Contracts in Trade and Transition
This page intentionally left blank
Contracts in Trade and Transition The Resurgen...
29 downloads
817 Views
1MB 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
Contracts in Trade and Transition
This page intentionally left blank
Contracts in Trade and Transition The Resurgence of Barter
Dalia Marin and Monika Schnitzer
The MIT Press Cambridge, Massachusetts London, England
( 2002 Massachusetts Institute of Technology All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher. This book was set in Palatino in 3B2 by Asco Typesetters, Hong Kong. Printed and bound in the United States of America. Library of Congress Cataloging-in-Publication Data Marin, Dalia. Contracts in trade and transition : the resurgence of barter / Dalia Marin and Monika Schnitzer. p. cm. Includes bibliographical references and index. ISBN 0-262-13399-7 (hc. : alk. paper) 1. Barter. 2. Countertrade. 3. BarterÐEurope, Eastern. 4. CountertradeÐEurope, Eastern. 5. International trade. I. Schnitzer, Monika. II. Title. HF1019 .M37 2002 3320 .5Ðdc21 2001044323
to Martin to Klaus, Klara, Hanna, and Laura
This page intentionally left blank
Contents
List of Tables List of Figures Preface xiii
ix xi
1
Introduction
1
I
Barter Contracts in International Trade
2
Puzzling Facts and Popular Explanations
3
Creating Collateral
4
The Export Pattern: Paying with Goods
5
The Import Pattern: Securing Technology Transfer Appendixes to Part I
17
35 61 77
93
II Contract Enforcement in Transition Economies 6
Stylized Facts and Competing Explanations
103
7
Arrears and Barter: Extending Trade Credits
8
The Output Decline: The Lack of Trust and Liquidity
9
Confronting the Data
153
Appendixes to Part II
167
127 137
viii
Contents
10 Contracts in Trade and Transition: An Outlook Notes 185 References 193 Index 201
173
List of Tables
Table 1.1
Cross-country pattern of barter in transition economies 4
Table 1.2
Estimates of countertrade in world trade
Table 2.1
Contract type and compensation ratio
Table 2.2
Pattern of trade with Eastern Europe in 1987
Table 2.3
Export promotion
Table 2.4
Contract type and region
Table 2.5
Market distortion and compensation ratio
Table 3.1
Value of collateral
Table 4.1
Ranking of commodity money
Table 4.2
Liquidity of commodity money
Table 4.3
Choosing collateral goods
Table 4.4
Creditworthiness and trade pattern
Table 5.1
Counterpurchase versus barter
Table 5.2
Determining the degree of tying
Table I-B
De®nition of variables and sample statistics
Table 6.1
Barter in transition economies
105
Table 6.2
Barter and ®rm characteristics
107
Table 6.3
Commodity composition of barter deals
Table 6.4
Ownership, debt, and barter
Table 6.5
Barter, arrears, and ef®ciency
Table 6.6
Barter, arrears, and ef®ciency by ®rm size
Table 6.7
Terms of trade of the non-cash economy
Table 6.8
Pricing behavior of sectors
5 20 21
24 27 28
58 66 68
70 74
88 91 97
107
110 113
116
114 115
x
List of Tables
Table 6.9
Is Russia's economy virtual?
120
Table 6.10
Motives for barter
Table 7.1
Barter as credit
Table 7.2
Trade credit
Table 7.3
Barter and creditworthiness
Table 7.4
Barter, arrears, and liquidity squeeze
Table 7.5
Solving creditworthiness in cash deals
Table 7.6
Inverse U-curve
Table 8.1
Lock-in
Table 8.2
Close business ties
Table 9.1
Output decline in the former Soviet Union
Table 9.2
Terms of trade effects
Table II-B
De®nition of variables and sample statistics
Table 10.1
Debt to GDP ratio in selected countries
Table 10.2
Growth of real GDP in transition economies, 1990 to 1997 176
Table 10.3
Institutional structure of East±West trade
Table 10.4
Macro- and microeconomic determinants of barter 181
123 128
129 130 132 133
134
138 138 156
160 168
174
178
List of Figures
Figure 1.1
Time pattern of barter in Russia
3
Figure 3.1
Time structure
41
Figure 5.1
Time structure
82
Figure 6.1
Output decline in transition economies
Figure 6.2
Total arrears as percentage of annualized GDP in Russia 104
Figure 8.1
Time sequence at production step 1
104
142
This page intentionally left blank
Preface
Contracts are essential for many economic activities. Any economic relationship that involves some quid pro quo needs to be supported by some contractual arrangement to make sure that the parties involved abide to their obligations. But contracts are worth only as much as their enforcement can be relied upon. Dif®culties in contract enforcement are an important impediment to international transactions in the world economy and to domestic transactions in transition economies. In international trade, national sovereignty interferes with contract enforcement because national borders demarcate national jurisdictions. Such demarcations segment markets and impose severe transaction costs on exchanges across national jurisdictions. The hazards involved in international transactions are often disregarded, but they make headlines each time a sovereign debtor threatens to stop servicing its debt, as it happened in the international debt crisis in the 1980s or in the Russian ®nancial crisis in 1998. In transition countries, poorly developed legal and ®nancial institutions make contract enforcement unreliable and impose severe transaction costs on any economic activity. These costs become prohibitively large in times of historic change and revolution. Unstable business partner relationships and rapidly changing social norms limit the extent to which economic exchanges can be sustained by reputation, by repeated interaction or by embedding them in social networks. This book is about how barter as an economic institution can help deal with the problem of contract enforcement across national borders in international trade and within borders in transition economies. The book is the result of almost a decade of joint research. We started this collaboration with the intention to write a short paper
xiv
Preface
addressing one rather speci®c question. But, in the process of writing the paper, we touched an increasing number of issues that seemed worth exploring in greater detail. So one paper led to many others. Over time, a broader picture of how to interpret countertrade and barter emerged from our work, and we felt it would be worthwhile to put all these explanations in a broader perspective. So when Terry Vaughn, who at that time was the economics editor at The MIT Press, encouraged us to write a book on this subject, we were very enthusiastic to pursue this idea. From this ®rst idea of writing a book to actually ®nishing the manuscript took much longer than we had originally planned. Moving places and new responsibilities, both professional and private, delayed the project. Most important, however, in the wake of the political and economic changes in Eastern Europe since 1989, domestic barter trade became a dominant feature of the transition from the plan to the market in many of the transition economies. This development was quite puzzling to economic experts on transition. In 1996, when one of us was invited to a World Bank conference in Kiev to consult on barter, we quickly realized that forces similar to those driving barter in international trade are at work in transition economies. Several discussions with Daniel Kaufmann of the World Bank, who at that time worked at the Harvard Institute for International Development (HIID) in Cambridge, made us con®dent that we would be able to provide new insights on barter in transition economies. We conducted a joint survey with Daniel Kaufmann and with the help of the local of®ce of HIID in Kiev and one of our students, Bogdan Gorochowskij, to collect data on barter in transition economies. What started as a project of writing a book on international countertrade turned into a much larger project on contracts in trade and transition. Over these years, we have been deeply in¯uenced by discussions and collaborations with colleagues, coauthors, and friends. Dalia Marin owes much of her knowledge in international trade to Elhanan Helpman from whom she learned how exciting economics can be when he visited the Institute for Advanced Studies in Vienna in the mid-1980s to teach international trade. Since then, over the many years she exposed the ideas of the book to Elhanan's critical and wise judgment. Richard Caves initiated the research on countertrade when he visited the Institute for Advanced Studies in Vienna in 1987. Dalia Marin and Richard Caves started to work on countertrade and
Preface
xv
price discrimination. From this collaboration, and the joint work with Erwin Amann on countertrade and risk sharing, it became apparent that several of the stylized facts of international barter remained unexplained. This is how contracts and incentives became the major research agenda to which this book is devoted. Monika Schnitzer is highly indebted to Klaus Schmidt and Georg NoÈldeke. Innumerable discussions with them inspired her how to put contract theory to use, and their untiring critical questions very much contributed to making the analysis rigorous. She also bene®ted greatly from the unique research environment she experienced while working at Urs Schweizer's Institute at Bonn University. We are highly indebted to Daniel Kaufmann for inspiring discussions and enjoyable joint research. His expertise in transition economics in¯uenced the development of part II of the book. We are especially grateful to him for letting us borrow from the joint research with him in the writing of chapters 6 and 7. This book also owes much to Rachel Kranton. Rachel gave very detailed and thoughtful comments on the ®rst draft of our book which greatly helped to improve the exposition of our ideas. Furthermore the numerous discussions with her in the faculty research seminar of the University of Maryland or while walking in the beautiful forest in the Maryland area helped to clarify many of the ideas developed in part II of the book. In the course of so many years on this project, we have incurred debts to many other friends and colleagues. The research presented in part I of this book bene®ted from discussions and comments from Richard Baldwin, Avinash Dixit, Tore Ellingsen, Jonathan Eaton, Raquel Fernandez, Harry Flam, Avner Greif, Elhanan Helpman, Oliver Hart, Paul Krugman, Albert Ma, Georg NoÈldeke, Peter Neary, Martin Rein, Michael Riordan, Dani Rodrik, Lars-Hendrik RoÈller, Klaus Schmidt, and Urs Schweizer. The research presented in part II bene®ted from discussions and comments from Philippe Aghion, RuÈdiger Ahrend, Erik BergloÈf, Guillermo Calvo, Jiahua Che, Peter Clark, Simon Commander, Theo Eicher, Horst EidenmuÈller, Tore Ellingsen, Guido Friebel, Irena Grosfeld, Sergei Guriev, Elhanan Helpman, Bengt HolmstroÈm, Haizhou Huang, Barry Ickes, Michael Keren, Anna Meyendorff, Christian Mummsen, Peter Murrell, GeÂrard Roland, Luis Sanchez, Mark Schankerman, Paul Seabright, Claudia Senik-Leygonie, Judy Thornton, Ksenia Yudaeva, and Jeromin Zettel-
xvi
Preface
meyer. To this list, we should add the various anonymous reviewers who read a previous draft of this book. We thank them for all their comments and encouragement. Our joint research was started when Dalia Marin visited Harvard University and Monika Schnitzer visited Massachusetts Institute of Technology and Boston University, and it bene®ted from further research visits to Stanford University and the Science Center Berlin. We are grateful for the hospitality and stimulating environment we enjoyed at these places. We also bene®ted from the support of our home institutions, the Institute for Advanced Studies in Vienna, Humboldt University Berlin, University of Bonn, and University of Munich, and from the intellectual diversity of these institutions. Our research was supported by the German Science Foundation through Sonderforschungsbereich 303 at University of Bonn, and through grants Ma 1823/2-1, Ma 1823/2-2, Schn 422/2-1, and Schn 422/2-2, and by a Erwin Schroedinger Grant of the Austrian Science Foundation. The survey on international barter that was conducted in Austria was supported by the JubilaÈumsfonds of the Austrian National Bank. Angela KoÈppl, who at that time was a research assistant at the Institute for Advanced Studies in Vienna, was important for the success of the survey. The survey on barter in transition economies was conducted in Ukraine and was supported by the Harvard Institute for International Development. Special thanks go to Januz Szyrmer of the Kiev of®ce of the Harvard Institute for International Development and to Vira Nanivska of the International Center for Policy Studies in Kiev for logistic support for the survey among ®rms in Ukraine. The discussions with Luis Sanchez, who at that time worked as a consultant to the World Bank in Ukraine, helped us in adopting our ideas to the circumstances in Ukraine. We also thank Daniel Bauchet, Bogdan Gorochowskij, Alexis Giesen, and Thomas MuÈller for valuable research assistance for part II of the book. We are grateful to Terry Vaughn for encouraging us to write this book for The MIT Press and to Elizabeth Murry and her team from The MIT Press for their support in realizing and completing the book. The greatest thanks go, of course, to our families, who make it all worthwhile. To them we dedicate this book.
Contracts in Trade and Transition
1
Introduction
Barter trade has received much attention lately. But it is not a new phenomenon. In the 1980s, in the aftermath of the international debt crisis, barter became prevalent in international trade with developing countries and Eastern Europe. Since the 1990s, with the domestic debt crisis in transition economies, barter has continued to be a dominant phenomenon in domestic trade in these countries. What explains the appearance of barter in international trade in the 1980s and in domestic trade in transition countries in the 1990s? What makes barter, or countertrade, as it is more generally called, preferable to conventional forms of trade? In this book we will argue that in both environments, in international trade and in transition economies, contract enforcement is problematic, and hence conventional contracts cannot be relied on as the main mechanism to sustain economic exchange. As we will show, barter and countertrade can be explained as an institutional response to such contractual problems arising in imperfect capital and goods markets. Before we set out to explore these issues, we need to clarify the terminology used throughout this book, since the terms used in practice and in writings vary greatly. In particular, the term ``barter`` is used with different meanings, sometimes as a term that refers to countertrade transactions in general, sometimes in a more speci®c sense. In part I, in dealing with international trade, we will use the term ``countertrade'' as the general term to denote transactions in which a party from an industrialized country supplies goods, services, or technology to a second party in an Eastern European or developing country, and in which, in return, the ®rst party purchases from the second party an agreed amount of goods, services, or technology. Our terminology will take the point of view of the second
2
Introduction
party, and so the ®rst transaction will be called ``import'' and the second transaction ``export.'' A distinctive feature of countertrade is the existence of a link between the two transactions, the import and the subsequent export. Countertrade transactions take a variety of forms. The three forms we will distinguish throughout this book are ``barter,'' ``counterpurchase,'' and ``buyback.'' Barter in the strict sense of the word refers to an import that is paid entirely or partly with an export from the latter country without using foreign exchange. Counterpurchase refers to a transaction in which the import is paid with foreign exchange but the industrialized country commits to buy export goods from the Eastern European or developing country in return. Buyback is a transaction in which the ®rst party supplies a production facility and the parties agree that the supplier of the facility will buy goods produced with that production facility. All three forms of countertrade are frequently observed in international trade. In part II, in dealing with domestic barter in transition economies, we refer to transactions where the term barter is used in the strict sense of the word. This means a bilateral exchange of goods and services, without the use of money as a medium of exchange. Under central planning, countertrade was especially observed in international trade among CMEA (Council for Mutual Economic Assistance) countries as well as in East±West trade. Before 1989 barter and countertrade accounted for up to 40 percent of total trade between East and West.1 After 1989, domestic barter in Russia exploded after macroeconomic stabilization in 1994 from 5 percent of GDP to 60 percent in 1998. In Ukraine the share of barter in industrial sales is estimated to have been 51 percent in 1997.2 The survey of 200 ®rms by the Russian Economic Barometer (1999) since 1992, in turn, suggests that noncash payments rose steadily from 8 percent in 1992 to 54 percent in mid-1998 (see ®gure 1.1). Since the ®nancial crisis in August 1998, barter and the use of other money surrogates have started to decline, accounting for 43 percent of sales of industrial ®rms in February 1999. The importance of barter varies across transition economies. Table 1.1 gives ®gures of the World Business Environment Survey in twenty transition economies and shows that Croatia exhibits the highest share of barter of 33 percent in 1999. Russia and Ukraine show a barter share of about 24 percent and Central European
Introduction
Figure 1.1 Time pattern of barter in Russia. Source: Russian Economic Barometer, various years, Moscow.
