct. EmergingTextiles.com, an online trade journal, observed that ﬁrst-quarter Turkish apparel ‘‘exports are far from decelerating’’ (‘‘Turkey’s Textile and Apparel Exports Not Yet Affected,’’ 2006). During the ﬁrst quarter of 2005, knit apparel exports rose 5 percent, while woven apparel exports surged 17 percent. Despite these ﬁrst-quarter gains, the Turkish government continued to lobby the EU and the US to impose quota restrictions against China. In May 2006, the US imposed quota limits on China on three apparel categories; these quotas are set to expire in 2009. In June 2006, the EU established quota limits on Chinese apparel imports in 10 clothing categories; these quotas expire in 2008. Yet, the imposition of quotas did little to ease the increasing suffering of Turkish apparel exporters. By May, there were reports of a ‘‘looming crisis’’ in Turkey as the automotive sector overtook apparel as the country’s top export (Erdem, 2006). Turkish apparel exports not only faced steep competition from China and other countries, but they also struggled to remain competitive in light of a strengthening new lira and weakening Euro and US dollar. High domestic labor costs brought about by a 40 percent tax on workers’ wages also exacerbated the issue. Production decreased in response; May 2005 apparel production declined 21 percent from May 2004 levels (‘‘Turkey’s Exports Still Decelerating,’’ 2006). Although there was a minor rebound in Turkish apparel exports in September, a steady stagnation in export sales began in October that continued through the end of the year. Overall, in 2005, knitted apparel grew only slightly (0.39 percent), and non-knitted apparel declined fairly substantially (–4.49 percent). This performance contrasts sharply with the double-digit growth Turkish apparel experienced in previous years. But as Ismail Gulle, chairman of the Istanbul Textile and Clothing Exporters’ Union noted, ‘‘there had been predictions of a 30 percent shrinkage in exports, so to see even a small amount of increase in such a difﬁcult year such as 2005 was very important to us’’ (Erdem, 2006).

584

LIESL RIDDLE

Turkey managed to sustain most of its market share in the EU market, but did witness a slight 3 percent market-share decline in 2005. Categories of Turkish synthetic-ﬁber apparel were especially hard-hit in the EU market, many losing 50 percent or more market share in 2005. But signiﬁcant gains were observed in several categories, particularly wool apparel products. But the story was more disturbing in the American market. Imports of Turkish apparel to the US dropped dramatically in 2005, decreasing 19.24 percent in 2005. Although imports of Turkish wool apparel into the US increased 11.81 percent, substantial losses were incurred in cotton apparel (–25.29) as well as in silk and vegetable-ﬁber apparel (–22.73). In response to 2005 export performance, many Turkish apparel producers began to move operations to foreign markets, such as Jordan and Uzbekistan, to take advantage of lower labor costs and other advantages. Other ﬁrms began to seek ways to upgrade product quality and signal their higherquality production standards to potential buyers. For example, Ekoteks, a testing laboratory in Istanbul, noted that in 2005 there was a 120 percent increase in quality certiﬁcation applications among Turkish apparel producers (Erdem, 2006).

CONCLUSIONS The Turkish case provides insight into the experience of a country for which the post-quota elimination future was uncertain but for whom the apparel industry was vital in 2005. At the dawn of quota elimination, there was wild speculation about Turkey’s future. In January 2005, most Turkish apparel exporters expected that their market share in the EU and US would increase or at least stay the same in 2005. But by mid-year, it had become clear that export windfalls would not emerge, and most Turkish apparel exporters were ﬁghting for survival. At the end of 2005, the Turkish apparel industry had suffered losses in its two key export markets, and production had sharply declined. Losses were less severe in the proximal and more critical EU market, providing evidence for earlier hypotheses about the importance of geographic proximity and turnaround time in the post-quota competitive environment. But pessimism about the future of the Turkish apparel industry continues. For example, Onur Guner, assistant director of sales and marketing at Altinyildiz, one of Turkey’s leading brands, posited, ‘‘As the textile industry moved from Italy to Turkey, so now it is moving further east to China and India’’ (Erdem, 2006).

Strategy During an Industry Crisis

585

NOTES 1. In 2004 quotas were levied on 50 apparel categories in the EU and 59 in the US. Twenty countries faced quota limits in the EU; 42 nations were subject to quotas in the US (http://otexa.ita.doc.gov/). The number of quotas a country was subject to and its quantitative limit varied based on the terms of bilateral agreements between importing and exporting nations. 2. Once assigned quantitative quota limits, countries allocated quota percentages to textile and apparel ﬁrms in their country. Many nations developed elaborate decision-making mechanisms based on attributes such as export intensity, export performance, production capacity, and reputation (Dickerson, 1998). 3. Safeguards are available until 2008. Tariffs remain in place in many countries in textile and apparel categories, but, in most cases, tariff rates (average in the US is 17 percent) were not as burdensome as quotas. For example, the WTO reports that the tariff equivalent of quotas – the amount of tariff necessary to produce the same restrictive effect of quotas – is as high as 34 and 33 percent for India and China, respectively (Nordas, 2004). Post-2005 quotas will most likely remain in place for several non-WTO member nations, including Belarus, Laos (US only), Russia (US only), Serbia (EU only), Ukraine (US only), and Vietnam (Clothesource, 2004). 4. African Growth and Opportunity Act. 5. One percent of Turkish apparel exports are each sent to Canada, Switzerland, Norway, and Russia, while 8 percent is sent to other countries. 6. Out of 41 apparel-importing countries, Turkey was ranked 11th in terms of the number of quotas to which it was subjected. China was at the top of the list with 71 categories (Clothesource, 2004, p. 53). 7. Turkey vigorously lobbied the US for quota removal in exchange for their support of US efforts during the second invasion of Iraq in 2003. When Turkey refused to allow the US to deploy their troops for a northern front, the Bush administration retracted a multi-billion dollar aid package that included quota elimination for key Turkish apparel products (Ellis, 2003b). 8. Turkey’s average ﬁll rate was near the median (ranked 20th) among the 41 apparel-importing countries into the US market (Clothesource, 2004, p. 53). 9. SME is deﬁned here as a ﬁrm with less than 100 employees.