3
4
Introduction
Table 1.1 Cross-country pattern of barter in transition economies (% of sales) 1996
1999
Percent change
Armenia
2.9
2.9
0.0
Azerbaijan
5.1
4.0
ÿ21.6
Belarus
13.1
13.9
6.1
Bulgaria
4.0
4.2
5.0
Croatia
21.7
32.8
51.2
Czech Republic Estonia
3.8 5.5
3.3 4.1
ÿ13.2 ÿ25.5
Georgia
6.8
5.2
ÿ23.5
Hungary
1.7
0.8
ÿ52.9 ÿ13.5
Kazakhstan
20.7
17.9
Kyrgyzstan
16.5
17.4
5.5
Lithuania
3.1
2.8
ÿ9.7
Moldova
29.6
26.3
ÿ11.1
Poland Romania
3.9 8.6
4.7 7.3
20.5 ÿ15.1
Russia
23.5
24.1
2.6
Slovakia
19.2
19.2
0.0
Slovenia
17.4
16.3
ÿ6.3
Ukraine
20.3
24.0
18.2
Uzbekistan
23.2
10.2
ÿ56.0
Total
12.5
12.1
ÿ3.2
Source: World Business Environment Survey, World Bank-EBRD 1999.
countries like Hungary, Poland, and the Czech Republic have barter shares between 0.8 and 4.7 percent. It is also interesting to see that some of these countries experienced an increase in the importance of barter over time, like Croatia, and Ukraine, while barter declined in Uzbekistan, Kazakhstan, and Moldova. In Uzbekistan, the fall of barter was particularly pronounced.3 In international trade a similar boost of countertrade was experienced in the 1980s, when the international debt crisis led to a dramatic decline in private lending to developing and Eastern European countries. Even though debtors hesitated to invoke total repudiation, commercial banks were reluctant to provide new loans, skeptical that they would ever be repaid in full.4 As highly indebted countries found it increasingly dif®cult to ®nance their imports, unconventional forms of trade and trade ®nancing experienced a resurgence,
Introduction
5
Table 1.2 Estimates of countertrade in world trade Source IMF (1980, 1984) Business International (1983) GATT (1984) Economic Intelligence Unit (1984) OECD (1985)
Barter and countertrade in percent of world trade 1±10 10 8 15±20 4±5
Group of Thirty (1985)
8±10
UNCTAD (1987/1988)
15±20
ifo Institut (1988)
10±12
including countertrade. Bussard (1987) reports that the number of countries engaged in countertrade rose from 27 in 1979 to 88 in 1984. Likewise the number of countertrade transactions that was reported by a group of survey respondents increased on average by 50 percent between 1980 and 1981, by 64 percent between 1981 and 1982, and by 117 percent between 1982 and 1983.5 Hammond (1990) observes that precedents of this striking comovement of debt and countertrade can be found in the late nineteenth century and in the depression of the 1930s. Table 1.2 summarizes different estimates of countertrade in world trade. It highlights the importance of countertrade in the international trade of the 1980s. More recent estimates suggest that international countertrade has not declined in East±West trade.6 What explains these two waves of countertrade and barter in the last two decades of the twentieth century, in international trade in the 1980s and in domestic trade in transition economies in the 1990s? What are the common themes of trade and transition that help us to understand why ®rms turn in large numbers to unconventional trade forms of notoriously ill repute? Why would parties in a fully monetized economy want to pay in goods rather than money? Is it indeed, as many experts have argued, a regression to bilateralism and reciprocity, or can we ®nd ef®ciency reasons for this type of trade? Our view is that we need to take an institutional approach in order to explain barter and countertrade. Traditional trade theory does not provide an answer to these questions. In this book we introduce contracts and institutions into trade theory and explain barter and countertrade as institutions that are designed to deal with contrac-
6
Introduction
tual problems. Thus the answer we propose to the questions raised above is that parties might want to pay in goods rather than cash or link an import with an export as in countertrade if, by doing so, they can solve incentive problems that otherwise would prevent any trade from taking place. This institutional approach will help us see the common themes of international trade and economic transition that explain the emergence of barter and countertrade in the 1980s and the 1990s. In both environments, contract enforcement is problematic, and hence conventional contracts cannot be relied on as the main mechanism to sustain economic exchange. In international trade, the enforcement of contracts is in the hand of national authorities. If, for whatever reason, these authorities fail to perform this function, trade partners have no higher instance to which they can turn to enforce the law due to the lack of supranational authorities. Similarly, in transition economies, relying on contracts is problematic as long as state authorities have not established a functioning legal system, and hence contract enforcement through state authorities cannot be counted on.7 If contract enforcement through state authorities cannot be relied on, trade partners have to seek other means of protecting their interests in business transactions. One possible substitute for formal contracts would be relying on reputation as an enforcement mechanism. Trade partners voluntarily stick to the terms of the contract if the risk of losing their reputation and of not being able to do business in the future is a suf®ciently large punishment for misconduct. However, the extent to which reputation can help as an enforcement mechanism is limited. In the context of international trade, Bulow and Rogoff (1989a) have shown that reputation may fail to induce sovereign countries to repay their debt. And even if this reputation mechanism sustains some international trade, ®rms may look for means to overcome the limits set of this mechanism. In transition economies the problem is even more conspicuous, since in times of historic change the future is uncertain and business partners may not know whether they or their partners will be in business in the future for a variety of reasons. Thus, in both environments, trade partners may look for alternative means to enforce contracts and to sustain their economic activities. If contract enforcement is weak, problems may arise on both ends of a business transaction: the seller may fail to deliver the good, and
Introduction
7
the buyer may fail to pay for the goods. If buyers have no cash to pay, and thus face liquidity constraints at the time of delivery, the business transaction can take place only if the seller can trust the buyer to pay in due course. On the other hand, the buyer is willing to engage in a business transaction only if she can trust the seller to deliver the right goods. Both problems are prevalent in trade and transition. In each environment, enforcing the payment of goods can pose serious problems. In the aftermath of the debt crisis in the 1980s, highly indebted countries were liquidity constrained and could not ®nance necessary imports. Given their level of indebtedness, debt repayment could not be relied on. The debtor country could create more liquidity by not repaying its debt rather than receive a new loan. Similarly ®rms in the former Soviet Union are highly indebted, vis-aÁ-vis each other, vis-aÁ-vis the state and vis-aÁ-vis their workers. The phenomenon of inter-®rm tax and wage arrears exploded in the 1990s, making it more likely that these ®rms would repudiate on their payments. As we will argue, in both environments, countertrade and barter emerged as institutions to deal with the lack of creditworthiness of countries and ®rms. Barter and countertrade introduce a deal-speci®c collateral that serves to protect the interests of the creditor for one particular business transaction. In this way barter and countertrade can mitigate the contractual hazards associated with indebtedness. There are also important problems arising on the seller's side. In international trade, the most conspicuous example is the technology transfer problem. It is often reported that explicit contracts cannot be relied on to make sure that developing countries receive the advanced technology as promised. These countries often complain that ®rms from industrialized countries sell inferior technology to them, a technology that is outdated and cannot be sold on Western markets. In transition economies, the dominant problem for ®rms seems to be to receive the inputs they need in order to produce. Often they need to invest in ®nding the right input supplier and then face the risk that the seller will exploit the fact that there are no other sellers around from which the ®rm can buy. In market economies long-term relationships and explicit contracts help to protect the ®rm's investment. But in transition economies these alternatives are less reliable, as we have just argued. As we will show, the very lack of liquidity that makes it dif®cult to ®nance imports and input purchases can actually help when it comes
8
Introduction
to dealing with problems on the supplier's side. We ®nd evidence for this in both contexts, international trade and transition economies, even though the actual mechanics are very different. This goes to prove that in an imperfect world in which contract enforcement is weak, like the former Soviet Union or imperfect capital markets, something that seems to be worseÐthat contractual problems arise on both sides of the business transaction rather than only on one sideÐcan improve contract enforcement. In international trade, the liquidity constraint helps to solve the technology transfer problem. In transition countries, ®rms can take advantage of their lack of liquidity to protect themselves against sellers who would cheat them by not providing the correct input. Our analysis of countertrade and barter proceeds on two fronts. We ®rst develop a theoretical framework in order to explain barter and countertrade as optimal institutional responses to contractual hazards and incentive problems in international trade and economic transition. Part I of the book deals with the problems of contract enforcement in international trade, part II with the problems of contract enforcement in transition economies. Second, we confront these theoretical explanations with empirical evidence. For this purpose we draw on two unique data sets. The empirical analysis of part I examines a sample of 230 countertrade contracts signed by ®rms from OECD countries and trade organizations in Eastern Europe or developing countries in the period 1984 to 1988, when industrialized countries were struggling with the consequences of the international debt crisis. In part II our empirical study is based on a survey among ®rms in Ukraine in 1997, from which we obtained information on 165 domestic barter deals. Both data sets contain contract speci®c information that allows us to test for the contractual hazards put forth in our theory. The issues we deal with in this book have received relatively little attention in the literature on international trade and in the reform programs of transition countries. In traditional trade theory, contracts and institutions as a determinant of trade do not ®gure prominently. A notable exception is Greif, Milgrom, and Weingast (1994) who analyze merchant guilds in the middle ages as an ef®ciency enhancing institution to deal with moral hazard problems in ``international'' trade.8 The main reason for this neglect seems to be that contract theory and international trade theory belong to two strands of the literature with separate
Introduction
9
traditions. Traditional trade theory is based on general equilibrium with perfect competition. In the last ®fteen years imperfect competition, product differentiation, and economies of scale have been incorporated broadening the range of insights considerably. But the treatment of the decision whether to organize an activity within the ®rm or via the market, which is relevant for such international business activities as foreign direct investment, joint ventures, alliances, barter, countertrade and other forms of international partnerships, has remained limited. The ``internalization question'' belongs to the domain of contract theory, a theory that has made much progress in recent years (Hart 1995; HolmstroÈm and Tirole 1989; Tirole 1999). As we will show in this book, much is to be gained by merging these two ®elds. Incentives and economic institutions can play an important role in explaining international trade beyond the known determinants of trade like factor endowment, productivity, preferences, imperfect competition, and economies of scale. Similar to the way the incomplete contracts literature revolutionized our thinking about ``institutional arrangements'' like ®rms, we expect it to revolutionize our understanding of ``institutional arrangements'' in international trade. As we will see, this contractual approach is particularly relevant for the explanation of international business activities such as barter and countertrade. But it will also in¯uence the way we think about international capital and goods ¯ows in general, and help us to understand why it takes particular forms like foreign direct investment, portfolio investments, joint ventures, or international alliances. In the reform programs of transition countries, contractual issues have been thoroughly ignored. As Clement and Murrell (2000) note, Leszek Balcerowicz's description of his early reform program in Poland contains hardly any reference to legal and institutional reform (Balcerowicz 1994). Similarly small is the emphasis Vaclav Klaus puts on the rule of law in his discussion of economic reforms (Klaus 1997). And Yegor Gaidar in his description of Russian reform does not mention institutional reform at all (Gaidar 1995). Ten years after the fall of communism, it becomes apparent that one of the biggest mistakes in the assessment of how to organize the transition from the plan to the market was the strong emphasis on stabilization policy, price, and trade liberalization and privatization, while the importance of legal and institutional issues was underrated. Today it seems that economic experts consulting on the transition reforms
10
Introduction
have underestimated the importance of microeconomic and institutional issues relative to macroeconomic issues.9 Contract enforcement problems and nonfunctioning legal institutions have become the key issue in explaining differences in the performance in transition countries. We will look at the differences in contract enforcement problems that can explain why the economic performance in Russia and the former Soviet Union was so much worse than that of the early transition countries. Outline of the Book In part I we consider the problems of contract enforcement in international trade. We start by presenting in chapter 2 what may be called the ``stylized facts'' of countertrade and discuss some popular explanations. The following chapter 3 studies the problem of enforcing debt repayment if the importing country is liquidity and credit constrained. In this chapter we show why it makes a difference if trade partners promise a payment in goods rather than a payment in cash. The basic idea is that goods can be used as a deal-speci®c collateral for the ``credit'' granted for the original trade. Chapter 3 also confronts the theoretical predictions with empirical evidence on actual countertrade contracts. We show that the size of the deal-speci®c collateral will be the larger the less creditworthy the country and the less reputation can be relied on as an enforcement mechanism. Our argument that payments in kind may have advantages over payments in cash contradicts the conventional wisdom in the theory of money. The common view is that barter is inef®cient because it does not overcome the double coincidence of wants problem as money does. A seller may need to accept goods for which he has no use himself. Moreover payments in goods are problematic, since it is dif®cult to judge the quality of the goods offered as means of payment. In chapter 4 we rank goods with respect to their liquidity and anonymity properties, and we show that these incentive problems have natural implications for which type of goods qualify as means of payment in international countertrade. This will explain the export pattern of countertrade. We ®nd evidence that in contrast to the descriptive literature on countertrade, ``new'' goods may be chosen as export goods in countertrade because these goods are speci®cally equipped to mitigate contractual hazards.
Introduction
11
In chapter 5 we consider the import pattern of countertrade and thus the problems that arise on the seller's side of a transaction. We show that imports from industrialized countries are predominantly technology goods. We argue that tying an import to an export induces the technology seller in an industrialized country to deliver high-quality technology, even though the original import and the subsequent export are not technologically related. Interestingly we ®nd that the lack of creditworthiness of the technology-importing country can help give the technology supplier the right incentives. Introducing a second transaction in the form of an export to an industrialized country can serve as a hostage that deters cheating on the quality of the technology. This contractual arrangement makes the technology supplier internalize the externality her technology has on the Eastern European or developing country. Thus countertrade is a ®rst-best substitute for foreign direct investment when these countries are reluctant to give access to foreign ownership in their markets. In part II of the book, we turn to contract enforcement in transition economies and thus to the phenomenon of domestic barter. In chapter 6 we start by reporting some ``stylized facts'' of transition economies that distinguish the development in the countries of the former Soviet Union from that in the early transition economies. It seems that the countries of the former Soviet Union have experienced a more pronounced output decline, have larger inter-®rm debt, and have more domestic barter trade than the early transition economies in Central Europe. We ®rst trace the connections among the output decline, arrears, and barter, discuss some competing explanations for these phenomena, and confront them with data on bartering ®rms in Ukraine. In chapter 7 we turn to the question why ®rms are willing to give loans to other ®rms in form of trade credits when the banking sector is reluctant to provide capital. As we show in this chapter, the possibility of undertaking a business in form of barter trade becomes important in this context. As in international trade, one special advantage barter trade offers in transition economies is that it can be used to collateralize trade credits. In chapter 8 we present a theoretical framework that explains the output decline in the former Soviet Union with problems that arise on both the seller's and the buyer's side. On the seller's side we discuss the ``lack of trust'' problem, or, to use the terminology of con-
12
Introduction
tract theory, the ``holdup problem.'' A ®rm that needs to invest in ®nding the right supplier in order to produce will potentially be exploited by the input supplier when there are no or only a few alternative suppliers. This problem has been discussed in the transition literature under the heading of ``disorganization.'' Disorganization arises when old relationships break down before new ones can be established, as is the case in the former Soviet Union when the system of centralized planning was abolished. In such a ``no future'' environment, ®rms cannot trust their input suppliers and cannot rely on legal support to protect their interests. Output collapses due to an input shortage. On the buyer's side we discuss the ``liquidity problem.'' Firms are liquidity constrained and banks in transition countries are reluctant to provide capital. In order to alleviate the ®nancial squeeze, ®rms turn to other ®rms for trade credits. This has led to the phenomenon of inter-®rm arrears in transition economies. Interestingly we ®nd that when both problems are present, the holdup problem on the seller's side and the credit problem on the buyer's side, one can help with the other, rather than making things worse. More speci®cally, we ®nd that inter-®rm arrears can alleviate contract enforcement problems in the economies of the former Soviet Union. When the input buyer is short of cash, the fact that the input seller has to worry about getting paid gives the input buyer bargaining power. This bargaining power reduces the supplier's ability to exploit the input buyer's need for the input. We also show that when the liquidity problem gets very large and thus credit enforcement becomes too costly for the supplier, barter can be used as a commitment device by the buyer not to exploit his bargaining power. This way barter provides a mechanism to deal with disorganization when the problem of credit enforcement becomes very large. Chapter 9 confronts our theoretical predictions with data on bartering ®rms in Ukraine. It summarizes the links between output decline, inter-®rm arrears and barter and shows how our theory is able to explain the cross country as well as the time pattern of arrears and barter in transition countries. Chapter 10 is the concluding chapter of our book in which we speculate about the contracts that may arise in the future to deal with the capital transfer and technology transfer problem. We establish that for many countries creditworthiness has declined and income gaps between poor and rich countries have widened in the last dec-
Introduction
13
ade, creating an even stronger demand for institutions to deal with these problems than in the past. Based on our analysis of the 1980s, we speculate that foreign investment is likely to replace countertrade as a contract in international trade. The perspective of foreign investment as an institution to evolve has shifted since the fall of communism. Under central planning foreign investment was rare because of a political ownership constraint. Foreign ®rms were not allowed to own assets in these countries. Today foreign investment is increasing surprisingly slowly because of an uncertainty constraint. Foreign ®rms are reluctant to invest in these countries because of insecure property rights. For transition economies we are less optimistic. We discuss the long-term costs of barter in the danger of an institutional trap that hinders the banking sector from developing. This might make it more dif®cult to remonetize the economies of the former Soviet Union. We also discuss the role of monetary and exchange rate policy in a barter economy such as Russia's as well as policy lessons learned from the ®nancial crisis in Russia.
This page intentionally left blank
I
Barter Contracts in International Trade
This page intentionally left blank
2
Puzzling Facts and Popular Explanations
In this chapter we set the stage for our analysis of countertrade by reporting what may be called the ``stylized facts.'' Further we review the explanations offered in the literature of this reciprocal form of exchange. In a countertrade transaction, an Eastern European or developing country imports from an industrialized country and the industrialized country commits to buy export goods. The three main forms of countertrade are barter, counterpurchase, and buyback. In barter no foreign exchange is used. The import is paid for with the export. In counterpurchase and buyback each ¯ow is paid for in foreign exchange. With buyback there is also a technological relationship between the original import and the subsequent export, one ¯ow consisting of plants, technology, or machinery and the other of output produced with this equipment. Even though most experts agree that in the aftermath of the debt crisis, countertrade accounted for a large and rapidly growing proportion of international trade (according to most estimates between 10 and 20 percent of world trade, as reported in table 1.2), it is dif®cult to obtain reliable data on the exact volume and features of these transactions. The problem is that countertrade does not ®gure in of®cial trade statistics. Furthermore governments have an incentive to downplay the extent of countertrade because the use of it is regarded as a potential trade barrier by GATT. In order to obtain data on the speci®c features of countertrade arrangements, we carried out a survey among countertrading companies located in Austria (the majority of which are trading companies or subsidiaries of multinationals). 2.1
The Data
We conducted a survey in Austria by personal interviews among ®rms or trading companies specializing in countertrade. In the 1980s,
18
Barter Contracts in International Trade
before the fall of communism, about 40 percent of East±West trade was conducted in the form of countertrade. Because of its large exposure to the East, Austria has become a center with experience and accumulated knowledge in countertrade. The sample consists of 230 countertrade contracts signed by ®rms from OECD countries and trade organizations in Eastern European or developing countries in 1984 to 1988, and it contains detailed information on each contract. The companies included are ®rms producing in Austria, or subsidiaries of multinational enterprises with their own in-house countertrade division located in Austria, or other ®rms in OECD countries using an international trading ®rm in Austria to carry out the barter transaction. Each ®rm was asked for information on about 40 dimensions of each countertrade contract. Among these dimensions are the parties in the East and West involved in the deal, the goods exchanged, the size of the deal, the competitive conditions for the products exchanged, the existence of a futures market for the goods exchanged, duration of the contract, ®x versus ¯ex price contracts, quality and degree of differentiation of the goods exchanged, and history and quality of the relationship between the trade partners in East and West. We obtained on average information on three deals. Very large ®rms provided us with information of up to ten deals. Thirty percent of the sample are ®rms in the European Community, and 62.7 percent in other industrialized countries including Austria, Sweden, Japan, and the United States. The remaining 7.3 percent are coming from another region outside the OECD. The countertrade partner in Eastern Europe or developing country was a state agency (85.2 percent of the cases), a state-owned enterprise (9.1 percent), or a private ®rm (5.7 percent). Because of Austria's proximity to Eastern Europe, East±West countertrade accounts for 86.5 percent of all deals in the sample. More speci®cally, 14.8 percent of the transactions are with the former Soviet Union, 24.8 percent with the former Czechoslovakia, 14.3 percent with Hungary, 7 percent with Poland, 4.3 percent with Romania, 6.5 percent with former East Germany, 6.5 percent with Bulgaria, 6.1 percent with the former Yugoslavia, and 0.9 percent with Albania. Thus our sample can be considered to be representative for East±West countertrade. North±South countertrade, in contrast, is underrepresented in the sample. Only 5.7 percent of the transactions took place with Africa, 3
Puzzling Facts and Popular Explanations
19
percent with Asia, 2.6 percent with South America, and 2.2 percent with China. The deals are mostly very large in size, ranging from U.S.$ 8,400 to U.S.$ 635 million with a mean of U.S.$ 11.1 million. All statistics presented in this book are based on the number of contracts, rather than trade volume, as the unit of analysis. As we show in the following sections our micro data set allows us to test a much richer set of predictions than previous studies.1 In particular, we can consider predictions of the optimal design of countertrade contracts, using a comparatively large number of observations. 2.2
Stylized Facts about Countertrade
The following observations on countertrade have been made by many experts and can be found in most empirical studies on the subject, including ours. observation 2.1 indebted.
Countertrading countries tend to be highly
The debt±GDP ratios of the countries engaged in countertrade range from 4.5 percent (Iran) to 326 percent (Zambia), with an average debt±GDP ratio of 71.5 percent in 1987. A comparison with the debt ratio of Poland (69.9 percent), the country whose debt problems in 1981 anticipated the international debt crisis, suggests that this number is rather large. The average debt±GDP ratio for all developing countries in 1987 was 39.6 percent.2 Casson and Chukujama (1990) report evidence (based on 35 observations) that countries with higher debt ratios are more strongly engaged in barter. Hennart (1989) examined the relationship between the extent to which countries used countertrade and their debt±service ratio in 1983 for 40 countries, but could not ®nd a clear relationship between indebtedness and countertrade. Hennart and Anderson (1993) used different aggregate variables and found (on the basis of 40 observations) that a country's creditworthiness, as measured by a composite of ratings of international banks is positively correlated with its barter activities. observation 2.2 The degree of tying of the two trade ¯ows in countertrade varies considerably.
20
Barter Contracts in International Trade
Table 2.1 Contract type and compensation ratio Compensation ratio
Counterpurchase
Buyback
2 [7.7]
104 [59.1]
7 [25.0]
113 [49.2]
100%
18 [69.2]
52 [29.5]
14 [50.0]
84 [36.5]
101±400%
6 [23.1]
20 [11.4]
7 [25.0]
33 [14.3]
Total
26 (11.3)
176 (76.5)
28 (12.2)
2±99%
Barter
Total
230 (100)
Source: Sample of 230 countertrade contracts. Notes: The compensation ratio is the value of export of an Eastern European country, or a developing country, as a percentage of the import value from the industrialized country ®rm. Numbers without fences are absolute numbers of cases. Numbers in brackets are column percentages; numbers in parentheses are row percentages. Chisquare value 33.82; marginal signi®cance level 0.000.
In countertrade an import is tied to an export. The degree by which this tying takes place is called by the traders the ``compensation ratio.'' The compensation ratio is the value of the export of the Eastern European or developing country as a percentage of the original import value from the industrialized country. Thus a compensation ratio of 100 percent means that the industrialized country purchases export goods of equal value to the original import to the developing or Eastern European country. The compensation ratio varies between 2 and 400 percent, as shown in table 2.1. In almost 50 percent of the contracts the parties opted for less than complete tying of trade ¯ows by choosing a compensation ratio below 100 percent. Buyback contracts tend to have repurchase values that are equal or are higher than the original import value (75 percent of all buyback agreements have a compensation ratio of 100 percent or higher). Counterpurchase contracts instead tend to have smaller repurchase values (59 percent of all counterpurchase contracts have a compensation ratio below 100 percent). observation 2.3 purchase.
The dominant form of countertrade is counter-
Recall that counterpurchase is a form of countertrade in which the import is paid with foreign exchange but the industrialized country
Puzzling Facts and Popular Explanations
21
Table 2.2 Pattern of trade with Eastern Europe in 1987
Machinery and equipment Consumer goods
Imports from OECD countries (in %)a
Exports to OECD countries (in %)a
Countertrade (1)
Total Tradeb (2)
Countertrade (3)
Total Tradeb (4)
70.9 8.8
35.7 8.2
35.5 31.9
17.7 16.4
Basic goods Chemicals
3.6d 8.8
43.7c
18.4d 8.2
47.9c
Services
4.0
NA
6.0
NA
Foods and agricultural products
10.4
10.9
Sources: Sample of 230 countertrade contracts; Comecon Data 1989; Vienna Institute for Comparative Economic Studies, Vienna. a. Percentages of columns 1 and 3 refer to the total of 230 countertrade contracts of the sample. Percentages of columns 2 and 4 refer to total imports and exports of Eastern Europe. Eastern Europe is here de®ned as Poland, Hungary, Czechoslovakia, and the Soviet Union. These four countries cover more than 60 percent of the 230 countertrade contracts of the sample. Their respective weights are 0.115, 0.235, 0.407, and 0.243. These weights correspond to the importance of the respective country in the countertrade sample. b. This column gives the CMEA classi®cation of commodity groups called CMEA trade nomenclature (CTN): CTN1 machinery and equipment, CTN9 consumer goods, CTN2±5 basic goods, and CTN6±8 food and agricultural products. c. Includes chemical products. d. Includes food and agricultural products.