REFERENCES Abernathy, F., Volpe, A., & Weil, D. (2005). The apparel and textile industries after 2005: Prospects and choices. Working Paper No. 2005-23, School of Management, Boston University: Harvard Center for Textile and Apparel Research. American Textile Manufacturers Institute (2004). The China threat to world textile and apparel trade. Washington, DC: American Textile Manufacturers Institute. Anderson, C. D. (2000). The social consequences of economic restructuring in the textile industry: Change in a southern mill village. New York: Routledge. Appelbaum, R. P. (2004). Assessing the impact of the phasing-out of the agreement on textiles and clothing on apparel exports on the least developed and developing countries. Center for Global Studies, Paper 05. Retrieved May 10, 2004, from http://repositories.cdlib.org/isber/cgs/05

586

LIESL RIDDLE

Buckman, R. (2004, September 10). Navigating China’s textile trade. Wall Street Journal, A10. Clothesource guide to a quota-free world (2004). Charlbury: Clothesource Limited. De Jonquieres, G. (2004, July 19). Clothes on the line: The garment industry faces a global shake-out as quotas end. Financial Times, A11. Dickerson, K. G. (1998). Textiles and apparel in the global economy (3rd ed.). New York: Prentice Hall. Ellis, K. (2003a, August 26). Quota phaseout poses worry for some. Women’s Wear Daily, 2. Ellis, K. (2003b, July 1). Turkey seeks thaw. Women’s Wear Daily, 6. Ellis, K. (2005, January 3). A new trade era dawns: Global uncertainty reigns as quotas on apparel end. Women’s Wear Daily, 1. Erdem, S. (2006). Turkish industry battles growing China threat. Women’s Wear Daily, 191(6), 9. Ghemawat, P., & Baird, B. (1998). International competitiveness: Turkey and its garment industry. Harvard Business School Case, N9-799-033. Gresser, E. (2004). The big bang: Ending tariff and quota policies. Washington, DC: United States Association of Importers of Textile and Apparel. Ikenson, D. (2003). Threadbare excuses: The textile industry’s campaign to preserve import restrictions. Working Paper No. 25. Cato Institute, Washington, DC. International Trade Commission (2004). Textiles and apparel: Assessment of the competitiveness of certain foreign suppliers to the US market (Vol. 1), Investigation No. 332–448, USITC publicaton 3671. Electronic version accessed on 6 December 2005 at http:// hotdocs.usitc.gov/pub3671/pub3671.pdf Malone, S. (2004, September 28). Winners and losers. Women’s Wear Daily, 10B. Munir, M. (2004, July 23). Turks say speed will keep them ahead of China. Financial Times, 10. Murphy, R. (2001). Garment trade in Turkey estimated to have exceeded $30 bil in sales in 2000; garment industry accounts for 28% of exports, although only 10% of garments exported from Turkey carry a Turkish brand name. Women’s Wear Daily, 181(45), 12. Murphy, R. (2002, July 16). Turkey aiming for higher proﬁle. Women’s Wear Daily, 20. National Council of Textile Organizations (2004). The China threat to world textile and apparel trade [An electronic version]. Washington, DC. Accessed on 5 December, 2005 at http:// www.ncto.org/newsroom/chinarpt.pdf Neidik, B. (2004). Organizational foundations of export performance: The case of the Turkish apparel industry. Journal of Fashion Marketing and Management, 8(3), 279–299. Nordas, H. K. (2004). The global textile and clothing industry post the agreement on textiles and clothing. WTO Discussion Paper, Geneva, Switzerland: World Trade Organization. Ostroff, J. (1999, September 9). Turkish textile talks set on post-quake quotas. Women’s Wear Daily, 18. Owen, R., & Pamuk, S. (1999). A history of the Middle East economies in the twentieth century. Cambridge, MA: Harvard University Press. Padhi, A., Pauwels, G., & Taylor, C. (2004). Freeing India’s Textile Industry [Special Edition: What Global Executives Think]. McKinsey Quarterly, 9–11. Riddle, L. A. (2001). The social embeddedness of export promotion organization in the Turkish clothing industry. Ph.D. thesis, The University of Texas at Austin. Riddle, L. A., & Gillespie, K. (2003). Information sources for new ventures in the Turkish clothing export industry. Small Business Economics, 20(1), 105–120. Rossen, E. I. (2004). The globalization of the US apparel industry: Making sweatshops. Berkeley, CA: University of California Press.

Strategy During an Industry Crisis

587

Smeets, M. (1995). Main features of the Uruguay round agreement on textiles and clothing, and implications for the trading system. Journal of World Trade, 29(5), 97–110. Smid, S., & Taskesen, F. (2002). Textile, apparel, and leather sector in Turkey. Amsterdam: PWC Consulting. Turkey’s exports still decelerating. Retrieved January 14, 2006, from http://www.emergingtextiles.com Turkey’s textile and apparel exports not yet affected. Retrieved January 14, 2006, from http:// www.emergingtextiles.com Wilson, E. (1999, August 16). Turkish delight. Women’s Wear Daily, B2. Zarocostas, J. (2004, September 28). The Istanbul crusade. Women’s Wear Daily, 20B. Zarocostas, J. (2005, March 22). Safeguard showdown. Women’s Wear Daily, 18B.

This page intentionally left blank