commits to buy export goods from the Eastern European or developing country in return. The prevalence of counterpurchase has been reported by many studies (see OECD 1981, 1985; Bussard 1987; Hennart 1990) and can also be seen in table 2.1. In our survey, counterpurchase accounts for 76.5 percent of all countertrade contracts, compared to 11.3 percent for barter and 12.2 percent for buyback agreements.3 observation 2.4 The trade pattern in countertrade differs signi®cantly from that in untied trade. As shown in table 2.2, in our sample around 68 percent of all countertrade exports from developing and Eastern European countries are manufactured investment and consumption goods as compared to only 34 percent in total trade. Similar numbers are reported by Hennart (1989). On the import side, 70 percent of all countertrade
22
Barter Contracts in International Trade
imports from ®rms from industrialized countries were technology related and investment goods, as opposed to only 36 percent of total trade. Similar evidence is found by Bussard (1987). Note that this discrepancy between import patterns of countertrade and total trade can also be found in country speci®c comparisons. As an example, consider the former CSFR (57 of 230 cases in the sample) where 66.7 percent of all imports in countertrade were technology goods as opposed to 42.7 percent in total trade of this country. Similarly, for Hungary (33 of 230 cases), the respective percentages are 51.5 for countertrade versus 33.7 for total trade, and for Poland (16 of 230 cases), 87.5 as compared to 31.8. The numbers for total trade are taken from OECD trade statistics. If we are to understand the economic motivation for international barter and countertrade, we need explain these observations in a consistent way. What is the reason that mainly highly indebted countries turn to countertrade? Why is an import tied with an export and what explains the extent to which the two trade ¯ows are tied? What determines the trade pattern in countertrade, that is, the choice of import and export goods? A number of explanations have been put forth in the literature. We report now four of the most popular explanations, and confront them with the data on actual countertrade contracts. We ®nd some explanations are borne out by the data while others are not.4 Foreign Exchange Shortage One of the most frequent explanations of countertrade is that it allows countries to overcome the constraint on development imposed by a shortage of hard currency. Observation 2.1, that countertrading countries are highly indebted, is taken as evidence that these countries face a shortage of foreign exchange and that their low creditworthiness makes it impossible to ®nance imports with a simple loan from an international bank. Countertrade is interpreted as an arrangement that allows an importing country to save on foreign exchange (e.g., OECD 1981, 1985; Welt 1984). This interpretation is not fully plausible for the following reasons. First, countertrade uses export goods that otherwise could have been used to generate foreign exchange to pay for future imports (Banks 1983, p. 165). Thus one can hardly argue that it actually relaxes the foreign-exchange constraint. Furthermore, if the foreign-exchange shortage were the main explanation of countertrade, we would ex-
Puzzling Facts and Popular Explanations
23
pect barter to be the main form of countertrade since it is the only form of countertrade that does avoid the use of hard currency. However, table 2.1 shows that barter accounts for only a small portion of total countertrade. Finally, if saving on foreign exchange were the main motivation behind countertrade, we should expect the export value to equal the original import value. Table 2.1, however, states that the compensation ratio differs considerably and only 36.5 percent of the countertrade contracts have a compensation ratio of 100 percent. Mirus and Yeung (1987) examine whether countries can use countertrade to ameliorate their hard currency position. They ®nd that countertrade in the form of simple barter or counterpurchase does not improve a country's foreign exchange position unless it improves economic ef®ciency in the sense that it leads to an increase in national income. Since countertrade in the form of buyback involves the construction of capital goods it reduces a country's foreign exchange holdings in the short run but increases them in the longer run. Mirus and Yeung interpret their results as evidence that if countertrade is a response to foreign exchange shortages, then it is only because it improves economic ef®ciency. Export Promotion Observation 2.4 on the trade pattern in countertrade serves as the basis for the second popular explanation of countertrade. It is argued that developing countries use countertrade transactions to promote the export of ``new'' goods, goods they did not export to industrialized countries before, in order to gain access to new markets and to diversify their exports (OECD 1981, 1985). How valid is this explanation? Countertrade will promote exports when one of two conditions are met: (1) Countertrade allows the countries to export goods they could not otherwise export. (2) Countries can effectively segment the market to create a market in a new region for a product that is sold already in other regions. Sales of the countertrade goods could cause displacement of traditional exports when they are sold at lower prices in the countertrade transaction than in traditional trade, and when the contracts are not endorsed with destiny clauses constraining or prohibiting the selling of the countertrade goods to third markets and third parties. Whether countertrade has been effective in stimulating the export of nontraditional products is examined in table 2.3. In 40 percent of
24
Barter Contracts in International Trade
Table 2.3 Export promotion LDC/EE trade partner First time/ occasional exportera
Goods and contracts characteristics
Regular exportera
Restriction on sale Yes
40.0
60.0
No
41.1
58.9
Investment/technical goods
40.8
59.2
Chemicals Consumption goods
59.1 35.1
40.9 64.9
Marginal signi®cance
0.88
Chi-square value
0.024
Category of countertrade goods
Basic sector goods
33.3
66.7
Services
71.4
28.6
Marginal signi®cance
0.105
Chi-square value
7.66
Countertrade goods exported for the ®rst time or occasionally
40.2b
Higher price in countertrade export
26.6
Lower price in countertrade export
10.6
Restriction on resale
36.8
Countertrade goods differentiated in design/quality c
51.8
Source: Sample of 230 countertrade contracts. a. Numbers are row percentages. b. In percent of total countertrade. c. Countertrade specialists' judgment of whether the countertrade goods are standardized or differentiated with respect to design and/or quality.
the transactions in our sample the Eastern European or developing countries exported goods that they have not (or only occasionally) sold in the respective export market before. The table provides evidence also on the second prerequisite for additionality of exports. In 10.6 percent of the contracts the countertrade goods have been potentially competing with non-tied exports by being sold at a lower price within countertrade. In order to avoid the displacement of traditional exports, the contracts have been furnished with destiny clauses in 36.8 percent of all countertrade contracts. The table also looks at whether regularly exported countertrade goods are more often endorsed with resale restrictions than newly
Puzzling Facts and Popular Explanations
25
exported goods. One would expect this to be the case since goods that are exported for the ®rst time are not potentially competing with non-tied exports. The null hypothesis of no association between status of countertrade export and resale restriction, however, cannot be rejected at conventional levels. Consumption and basic sector goods have tended to be sold less often by countertrade when the Eastern European or developing country trade partner is a ®rst-time exporter than when it is a regular one, which is some indication of countertrade pushing nontraditional exports (the relationship is signi®cant at about the 10 percent level). Summing up, the empirical evidence presented gives some support for the view that countertrade has helped to stimulate and diversify exports. 51.8 percent of the transactions involved goods that were differentiated in quality or design or both, 40.2 percent of the countertrade goods have not or only occasionally been exported before, and in 36.8 percent of the cases countertrade has been used as an attempt to create a new market.5 Other studies con®rm that the goods exported by developing countries through countertrade arrangements are often goods for which export markets have yet to be established. Readily marketable products, like raw materials, are usually not available for countertrade. It can also be observed that a country removes goods from the countertrade shopping list once it has gained some experience with exporting these particular goods (Banks 1983, p. 167). However, opinions diverge how to interpret this phenomenon. Some authors claim that developing countries use their purchasing power to force undesirable goods on Western ®rms. One example is McVey (1985, p. 10). But it seems rather questionable that developing countries indeed have monopsony power on the world market. Even if this were the case they could just as well use their bargaining power to enforce lower import prices (Banks 1983, p. 169). Most important, the available evidence in our sample of 230 countertrade contracts does not support the view that the countertrade goods exported to industrialized countries are hard to market. Most Western trading partners in the sample con®rmed that the goods they received were of good quality, matching Western standards.6 A second interpretation is that the goods sold through countertrade are of high quality, but developing countries cannot sell them on their own because they lack the marketing expertise and have dif®culties gaining access to Western markets. Even in this case,
26
Barter Contracts in International Trade
countertrade seems to be a rather inef®cient way to promote a country's exports, given that the Western ®rm involved is not necessarily expert in marketing the goods. In fact most Western ®rms employ the help of specialized trading companies to market the goods they receive. This raises the question why developing countries do not choose specialized trading companies as trading partners for their exports in the ®rst place (Banks 1983). Murrell (1982) argues that buyback arrangements can help Eastern European enterprises to signal high-quality products by using capital equipment made by a Western company. Furthermore, by making itself dependent on Western technological updates and after-sales service, the Eastern European ®rm gives a signal of the reliability of its future conduct. Using data on countertrade contracts from Business Eastern Europe, Murrell ®nds supportive evidence for the theory that informational dif®culties and therefore market signals affect the pattern of East±West buyback trade. But since only a small fraction of countertrade are buyback agreements, as seen in table 2.1, this cannot account for the trade pattern of countertrade as a whole. Market Distortions A third popular explanation given for countertrade is that it corrects distortions in noncompetitive markets. To illustrate this, consider the following two examples. Suppose that the countertrade partner is a member of an international price agreement (like OPEC or an International Commodity Agreement) that has surplus capacity. Under these circumstances this country will want to undercut the cartel price. One way of doing so without openly violating the collusive agreement is to use barter. The developing country sells the countertrade goods at the of®cial price and takes an overpriced good from the industrialized country ®rm. The lack of transparency in barter makes it an effective route for chiseling. Circumventing collusive agreements via countertrade might account for why the developing countries use the barter form of countertrade more frequently than the EE countries, as shown in table 2.4. 83.3 percent of the agreements in which the countertrade partner is a member of a cartel are signed with developing countries.7 Another example in which barter becomes attractive is when the market distortion is an industrialized country that price-discriminates, selling to an Eastern European or developing country trade partner (Caves 1974). As is well known, the price-discriminating monopolist
Puzzling Facts and Popular Explanations
27
Table 2.4 Contract type and region Eastern Europe
Developing countries
Total
Barter
18 [69.2] (9.0)
8 [30.8] (25.8)
Counterpurchase
158 [89.8] (79.4)
18 [10.2] (58.1)
176
23 [82.1] (11.6)
5 [17.9] (16.1)
28
199 [86.5]
31 [13.5]
Buyback
Total
26 (11.3)
(76.5)
(12.2) 230 [100]
Source: Sample of 230 countertrade contracts. Notes: Numbers without fences are absolute numbers of cases. Numbers in parentheses are column percentages; numbers in brackets are row percentages.
can increase pro®ts by asking different prices from customers with different levels of willingness to pay. Assuming that consumers in Eastern European or developing countries have a lower willingness to pay for Western products, and that the industrialized country ®rm is constrained in the exercise of its monopoly power, barter allows the monopolist to discriminate by obscuring the effective price from disfavored customers. From the perspective of Eastern European or developing countries, barter is a way to extract monopoly rents from the industrialized country ®rm. Magenheim and Murrell (1988) see barter as a vehicle for price discrimination that allows to hide the effective terms from other agents and thus to maintain a strong bargaining position vis-aÁ-vis other customers. The authors examine when barter can be mutually bene®cial to both trade partners and which ®rm characteristics are likely to be correlated with barter activity. Caves and Marin (1992) test for price discrimination in countertrade transactions and ®nd supporting empirical evidence. In particular, they ®nd that industrialized country ®rms possess ample market power to discriminate. Better terms go to novice partners and others with presumed low reservation prices, who should bene®t from discrimination.8 If price distortion plays a role as a source for barter, we expect to ®nd that the size of the compensation ratio will vary systematically with the industrialized country ®rm's market power and with
28
Barter Contracts in International Trade
Table 2.5 Market distortion and compensation ratio Compensation ratio <100%
100%
>100%
49.1
36.5
14.3
(67.0)
56.6
32.9
10.5
(9.3)
38.1
52.4
9.5
(23.8)
35.2
37.0
27.8
Countertrade Industrialized country ®rm's market positiona Leading producer Follower One among many Marginal signi®cance Chi-square value
0.005 14.86
Status of party in Eastern Europe or developing country (2.6) Member of cartelb
0.0
83.3
16.7
Administered pricec
(77.3)
53.1
32.8
14.1
None of above
(20.1)
39.1
45.7
15.2
Marginal signi®cance
0.043
Chi-square value
9.83
Source: Sample of 230 countertrade contracts. Notes: Numbers are row percentages; numbers in parentheses are column percentages. a. Countertrade specialists' judgment of whether the industrialized country ®rm is a leading producer in terms of market share and/or technology. b. Oil cartel or commodity agreement. c. Regulated prices in agriculture, and administered prices in planned economies.
whether or not the Eastern European or developing country party is a member of a cartel or a collusive agreement. In barter transactions with no money involvement, the compensation ratio determines the terms of trade of the transaction. A low repurchase value as a percentage of the import value means unfavorable terms of exchange for the industrialized country ®rm and favorable ones for the Eastern European or developing country party. Table 2.5 provides evidence on the association between the compensation ratio, the industrialized country ®rm's market position, and the status of the Eastern European or developing country party. Leading producers tend to have contracts with compensation ratios below 100 percent (56.6 percent of all contracts signed with ®rms from industrialized countries that are leading producers have a compensation ratio between 2 and 99 percent), while the reverse is the case for ®rms from industrialized countries with negligible market shares. The null hypothesis of independence between the compensation ratio and the industrialized
Puzzling Facts and Popular Explanations
29
country ®rm's market position is rejected by the chi-square test. A similar result is obtained for the industrialized country ®rms' competitive conditions (not shown). If the Eastern European or developing country party is a member of a cartel the contract is more likely to specify compensation ratios of 100 percent or above. The relationship is signi®cant at the 5 percent level. The results given in table 2.5 provide some con®rmation of the hypothesis that barter is used as a vehicle to change the terms of trade to allow price discrimination of Western monopolists. Technology Transfer Why are countertrade imports predominantly technology goods? Parsons (1985, 1987) reports that the sale of technology is particularly prone to incentive problems. The value of a plant or some equipment to the buyer is directly affected by actions of the seller that are dif®cult to specify in a contract, such as the level of care with which a plant is set up and workers are trained, or the supply of after-sale services and technological advances. Similarly the seller of a production plant may have superior information regarding the quality and/or future value of the commodity to be produced with the plant, which leads to a problem of adverse selection. One obvious solution to this kind of incentive problems is for the seller to engage in foreign direct investment. But in many Eastern European and developing countries this was not a feasible option, due to political and ownership constraints. Thus alternative ways to solve the technology transfer problem had to be found (Kogut 1986; Hennart 1989). In theory, a contract could be written that would stipulate that no payment be made until the facility works as intended and that would obligate the technology seller to provide continuous after-sales assistance. But in practice, contracts provide only limited protection, as the example of turnkey contracts show (Hennart 1989). These contracts were developed in the 1960s to sell complete plants to Eastern Europe and have been used extensively by other countries, that have discouraged incoming foreign direct investment, like Algeria. Abdallah-Khodja (1984) who was Algeria's minister of planning and one of the main forces behind Algeria's reliance on turnkey contracts, reports that turnkey contracts do not seem to have effectively protected technology buyers against opportunism by the seller.
30
Barter Contracts in International Trade
Parsons (1985, 1987) proposes buyback contracts as a solution to this moral hazard problem. A buyback contract gives the ®rm from an industrialized country an incentive to spend more effort in setting up the plant properly in order to improve the quality of the output it is paid with. Similar arguments have been made by a number of authors who emphasize the negative implications of the incentive problems described above on international technology transfer. Mirus and Yeung (1986, 1993) see countertrade as a rational economic response to transaction costs and information asymmetries. Like Parsons, they argue that buyback contracts are incentive contracts that ensure the transfer of desirable quality technology and postinstallation service performance if standard forms of internalization, like joint ventures or foreign direct investment, are not possible due to political and ownership constraints. Hennart (1989) argues similarly that buyback contracts give the buyer of technology a signal about its quality, since a plant supplier would be unwilling to contract for the product if he intends to sell obsolete equipment and thus knows that the plant will yield substandard output. Forcing the seller to buy back part of the plant's output at contracted prices also ensures that the plant seller will make realistic estimates of the level of future demand for the product. By agreeing to take some of the plant's output, the supplier of the plant demonstrates his con®dence in his forecasts of future demand. This argument has been modeled formally by Chan and Hoy (1991) and Choi and Maldoom (1992). Chan and Hoy (1991) consider an environment in which the production of an output gives rise to a double moral hazard problem when the output depends on inputs provided by both the technology supplier and the host country. They argue that a sharing rule for output or revenues, as in buyback contracts, in conjunction with a set of minimum standards requirements on some inputs can be used as complementary instruments to reduce the ef®ciency loss due to the double moral hazard problem. Choi and Maldoom (1992) study a similar double moral hazard environment and show that a buyback arrangement can produce higher pro®ts for the parties involved than one in which the good produced is sold in various markets to maximize revenue. The reason is that under the buyback contract the market allocation is ®xed, and so pro®ts are more responsive to changes in quality than under a market allocated contract that allows for adjustment. Thus buyback contracts provide
Puzzling Facts and Popular Explanations
31
a commitment to greater potential loss of pro®t in case of quality shirking, and this has positive incentive effects ex ante. All these suggestions see buyback as a potential solution to the technology transfer problem if foreign direct investment is not an option. For the argument to work, it is essential that there be a technological relation between the two goods to be traded. However, as we saw above in table 2.1, buyback accounts for a surprisingly small fraction of all countertrade transactions. Whereas more than 70 percent of all countertrade imports are technology goods, only 12.2 percent of all countertrade contracts are buyback agreements. Thus, even though this explanation is theoretically appealing, it cannot explain the great majority of technology imports that take place in form of counterpurchase. The question thus is how the technology transfer problem can be solved in counterpurchase as the dominant form of countertrade, where there is no technological relation between the import and export. Other Explanations There is a theoretical literature explaining barter in market economies that offers complementary reasons for nonmonetary trade but does not explicitly address the stylized facts we have reported above. Prendergast and Stole (1996) consider the chosen means of exchange when both monetary and nonmonetary exchange mechanisms are available. They illustrate three potential reasons for the emergence of barter. First, a willingness to barter may reveal information that cannot be revealed solely through monetary trade. Second, nonmonetary trade may constrain the ability of agents to engage in inef®cient rent-seeking activities. Finally, barter improves the ability of agents to impose trade sanctions on those who act dishonestly. Prendergast and Stole (1999) consider nonmonetary forms of exchange in the context of ®rms and organizations. They ask why ®rms often ban monetary exchange between their employees but encourage nonmonetary exchange. The additional reason they put forward here is that the use of barter may affect the allocation of rents in surplus-enhancing ways as agents respond strategically to the existence of these rents. Ellingsen and Stole (1996) discuss mandated countertrade as a policy to restrict unilateral imports. A country can use this policy
32
Barter Contracts in International Trade
to commit domestic ®rms not to purchase from a foreign trading partner unless there are reciprocal sales. The authors argue that the policy may be a rational response to fundamental contracting failures. The key assumption is that sellers are incompletely informed about buyers' valuations. The analysis suggests that an optimal mandated countertrade policy will target high markup imports and low markup exports. Ellingsen (1998) discusses barter as a screening device in case of uncertain liquidity constraints. The key assumption in this paper is that the debtor's liquidity is private information and that the debtor can signal his ®nancial constraint by making part of the payment in kind. Amann and Marin (1994) argue that barter can serve as an insurance mechanism against the risk from stochastic export revenues. In principle, countries can eliminate the risk from their foreign exchange earnings if they sell forward and thus have a guaranteed price over the period for which the futures market is open. However, futures contracts exist only for a narrow range of commodities, and most of them expire in six months or less.9 Barter allows forward selling of commodities where no organized futures market exists. By leading to more export stability, barter has arguably promoted growth in the Eastern European and developing countries. 2.3
New Interpretation of the Evidence
In the following chapters 3 to 5 we develop an alternative explanation of countertrade that is more consistent with the empirical observations and confront this theory with evidence from our data sample. In chapter 3 we deal with observation 2.1 that countertrading countries typically are highly indebted. We show that countertrade helps to solve the sovereign debt problem of developing countries and Eastern Europe, not because it avoids the use of foreign exchange altogether but because it helps the country to credits that otherwise would not have been granted. The second transaction, the export from the developing country, is used as a deal-speci®c collateral for the credit granted for the original import. Rather than solving the debt problem of the country overall, countertrade restores creditworthiness of highly indebted countries for one particular international trade transaction. The important question we address in this chapter is why it makes a difference if trade partners promise a pay-
Puzzling Facts and Popular Explanations
33
ment in goods rather than a payment in cash. We ®rst study this question in a framework with a one-shot credit interaction. Then we allow for repeated interaction to compare the enforcement mechanisms discussed by the sovereign debt literature such as reputation effects and trade sanctions with those if trade credits are collateralized through barter agreements. Furthermore we turn to a problem developing countries in particular are struggling with, ¯uctuations of world prices for commodity exports. We analyze how stochastic export revenues affect the problem of enforcing debt repayment and show that barter can help to overcome credit problems related to revenue uncertainty. Finally, we make a ®rst endeavor to confront our predictions with empirical evidence on actual barter contracts. We derive how the value of the deal-speci®c collateral is determined in response to variations in a country's creditworthiness, uncertainty, and reputation and test these predictions with the data. In chapters 4 and 5 we discuss the pattern of specialization of international barter. Our argument that payment in kind may have advantages over payment in cash contradicts the conventional wisdom in the theory of money. The common view is that barter is inef®cient because it does not overcome the double coincidence of wants problem as money does. More important, payment in goods is problematic because it is dif®cult to judge the quality of the goods offered as means of payment. These problems have natural implications for which type of goods will qualify as means of payment in international barter. This choice will explain the export pattern of barter. We ®nd evidence that in contrast to what the descriptive literature on countertrade argues, ``new'' goods may be chosen as export goods in countertrade exactly because developing countries ®nd it dif®cult to market them on their own. We show that this can help to mitigate contractual hazards. In chapter 5 we turn to the import pattern of barter and explain why countertrade imports are predominantly technology goods. In contrast to Parsons (1985, 1987), we show that a counterpurchase contract can solve the technology transfer problem even though the two goods are not technologically related. We show furthermore that like barter, counterpurchase can solve the problem of creditworthiness of highly indebted countries, even though in both transactions monetary payments take place. Why can the tying of two transactions, an import and an export, mitigate the moral hazard of technology transfer and credit repayment? We argue that the export from
34
Barter Contracts in International Trade
Eastern Europe or the developing country serves as a hostage (in the language of Williamson 1983) that deters cheating on quality and defaulting on the payment in the original import from the industrialized country. The export has to be pro®table to both the industrialized country ®rm and the Eastern European or developing country, and the contract is designed such that the export becomes suf®ciently less pro®table for either party that does not ful®ll its obligations in the original import, be it technology transfer or payment of import. Although the import and the export are not technologically related, the countertrade contract establishes a ®nancial link that improves the incentives of the parties involved. Interestingly we ®nd that the creditworthiness problem can help to solve the technology transfer problem of developing countries, and thus countertrade can be seen as an ef®cient contract to substitute for foreign direct investment when these countries are reluctant to give access to foreign ownership in their markets.
3
Creating Collateral
In the preceding chapter we argued that countertrade is not a reaction to a lack of hard currency but rather a response to a lack of creditworthiness. In this chapter we explore how barter can help highly indebted countries to overcome their problem of creditworthiness. The important question we address is why it makes a difference if trade partners promise a payment in goods rather than a payment in cash, as in barter, or, more generally, if trade partners agree to link an import with an export, as in counterpurchase and buyback. To study this question, we develop a simple theoretical model which we will build and extend throughout the rest of part I. In this chapter we model countertrade as a barter transaction, where no monetary payments are made. But as we will show, this argument can be extended for counterpurchase and buyback as well. 1 3.1
Paying with Goods versus Paying with Cash
Before we turn to a formal analysis, let us illustrate our arguments with an example. Consider two potential trade partners, A and B. Think of A (she) as a ®rm in an industrialized country, and B (he), as a ®rm or a trade organization in an Eastern European (EE) or a developing (LDC) country. The trade partner in the East wants to buy some import goods from A but does not have the foreign exchange necessary to pay for these imports. Of course, the natural thing to do would be to turn to a bank and to ask for a credit. Whether or not such a credit is granted depends on two conditions. Will B be able to generate enough export revenues in the future to repay this credit? And if B is able, is B also willing to do so? We will take it for granted that the ®rst condition is ful®lled, at least in expected terms.
36
Barter Contracts in International Trade
It is the second condition that will be the focus of our discussion. That is, we restrict our attention to trade partners who are potentially solvent but temporarily liquidity constrained. If this credit could be guaranteed by the government of B's country, there would be no problem. If, however, B's country is highly indebted it is likely to be credit-constrained due to the sovereign debt problem. The problem is that B's government cannot be forced by the courts in A's country to repay its debt. It has to meet this obligation voluntarily, either for reputational reasons in order to maintain future access to international ®nancial markets or because it wants to avoid the international trade sanctions that may be triggered by default on a loan. If a developing country is highly indebted already, the gains from default on existing debt may outweigh the expected future losses, and Western banks will be reluctant to offer additional credit (Eaton 1991). In our story there is no difference whether B gets a credit from A or from a commercial bank. For simplicity we will assume that any credit is provided by A directly. We want to allow for the possibility that A can enforce at least some repayment. This could be guaranteed, for example, by assets that B's country holds abroad and that A can seize in case of default. However, the more severely indebted B's country becomes, the more foreign creditors have to share any assets that can be seized in case of repudiation, and thus the smaller will be the repayment that can be enforced by A. Our problem gets interesting if B's assets abroad are not valuable enough to cover B's import cost. This implies that B is not suf®ciently creditworthy to receive a credit for his imports. Of course, A will not deliver any goods unless she is sure to be ultimately paid. The interesting question is now whether B's creditworthiness can be improved if B turns to barter trade. Barter means that instead of selling export goods and using these export revenues to repay import credits, B commits to repay his debt by delivering export goods directly to A. But with a barter contract, a new incentive problem arises. Since B cannot be forced to deliver the export good, he must be induced to do so voluntarily. Why should this make any difference to A? As we will show, the advantage of such a barter contract is that the barter goods can now serve as an additional deal-speci®c collateral. This means that A in fact acquires property rights on the barter goods. If B refuses to deliver the export good destined for paying the imports from A, he may not be able to sell them at all. A
Creating Collateral
37
can use the courts in her own country (or in other industrialized countries) to enforce her claim and seize the export goods when they are shipped to some third party. With a credit arrangement, instead, A does not have a direct claim on the export good. Furthermore, even if the good is seized, the returns from the good have to be shared with all of B's creditors. Thus a collateral right on the barter good gives A a larger return than a claim on cash. This in turn increases A's incentives to track down B's export of the barter good. Of course, A's control over B's exports will typically be not perfect, and she may succeed in tracking down her collateral only with some positive probability. The important point, however, is that B is not free anymore to do as he likes, which reduces his incentive to cheat on A. How successfully A can label B's export goods as her own property determines how much barter can contribute to restoring B's creditworthiness. In chapter 4 we investigate the implications of this argument for the choice of countertrade goods. Legal Issues Before we turn to a formal analysis of the issues raised, we want to elaborate on the legal background for the argument just made. As we have just seen, barter means that instead of selling export goods on the world market and using the export revenues to repay the import credit, B delivers the export goods directly to A. In this case, no monetary payments are made at all for the import. In case of counterpurchase or buyback, monetary payments are to be made for both the import and export transactions, but they can be set off against each other. The crucial feature is that the export goods can be speci®ed to serve as collateral. This view of barter as a possibility to collateralize trade credits is documented by the legal literature. The Legal Guide on International Countertrade Transactions, prepared by the United Nations Commission on International Trade Law describes how trade partners can use barter contracts to protect the developed country ®rm against default on the payment of the original import by using barter goods as collateral (UNCITRAL 1993, p. 158). 57. It may be agreed, however, that, if a supplier has not been paid for goods delivered in one direction, that supplier is entitled to withhold payment for goods delivered in the other direction up to the amount of the outstanding claim or to set off the two countervailing claims. . . .
38
Barter Contracts in International Trade
59. The advantage of making the payment obligations interdependent is that of security to a party who does not receive payment for the goods it has supplied. . . . 60. When it is agreed that a party is entitled to withhold payment or to set off the two countervailing payment obligations, it is sometimes also stipulated that the party who delivered goods ®rst (the exporter) is entitled to take possession of the goods that are to be delivered by the other party (the importer). Taking possession of the goods would enable the exporter, who is holding the outstanding claim, to obtain value and establish a payment obligation that could be set off against the outstanding claim. Such a stipulation is possible where the countertrade agreement speci®es the goods that are to be counter-exported. In order to implement such an approach, it is advisable to identify clearly the goods and their location and to consider taking such additional measures as granting the exporter a security interest in those goods and giving the exporter an express right to claim their possession. . . .
These paragraphs specify how countertrading parties can deal with the issue of failure to complete countertrade transactions. In case of barter, the two parties can agree on using the barter goods as a collateral. In case of counterpurchase, the outstanding payment for the counterpurchase goods can serve as a collateral, as well as the counterpurchase goods themselves, if B fails to deliver them as speci®ed in the contract. It appears that there is no uniform international terminology regarding collaterals, nor is there a uniform international practice (Welter 1989). Some jurisdictions are more creditor friendly (notably English-based states and the United States), allowing security over all the assets of the debtor, including future assets, others are more restrictive (in particular, the more traditional members of the FrancoLatin group, subject to wide exceptions) (Wood 1995a, b). The international tendency seems to be to encourage security (Wood 1995b). That goods from countertrade transactions are in fact frequently used as a collateral has been reported such as by Welt (1984), Verzariu (1985), and Barkas (1987). A related question is how A can enforce her collateral rights in her own and third countries. If a debtor defaults on his payment (in cash or in goods), the creditor can seek legal recourse in her own country. If she obtains a judgment in her local court, the question is whether that judgment has any value in other countries. If it is enforceable in other jurisdictions, then the holder of the judgment could follow assets of the borrower into those jurisdictions. Typically a lender cannot enforce a judgment directly in a foreign country: enforcement
Creating Collateral
39
almost always requires the ®at of the local court. But legal developments have greatly facilitated the free movement of judgments, and it is possible for a lender to obtain such enforcement with relative ease provided that the original court had proper jurisdiction (Wood 1995a). Of course, as long as the collateral goods are still in B's country, A needs the support by courts in B's country. In the countries we consider here the courts are not very reliable, as we argued above. This is exactly why it is not helpful to secure the import credit with the goods originally imported (known as ``retention of title''). The advantage of securing the import credit with export goods rather than the original import goods is that B intends to sell these export goods abroad, which means that A has a better chance to seize them. Most developed jurisdictions are prepared to recognize and enforce foreign judgments by permitting the lender to sue on the foreign judgment locally (Wood 1995a, p. 260). The lender may also be able to enforce her foreign judgment on the basis of a reciprocal enforcement convention or statute. A number of jurisdictions have entered into such treaties (Wood 1995a, pp. 260±64). Within Europe, enforcement of foreign judgments is made possible by the European Judgment Conventions. Under these Conventions, all judgments are enforceable, no special procedure is required, and a foreign judgment may not be reviewed as to its substance. The countries covered by these Conventions are Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, United Kingdom, Austria, Finland, Iceland, Norway, Sweden, and Switzerland (Wood 1995a, pp. 232±43). All of the leading jurisdictions contemplate an application to the court by a creditor prior to the commencement of proceedings for an order prohibiting a debtor from removing assets from the jurisdiction. Claims can be made even where the amount of the debt is not certain. A creditor must show that she has a good prima facie claim and the grounds for making an urgent application. That is to say, the circumstances are such that if an injunction is not granted, the defendant is likely to remove his assets from the jurisdiction or dissipate them within it. Subject to this, it would seem that in the leading developed nations, these applications can be made and granted within a matter of hours (Wood 1995a, pp. 276±78). What are the economic consequences of using barter goods as collaterals for the debtor? If he defaults on his payment, the fact that he
40
Barter Contracts in International Trade
has granted collateral rights on the barter goods to be paid with means that he can no longer freely use these goods. He cannot sell them to whomever he pleases because he must fear that they are going to be seized if he tries to do so. This means, the value of the goods to him is substantially reduced. Thus, even if the creditor cannot be certain that she will be successful with her legal attempt in each and every country, the debtor is certain to suffer an economic loss because he is restricted in where he can sell his goods without having to fear legal interventions. A Formal Analysis To make our ideas precise, we will study the questions raised so far in a stylized model that builds on our example of the two trade partners A and B. Suppose that B wants to buy one unit of good 1 from A in period 1 and his willingness to pay for good 1 is v1 . A's production cost for this good are denoted by c1 , with v1 > c1 . In period 2, B can produce one unit of good 2 at cost c2 and sell it on the world market. This generates foreign exchange revenues of value v2 . We assume that B's revenues v2 in period 2 are suf®cient to pay for A's production cost c1 in period 1, which means that v2 b c1 . Thus a price p1 could be found such that p1 covers A's production cost
c1 a p1 and such that B is able to pay p1 in period 2
p1 a v2 . As spelled out in our introductory discussion of this example, B is cash-constrained in period 1. In period 2, however, when he is able to pay p1 , he cannot be forced to use his export revenues to do so. If B repudiates on his debt, all A can do is to enforce some ®xed repayment a, which captures the value of B's assets abroad. This punishment potential imposes an upper bound on the maximum credit that B voluntarily repays, p1 a a. Note that A is willing to deliver good 1 on a credit basis only if c1 a p1 and if B will indeed pay p1 . Thus we say that B is ``creditworthy'' if and only if a b c1 :
3:1
If instead a < c1 , we say that B faces a credit constraint because there exists no p1 such that p1 b c1 and p1 a a are satis®ed simultaneously. In the following we will focus on cases where B is not creditworthy in the sense de®ned above. Consider now a barter contract where, instead of selling good 2 on the world market and using its revenues to pay for good 1, B agrees to deliver good 2 in period 2 to A. Since B cannot be forced to pro-
Creating Collateral
41
duce good 2, he must be induced to produce and deliver good 2 voluntarily. This corresponds to the problem to induce B to pay p1 in a simple credit arrangement. The advantage of the barter contract, however, is that A now holds property rights on good 2. If B should try to cheat on A and sell good 2 on the world market instead of delivering it to A as speci®ed in the barter contract, A can ask courts to intervene. We model this as follows: Given the possibility of legal action by A, the potential surplus to B from producing the good and selling it on the world market, v2 ÿ c2 , is reduced to p^
v2 ÿ c2 , 0 a p^ < 1. If the probability that A manages to seize good 2 is suf®ciently high, it is optimal for B not to produce good 2 at all but to save his production cost. In this case p^ 0. If A's legal action is less effective, however, it may become optimal for B (given that he wants to default) to produce good 2 and to try to sell it to a third party, which generates an expected surplus p^
v2 ÿ c2 . The term p^ can be thought of as a measure of the anonymity of good 2. If p^ 1, good 2 is as anonymous as cash. This means that A has no chance to enforce her claim on good 2 because it is not recognizable as her property. However, if A can successfully label good 2 as belonging to her, then p^ < 1. In this case good 2 is less anonymous than money and it is harder for B to use it for any other purpose than paying his debt to A. Good 2 functions like ``special purpose money,'' which makes it a better collateral than money. The sequence of events is summarized in ®gure 3.1. Before period 1, A and B negotiate the barter contract. Good 1 is delivered at the beginning of period 1. At the beginning of period 2, B decides whether or not to deliver good 2 to A, a decision denoted by d A f0; 1g. If d 0, B does not deliver and A takes legal action in order to seize assets a and to track down good 2. If, instead, d 1, B delivers.
Figure 3.1 Time structure
42
Barter Contracts in International Trade
Thus B voluntarily delivers good 2 to A if and only if v1 ÿ c2 b v1 ÿ a p^
v2 ÿ c2 :
3:2
A is willing to deliver good 1 to B if and only if condition (3.2) is satis®ed and in addition v2 ÿ c1 b 0:
3:3
For barter trade to overcome B's lack of creditworthiness both conditions have to be satis®ed. This is possible under the following condition. proposition 3.1 Barter trade can overcome B's lack of creditworthiness if and only if v2 b c1 ÿ a c2 p^
v2 ÿ c2 :
3:4
Proof The necessity of (3.4) is straightforward to see. To prove its suf®ciency, note that (3.3) is satis®ed by assumption. The only possible problem thus is that condition (3.2) is violated, that is, c2 p^
v2 ÿ c2 ÿ a > 0:
3:5
If this were the case, A could induce B to deliver good 2 by making a voluntary monetary side payment s conditional on B's delivery. As long as (3.4) holds, it is possible to ®nd a side payment such that v2 ÿ c1 ÿ s b 0 b c2 p^
v2 ÿ c2 ÿ a ÿ s:
3:6 r
Condition (3.4) shows that the moral hazard problem of debt repayment can be solved only if the problem of creditworthiness is not too severe, that is, if a is not too small, if the deal-speci®c collateral v2 is suf®ciently large, and if p^ is suf®ciently small (i.e., if A can successfully label good 2 as her property). The advantage of paying with goods rather than money is that goods are less anonymous than money and thus they can be earmarked as property of the creditor. If A could not label and identify good 2 as her own collateral in case of con¯ict, nothing could keep B from selling it to someone else. But, since goods act like special purpose money, they are better collaterals than cash. To conclude, in a one-shot credit relationship barter creates a dealspeci®c collateral and thus helps to overcome the credit constraint for one particular countertrade transaction. The reason is that by agreeing to repay the trade credit in goods rather than cash the
Creating Collateral
43
debtor country voluntarily reduces its ability to use the export goods for other purposes than to pay for the import. 3.2
Goods versus Reputation
So far we have emphasized the dif®culty of enforcing the repayment of international credits. The debt crisis of the 1980s stimulated proli®c research on the sovereign debt problem.2 In an important paper, Eaton and Gersovitz (1981) examined how ``reputation effects'' can sustain sovereign lending. They show that the threat of exclusion from credit markets provides an incentive for a country to repay its debt. If countries intend to borrow not just once but repeatedly, they are interested in maintaining a reputation for being good creditors. However, the amount of potential credits that can be sustained with such reputational concerns is limited. We want to investigate how barter can be used to sustain sovereign lending when reputation as an enforcement mechanism fails to be effective. The crucial advantage of barter turns out to be that it allows credit markets to establish seniority rights for new creditors. This means that in case of default the new creditors are served before other unsecured creditors. This allows creditors to overcome the debt overhang problem. To make this point, we compare the enforcement mechanisms discussed by the sovereign debt literature such as reputation effects and trade sanctions with those if trade credits are collateralized through barter agreements. Regarding our example of the two trade partners A and B, the question is to what extent reputational concerns can sustain international lending. Suppose that the trade partners of our example are engaged in a long-term relationship where B intends to buy import goods on a credit basis not just once but repeatedly. The idea of a reputation mechanism is that A threatens never again to deliver import goods if B repudiates once. Thus B would lose his discounted payoff from all future purchases of the import goods. However, as Bulow and Rogoff (1989a) pointed out, reputational concerns alone cannot enforce repayment if B can switch to cashin-advance contracts. In this case B can take the foreign exchange revenues that were destined for credit repayment and use them for importing goods instead. The problem is that even if everybody believes that B will not repay any debt in the future, A and A's competitors cannot commit not to deal with B if he offers to pay cash in advance for import goods. But A may have other possibilities to
44
Barter Contracts in International Trade
punish B and to recover some of her money. In particular, A may try to track down B's future exports and imports and take legal action in order to con®scate these goods or the payments associated with them.3 For instance, A may have goods seized that are shipped by B to A's country but not paid yet, or goods that are destined for B and paid but not yet shipped. Kletzer and Wright (1990) show that the problem caused by cashin-advance contracts can be mitigated if it is possible to give initial creditors a seniority right on any monetary transfers made by the country that defaulted on its debt, that is, if the initial creditor is served before later creditors are. For our example, this means that A should be given seniority rights on B's returns from selling his export goods. However, seniority rights on cash are notoriously dif®cult to enforce for two reasons. First, it is much more dif®cult to seize cash than to seize physical goods. Thus it may be impossible for A to enforce her claim. Second, if B is highly indebted already, there are other creditors having claims on B's cash who will refuse to give A seniority rights. This is the classical debt overhang problem (e.g., see Krugman 1992). An example of the ineffectiveness of such seniority rules is described in Bulow and Rogoff (1988). In February 1988, Mexico, as part of a buyback scheme, issued new debt, promising to treat it as senior to existing bank debt. However, the market reaction to this announcement indicated that creditors were not convinced of the enforceability of these seniority rights. The advantage of barter trade, as we have seen above, is that it gives creditors a property right on export goods rather than on cash revenues. This has two advantages. First of all, it is easier to establish seniority rights on goods because it is easier to seize physical goods than cash. Cash, or ®nancial transactions in general, are much easier to hide than physical goods. Second, the creditor has a larger incentive to pursue a nonpaying debtor because the seniority rights increase his expected return from legal action. Suppose B repudiates his debt and tries to sell the export goods to some other party C. If A has a property right on these goods, due to a barter contract, and if she manages to track them down, she has a direct claim on them and can ask the courts in C's country to seize them. With a credit arrangement A does not have a direct claim on these goods. If she manages to track them down, she will get the courts to seize them only if they do not yet belong to C. Furthermore, even if they are seized, the returns from the export goods have to be
Creating Collateral
45
shared by all of B's creditors. Thus a property right on some speci®c export goods gives A a larger return than a claim on cash. This in turn gives A a larger incentive to take actions in order to have her rights enforced. Thus both effects, that goods are easier to seize than cash and that A will spend more effort of having goods seized when a barter contract establishes property rights on them, gives barter an advantage over cash transactions, even in a repeated framework. A Formal Analysis To make this point more formally consider the following dynamic framework, which is an in®nitely repeated version of the credit relationship between the two trade partners, A and B, described in chapter 3. In each period, starting with period 1, B wants to buy one unit of good 1 at some ®xed price p1 from A. Suppose that p1 is the spot market price for good 1 and that there are other sellers in industrialized countries offering this good at the same price. Furthermore, starting with period 2, B can produce one unit of good 2 each period and sell it on the world market at price v2 . Recall that A's production cost for good 1 are c1 . To save on notation, but without loss of generality, we normalize production cost of good 2 to c2 0. We will consider a dynamic framework with in®nitely many periods where both parties A and B have a common discount factor d 1=
1 r, where r > 0 is the world interest rate per period. We assume that B's revenues v2 in period 2 are suf®cient to pay for A's production cost c1 in period 1, that is v2 b c1 =d. Thus a price p1 can be found such that p1 covers A's production cost
c1 a p1 and such that B is able to pay p1 =d in period 2
p1 =d a v2 . We focus again on the situation where a, the assets B holds in A's country and which A can seek to seize with the help of courts in A's country in case of default, is too small to guarantee a repayment suf®cient to cover A's cost, that is, aa
c1 : d
3:7
This means that B is not creditworthy in a one-shot credit situation. The question is whether there exists a self-enforcing implicit agreement between A and B (which cannot be enforced by the courts), saying that in each period A gives a trade credit to B in order to ®nance
46
Barter Contracts in International Trade
the purchase of good 1, and B repays p1 =d out of his revenues v2 one period later. An implicit contract is a subgame perfect equilibrium of the repeated game; namely it is optimal for each party to stick to the terms of the agreement on and off the equilibrium path. Abreu (1988) has shown that a path of behavior can be sustained as a subgame perfect equilibrium if and only if it can be sustained with the threat of the worst possible punishment equilibrium for each player. In our context the worst that can happen to B in case of default is that A seizes his assets a and refuses to trade with B in the future. B has the option to switch to other suppliers of good 1. While they may not be willing to offer a trade credit to B, they will not refuse to deliver good 1 if B pays cash in advance. However, given the possibility of legal action by A described above, the expected gains from trade of future exports and imports are reduced. Let p, 0 a p a 1, denote the probability with which B can cheat on A without being interfered with. This implies that his expected revenues from selling good 2 to a third party, once the default has been detected and A attempts legal action, are pv2 . His expected revenues from importing good 1 on a cash in advance basis is reduced to p
v1 ÿ p1 . Thus B's payoff in case of default is given by4 v1 ÿ da d
v2 p
v1 ÿ p1
y X
d t
p
v2 v1 ÿ p1 :
3:8
t2
Note that B can default only after the export in the second period has been carried out, since the revenue v2 from this export was supposed to be used to repay p1 =d. Hence A's punishment can, at the earliest, affect the import in the second period and the export in the third period. If B sticks to the terms of the implicit agreement, his payoff is y X p1
3:9 d t v1 ÿ v2 : v1 d t1 Comparing (3.8) to (3.9), we see that B will repay his debt in every period if and only if v1
dv1 ÿ p1 =d v2 1ÿd b v1 ÿ da d
v2 p
v1 ÿ p1
y X t2
d t p
v2 v1 ÿ p1 ;
3:10
Creating Collateral
47
which is equivalent to 1 1 p1 a
1 ÿ pv1
1 ÿ da d
1 ÿ pv2 : d 1 ÿ pd
3:11
On the other hand, A is willing to participate in the transaction if and only if p1 b c1 . The following proposition summarizes this discussion: proposition 3.2 B's creditworthiness can be restored through a repeated credit relationship if and only if 1 c1
1 ÿ pv1 d
1 ÿ pv2
1 ÿ da b : d 1 ÿ pd
3:12
Note that if p 1, which is if A cannot affect B's gains from future international trade, we are back to the condition a b c1 =d. This means that the repetition of the credit relationship does not improve B's creditworthiness as compared to a one-shot relationship. Now we want to ask whether B's creditworthiness can be improved if B turns to barter trade. Again, we assume that B cannot pay for good 1 in period 1, neither with money nor with goods. However, suppose that instead of selling good 2 on the world market and using its revenues to pay for good 1 B agrees to deliver good 2 in period 2 to A as a payment in kind. As we observed above, B cannot be forced to produce and deliver good 2 to A, he must be induced to do so voluntarily. But, if he refuses to deliver to A, he may not be able to sell good 2 at all because the barter contract gives A property rights on good 2. As argued above, A can use the courts in her own country (or in other industrialized countries) to enforce her claim and seize good 2 when it is shipped to some third party. Again, we assume that A may succeed in tracking down her collateral only with some positive probability. We model this as follows: Given the possibility of legal action by A, the potential surplus from selling good 2 on the world market, v2 , is reduced to p^v2 , 0 a p^ a 1. Suppose that A and B agree to repeat the following barter deal in®nitely often. In every period A delivers one unit of good 1, starting in period 1, and B delivers one unit of good 2, starting with one period delay. If B sticks to the barter agreement his payoff is v1
y X t1
d t v1 :
3:13
48
Barter Contracts in International Trade
What is the worst possible punishment if B deviates in period 2, refuses to deliver good 2 to A, and switches to cash-in-advance contracts thereafter? Again, A can seize B's assets a and try to con®scate some of B's future trades. This is modeled by assuming that B's future gains from exporting good 2 (if he defaulted on the barter deal) are reduced to p^v2 . We argued above that it is sensible to assume that p^ a p. Furthermore his expected payoff from imports is, as before, reduced to p
v1 ÿ p1 , since barter has no effect on the possibility to interfere with future imports. Thus B's payoff in case of default is v1 ÿ da
y X
d t p
v1 ÿ p1 p^v2 :
3:14
t1
In contrast to the case without barter, if A has a property right on the export good produced by B in period 2, she can affect B's export payoff in period 2 already. B's incentive to deviate depends on the price p1 he has to pay if he switches to another supplier of good 1. Let us consider the case where B's incentive to deviate is maximal, that is, where p1 c1 .5 Substituting p1 c1 in (3.14) and comparing this expression to (3.13), B will not deviate from the barter agreement if and only if
1 ÿ da b p^v2 ÿ
1 ÿ pv1 ÿ pc1 :
3:15
On the other hand, A is willing to participate in the barter agreement if and only if y X
d t c1 a
t0
y X
d t v2 :
3:16
t1
The following proposition shows under what conditions a barter contract can achieve ef®ciency: proposition 3.3 Suppose that B is not creditworthy in a one-shot transaction, meaning that a < c1 =d. Then there exists a barter agreement that restores B's creditworthiness and implements the ef®cient allocation if and only if 1 c1
1 ÿ pv1
1 ÿ p^v2
1 ÿ da b : d 1 ÿ dp Proof
3:17
See appendix I-A.
Condition (3.17) shows that the moral hazard problem of debt repayment can be solved only if the problem of creditworthiness is not
Creating Collateral
49
too severe, that is, if a is not too small and if the deal-speci®c collateral v2 is suf®ciently large. Let us compare barter, the promise on goods, to a credit arrangement, the promise on cash. Comparing the left-hand sides of (3.17) and (3.12), we ®nd that barter is advantageous if
1 ÿ p^v2 > d
1 ÿ pv2 :
3:18
Thus barter outperforms a simple credit arrangement if p^ < p, that is, if a claim on good 2 is easier to enforce than a claim on cash. Furthermore, since A has a property right on the export good produced by B in period 2, she can interfere with B's export in period 2 already. The less frequently the transaction is repeated, namely the smaller is d, the more likely this is to play a role. Why should A be more successful to interfere with B's exports if she has a barter contract? In other words, why should p^ be smaller than p? As we have argued already, one important reason is that it is easier to seize physical goods than cash, since cash, or ®nancial transactions in general, are much easier to hide than physical goods. Furthermore A's success in pursuing B hinges on the effort A spends on tracking down B's exports and on taking legal action against B. A's payoff depends on whether she succeeds in laying her hands on good 2 in case of B's default. To see this, consider A's payoff in case of default if she has no barter contract with property rights on B's goods but purely a ®nancial claim. In this case all returns from legal action are shared with other creditors. In contrast, in case of barter, A's property rights on good 2 give A seniority rights on the collateral good. Thus A does not have to share the revenues from tracking down good 2 with any other creditor who has purely ®nancial claims but no property right on good 2. This means that A's default payoff in case of barter is strictly higher than that in case of a credit relationship. As a consequence A has a larger incentive to spend effort on tracking down good 2 in a barter contract which in turn results in p^ < p. Thus, even if we allow for a long-term relationship and take reputational concerns and trade sanctions into account, we still ®nd that barter outperforms credit arrangements. Summing Up: The Advantages of Barter As we have seen barter extends the limits of the reputation mechanism. Reputation works as an enforcement mechanism for debt
50
Barter Contracts in International Trade
repayment only if the industrialized country ®rm can impose sanctions on the debtor country's future trade ¯ows. However, as Bulow and Rogoff (1989a) have shown future trade sanctions might not prevent the country from defaulting on debt when the country can turn to paying for imports in advance before the actual import takes place. In this case we are left with barter as the major enforcement mechanism for debt repayment. Furthermore, even if reputational effects are not completely ruled out by cash-in-advance contracts, we still ®nd that barter is advantageous because it increases the punishment potential in case of default. What are the main advantages a promise of goods (barter) offers over a promise of cash (credit arrangements)? advantage 1 Goods are less anonymous than cash, and therefore it is easier to establish property rights on goods than on cash. These property rights on the barter good make it easier for A to seize the collateral good when B defaults and tries to sell it to a third party, as we have discussed above. advantage 2 Property rights on the barter good give A's claim priority over other creditors who have purely ®nancial claims and therefore increase A's expected payoff in case of default. At the beginning of this chapter we assumed that without a barter contract it is impossible to seize some of B's returns from selling good 2. Now we have seen that even if this is possible, barter is still advantageous. The reason is that without a barter contract A does not have a direct claim on B's revenues from selling good 2. If she manages to track down some of these returns, the returns from good 2 have to be shared by all of B's creditors. Thus a property right on good 2 gives A a larger return than a claim on cash. advantage 3 Since A's payoff in case of default is higher if she has priority rights on the good, her incentive to seek legal recourse and to track down B's export of good 2 is larger. Therefore the chances of success are higher (i.e., p^ is smaller). As we have seen above, A needs legal support to enforce her claims on B, either on the barter good or, in case of a simple trade credit, on B's cash revenues. But the chances of succeeding with her claim depend not only on the legal system but also on her own effort in seeking legal recourse and providing evidence of the justi®cation
Creating Collateral
51
of her claims. Clearly, the higher A's payoff is from succeeding with her claim, the larger is her incentive to engage in this activity. This in turn has a positive impact on the probability of succeeding with her claim. advantage 4 Barter makes more ef®cient use of B's export goods as potential collaterals. Consider, as we have argued above, that it is easier to track down goods than to track down cash revenues. One might still argue that rather than tracking down the barter good, A may, in case of default, try to track down B's future exports. In a model with a one-shot transaction this possibility is not considered, since there is no further export. But in a model with repeated interaction, this possibility has to be taken into account. Interfering with the debtor's future exports is what Bulow and Rogoff (1989a) describe as the only viable way to punish for defaulting on debt repayment. The problem with this is twofold. In contrast to seizing B's cash revenues, seizing his export goods is possible only if they do not belong to some other customer already. And even if this is not the case, then A still has to share all potential revenues with many other creditors, as in the case of tracking down cash revenues if A cannot claim to have priority rights on the future exports of B. Second, this requires that B plans to export regularly, after the default occurs. If B, instead, exports rather infrequently, A may have to wait for quite some time before she is able to satisfy (part of) her claims. The advantage of barter is thus that A can interfere directly with the export that is supposed to pay for the import, without having to wait for future and perhaps uncertain exports. At the end of this chapter we will turn to the data and examine quantitatively how important each of these forces, barter and reputation, are as enforcement mechanisms in international trade with highly indebted countries. 3.3
Dealing with Uncertainty
So far we have assumed that all economic variables are deterministic, in particular future world market prices and export revenues. However, one of the problems of developing countries is that their export revenues depend mainly on commodities with huge ¯uctuations of world market prices. These ¯uctuations in world market prices for
52
Barter Contracts in International Trade
commodities make these countries suffer large ¯uctuations in export revenues. As a consequence these countries can ®nd themselves in the situation where they are not able to satisfy their debt obligations, even though they may be willing to do so. Of course, one way of dealing with this problem is to condition the debt repayments that are due in each period on the export revenues realized in this period. This means that in periods of low world market prices and thus low export revenues, only low repayments have to be made. In periods of high export revenues, instead, high repayments are required in order to compensate for the low repayments in periods with low revenues.6 In this section we study the question how stochastic export revenues affect the problem of enforcing debt repayment. As we will see stochastic revenues make it more dif®cult for the reputation mechanism of credit repayment to work. Thus stochastic revenues further restrict the credit limit for indebted countries. We ask how barter can help to overcome these problems related to revenue uncertainty and to extend the credit limit. Consider again the solution suggested above: a contingent debt repayment that is high if revenues are high and low if revenues are low. This solution has two problems. First of all, it gives the debtor country an incentive to misrepresent its export revenues, claiming they are low because only low prices could be realized. Second, and more important, it implies that the debtor's incentive to repudiate his debt is particularly high in periods when high revenues have been realized because then high payments are due. Since these high repayments in good times have to compensate the low repayments in bad times, they need to be larger than the necessary repayment if export revenues were nonstochastic and always large enough to cover the credit cost. This in turn implies that stochastic export revenues make it more dif®cult to ensure that the debtor has always the right incentive to make his repayments. To illustrate this point by way of an example, consider the following extension of the in®nite horizon model introduced in the previous chapter. This time B's future export revenues v2 are subject to uncertainty. This uncertainty can arise from ¯uctuations in the demand for the export good, in prices or in exchange rates. Now suppose that in each period, v2 takes a high value v2 with probability q and a low value v2 with probability 1 ÿ q. The distribution of revenues in each period is assumed to be i.i.d. The expected value of
Creating Collateral
53
export revenues in each period is thus v2e
1 ÿ qv2 qv2 . A and B are assumed to be risk-neutral. Consider ®rst the case where A and B are engaged in a repeated credit relationship, without payment in goods. Suppose that v2 < p1 , meaning that B cannot pay p1 in periods with low revenue realizations even if he is willing to do so. This implies that the two parties have to agree on two different prices p and p1 , to be paid in periods 1 with low revenues and in periods with high revenues, respectively. Of course, this requires that A can observe the realization of v2 , since otherwise payments cannot be contingent on revenue realizations. Since both parties are risk-neutral, A is willing to accept these prices if the expected payment p1e
1 ÿ qp qp1 covers her production 1 cost and if she can be sure that B will pay these prices as speci®ed. Consider now the case where export revenues v2 have been realized so that B has to pay p1 , where p1 > p1e > p . In this case B's pay1 off from defaulting is higher than if he were to pay p1e or p for that 1 matter. This means his incentive to default is larger than in case of nonstochastic revenues. This makes it more dif®cult to sustain a credit relationship with a reputation mechanism. By resorting to barter trade the trade partners can overcome both problems. Barter means that the debtor pays with goods rather than cash. This puts the creditor in a better position vis-aÁ -vis the debtor when it comes to judging the value of these export goods. She cannot be deceived about the value because she realizes the sales and the revenues herself. On the other hand, the debtor has to deliver the goods before knowing how much these goods are in fact worth in each particular period. This has the advantage that the incentive to deliver the goods as promised is affected only by their expected value, not by low or, in particular, high realizations. This means that the incentive to deviate is not different in case of stochastic export revenues as compared to deterministic revenues provided that the expected value of revenues is the same in both cases. Thus barter alleviates the credit repayment problem because it forces the debtor to make his payment in terms of goods before the uncertainty about their value is resolved. This arrangement avoids the temptation to cheat on payment when the realized value of the goods destined for payment is high. In case of cash payments this is not possible, since here the amount due does in fact have to condition on the realized revenues. Since the debtor is more tempted to deviate when a high payment is due, the credit limit without barter is lower than if barter is used.
54
Barter Contracts in International Trade
As we have seen, the incentive problems for debt repayment become more severe when export revenues are stochastic. Barter helps to resolve these problems arising from uncertain export revenues. In the following section we check how these issues affect the design of barter contracts. 3.4
Confronting the Data
In this section we make a ®rst endeavor to test the predictions of the preceding sections with empirical evidence on actual barter contracts. We argued that the fact that the country promises to repay the trade credits in goods rather than cash creates a deal-speci®c collateral. We have seen that reputation as an enforcement mechanism to repay credit is limited and by turning to barter the debtor countries can extend their debt capacity. We also noticed that the incentive problem for debt repayment in a reputation framework becomes more severe when export revenues are stochastic and that this incentive problem of debt repayment can be solved when the debtor country repays its trade credit with goods before their market value is fully realized. From our theoretical analysis we can derive the following empirical predictions. As we have reported earlier on, barter can be sustained only if condition (3.4) is satis®ed. Recall that this is possible only if the value of the collateral is suf®ciently large, that is, v2 b c1 ÿ a c2 p^
v2 ÿ c2 :
3:19
On the left-hand side we have the value of the collateral created by the barter contract. This collateral has to cover at least A's production costs, c1 , as well as B's payoff from defaulting and selling good 2 to a third party, p^
v2 ÿ c2 . The collateral can be smaller the more A can punish B when he defaults. This is captured by the term a. The larger is the right-hand side of condition (3.19), the larger the value of the collateral needs to be in order to overcome B's credit constraint. From this condition we can derive our ®rst two hypotheses on how the value of the collateral needs to be chosen for a given credit cheating problem. We measure the value of the collateral by the value of B's exports to A. The required value of the collateral will be in¯uenced by the size of the original trade credit which A gives to B. To control for size effects, we normalize the value of the collateral (export value) by dividing it through the value of the trade credit (im-
Creating Collateral
55
port value). Hence we use what the traders call the ``compensation ratio'' (the ratio of the export value to the import value, from the point of view of B) of each barter trade as the measure of the value of the collateral provided by the barter contract. The following hypotheses derive how the value of the collateral relative to the trade credit is determined in response to different exogenous parameters: hypothesis 3.1 (Restoring creditworthiness) The lower is B's creditworthiness, the larger will be the value of the collateral relative to the trade credit. In the model, B's creditworthiness increases with a, the assets hold by B abroad which can be seized in case of default. Note that a reduction of a has a positive impact on the right-hand side of (3.19). Intuitively, the smaller the collateral that B can provide via a, the larger the collateral generated through barter has to be in order to restore B's creditworthiness. hypothesis 3.2 (Creating collateral) The better B's export opportunities are in case of default, the larger will be the value of the collateral relative to the trade credit. In our model B's export opportunities in case of default are captured by the term p^
v2 ÿ c2 . The larger p^, the larger the left-hand side has to be to satisfy condition (3.19). We argue that the more B is integrated into the world market, the less dependent he is on A as a customer and the easier it is for him to ®nd alternative venues to sell good 2. Similarly, the more standardized are the goods delivered by B, the more dif®cult it is for A to label them as her own property and the more easily B can sell them to some third party in case of default. From our analysis of repeated credit relationships we can derive two more hypotheses. We demonstrated that if trade is repeated barter can be sustained only if equation (3.17) is satis®ed. Rearranging this expression shows that again, this is possible only if the value of the collateral is suf®ciently large. v2 b
c1 ÿ v1 p
v1 ÿ c1 ÿ
1 ÿ da p^v2 : d
3:20
The collateral on the left-hand side created by the barter contract has to cover at least A's production cost, which is now c1 =d, as well as B's payoff from defaulting and selling good 2 to a third party, p^v2 , and
56
Barter Contracts in International Trade
buying good 1 from a third party on a cash in advance basis in the future, p
v1 ÿ c1 . In case of default A can punish B. The larger are v1 and
1 ÿ da, the smaller the collateral needs to be. This argument is presented in the following two hypotheses. hypothesis 3.3 (Using reputation) The value of the collateral relative to the trade credit is smaller the more frequently the barter transaction takes place. The more frequently the transactions take place, the larger will be d and thus the smaller will be the right-hand side of (3.20). To see this, consider the derivative of the right-hand side with respect to d, which yields ÿ
c1 a < 0; d2
since a <
c1 c1 < 2: d d
3:21
When the transaction is repeated more often, reputation as an enforcement mechanism plays a larger role, and therefore the value of collateral can be smaller. hypothesis 3.4 (Using trade sanctions) The more B depends on his imports, the smaller will be the value of the collateral relative to the trade credit. B's bene®t from importing good 1 is given by v1 in the model. An increase of v1 reduces the right-hand side of (3.20) which allows for a smaller value of the collateral. The point is that the larger is v1 , the more B has to lose if he is (partially) cut off from future imports. Finally, we have seen that barter can be used to avoid problems related to enforcing debt repayment that arise if export revenues are uncertain. hypothesis 3.5 (Dealing with export revenue risk) As barter avoids exposure to export revenue risk, the value of collateral relative to the trade credit should be unaffected by the export revenue uncertainty. The mere fact that the parties engage in barter suf®ces to make sure that B is not exposed to export revenue uncertainty. If A is in fact risk-neutral, her collateral need not condition on the uncertainty, since B's expected payment is not a function of actual realizations of export revenues. To test hypotheses 3.1 to 3.5, we use the following variables.7 As a proxy for creditworthiness, we use the debt to GDP ratio DEBT as
Creating Collateral
57
reported by the World Debt Tables of the World Bank. The idea is that the more B is indebted already, the fewer assets remain to be seized by A in case of default, and thus the lower B's creditworthiness. As a proxy for how frequently the barter transaction takes place, we use the variable REPEAT. REPEAT measures how frequently A and B have interacted with each other in the past. The underlying presumption is that if A and B have traded regularly with each other in the past, they are more likely to continue to do so in the future. As a proxy for B's export opportunities in case of default, we use a country-speci®c and a deal-speci®c variable. As a countryspeci®c variable, we use the export ratio of B's country EXPORT. A high export ratio suggests that the country is well integrated into the world market indicating that B's outside option to barter is good. The deal-speci®c variable we use is BASICM. This is a dummy variable that takes the value of one if the good that is exported from B to A in the relevant barter deal is basic materials, agricultural goods, raw materials, or oil. These are very standardized goods that are easy to market for B and are dif®cult to label as A's property. As a proxy of B's bene®t from importing good 1, we use TECHIMP which is the ratio of technology imports over total imports in B's country. A large share of technology imports indicates that B depends essentially on A's imports and that it will be particularly dif®cult to ®nd substitutes. Finally, we have included RISK, a dummy variable of value 1 when the parties agreed on a ®xed price for the barter good over the duration of the contract. RISK captures the insurance motivation for barter. Table 3.1 presents the results of testing hypotheses 1 to 5. DEBT, our proxy for B's creditworthiness, has a positive and signi®cant coef®cient for all speci®cations. This con®rms hypothesis 3.1. The barter contract is indeed furnished with a larger deal-speci®c collateral when the country lacks creditworthiness. EXPORT and BASICM, our proxies for B's export opportunities, both show the expected positive and highly signi®cant coef®cient. This supports hypothesis 3.2 which predicts that the value of collateral will be larger the better B's export opportunities. When A's punishment potential in case of B's default is weak because B can easily trade with someone else, then the barter contract needs to create a larger collateral in order to restore B's creditworthiness. REPEAT is the variable that captures how frequently the barter transaction takes place. It has the predicted negative and signi®cant sign. Just as hypothesis 3.3 and the sovereign
58
Barter Contracts in International Trade
Table 3.1 Value of collateral: Dependent variable lnCOMP (1)
(2)
(3)
(4)
(5)
lnDEBT
0.38a (4.77)
0.35a (4.33)
0.36a (4.36)
0.24a (2.60)
0.25a (2.65)
lnEXPORT
0.22c (1.67)
0.25c (1.88)
0.32a (2.41)
0.56a (3.94)
0.57a (3.92)
0.30c (1.73)
0.34b (1.95)
0.22 (1.23)
0.19 (1.09)
ÿ0.38a (3.00)
ÿ0.46a (2.86)
ÿ0.48a (2.90)
ÿ0.89a (3.64)
ÿ0.87a (3.52)
BASIC REPEAT lnTECHIMP RISK
ÿ0.06 (0.41)
Intercept F 2
Adjusted R Number of observations
1.93a (4.57)
1.88a (4.45)
1.52a (3.54)
4.24a (4.59)
4.21a (4.51)
16.9a
12.3a
11.8a
13.2a
11.0a
0.13
0.13
0.18
0.23
0.23
224
223
205
201
197
Source: Sample of 230 countertrade contracts. Notes: Ordinary least square regressions of 230 observations. Numbers in parentheses are t-values. Levels of signi®cance: a 1 percent, b 5 percent, c 10 percent. COMP Export value from EE/LDC in percent of import value to EE/LDC DEBT Debt to GDP ratio in 1987 of EE/LDC country EXPORT Export to GDP ratio in 1987 of EE/LDC country BASIC Dummy variable equal to 1 when good exported from EE/LDC country is basic good or chemical REPEAT Dummy variable equal to 1 when industrialized country ®rm and LDC/EE country have interacted frequently TECHIMP Share of technology imports in percent of total imports in 1987 of EE/ LDC country RISK Dummy variable equal to 1 when contract is ®xed price contract
debt literature suggests, the value of the collateral is smaller the more frequently the barter transaction takes place. It appears then that B's repayment discipline has been indeed driven by his concern of not losing his reputation as a good debtor. TECHIMP, the ratio of B's country's technology imports, over total imports, is the variable we use for testing hypothesis 3.4, which states that the more B depends on his imports, the smaller will be the value of the collateral relative to the trade credit. This variable has the expected negative
Creating Collateral
59
and signi®cant coef®cient, con®rming hypothesis 3.4. The data suggest then that B's dependency on technology imports is indeed a powerful mechanism to secure debt repayment. This is why the parties opted for a smaller collateral in the countertrade contract. Finally, as stated in hypothesis 3.5, RISK has no signi®cant effect on the value of collateral. Since barter avoids B's exposure to export revenue risk, B's incentive to repay his debt is unaffected by this risk. Thus it appears that the parties did not need to take care of this risk in the contract. Overall the empirical evidence from contract speci®c data seems to support the rationale for barter we have put forth in this chapter. All speci®cations explain up to 23 percent of the variation of the data on the compensation ratio.
This page intentionally left blank
4
The Export Pattern: Paying with Goods
Our analysis so far suggests that barter has important advantages over traditional credit arrangements. The main difference between a promise on future money and a promise on future goods is that goods have superior credit enforcement properties. Money is an anonymous medium of exchange. This anonymity can prove disadvantageous in trade with countries that lack creditworthiness. In contrast, goods can be earmarked as property of the creditor in the industrialized country. Thus goods act as ``special purpose money.'' In this chapter we develop a theory that explains which goods qualify for being used as ``special purpose money'' in barter contracts.1 This will allow us to make predictions about the export pattern in international barter. Our argument that payment in kind may have advantages over payment in cash contradicts conventional wisdom. The common view is that barter is inef®cient because it does not overcome the double coincidence of wants problem as money does. A seller may be paid with goods for which he has no use himself. Money as a medium of exchange reduces the search and transaction costs associated with barter (Kiyotaki and Wright 1993) because it can be used for anything, whereas goods have to be unloaded on someone else. While eliminating the need for a ``double coincidence of wants'' is certainly an important advantage of money, we argue that incurring some additional transaction costs by reselling the goods received as payment may be a small price for overcoming a credit constraint that otherwise would prevent trade from taking place. A second objection to payment in kind is raised by Banerjee and Maskin (1996) in their adverse selection theory of money. They argue that if sellers are paid with goods which they do not use themselves, they will ®nd it dif®cult to judge the quality of the goods offered as payment. This asymmetry of information gives rise to moral hazard
62
Barter Contracts in International Trade
and adverse selection problems. The debtor has an incentive to pay with low-quality goods because creditors are less informed about the physical characteristics (different quality levels) and about the market characteristics (value when resold) of the goods in question. Thus goods tend to be less liquid than money. Banerjee and Maskin argue that these moral hazard and adverse selection problems lead to inef®ciencies which can be completely overcome only when ®at money is introduced, a ``good'' whose physical and market characteristics can be discerned by everybody. An important implication of this adverse selection problem is that not all goods are equally well quali®ed to serve as a medium of exchange in a barter contract. We explore this idea in this chapter, and we show how the institution of commodity money explains the export pattern (the goods offered as payment) of barter. Using data on actual barter contracts, we characterize different types of goods in terms of anonymity and liquidity, and we discuss which of them would be the best candidate for collateralizing future payments. Recall from chapter 3 that in the absence of quality problems, B's creditworthiness problem can be solved if and only if the value of the collateral is suf®ciently large, that is,
1 ÿ p^
v2 ÿ c2 a b c1 :
4:1
The term
v2 ÿ c2 captures the surplus generated by good 2, whereas p^ captures the degree of anonymity of good 2. The surplus v2 ÿ c2 tends to be larger, the more homogeneous the good. Homogeneous goods tend to be more liquid, because due to their standard characteristics it is harder to cheat on quality. On the other hand, homogeneous goods have the disadvantage of being relatively anonymous, meaning that
1 ÿ p^ is relatively low. Differentiated goods in turn are less anonymous, meaning that
1 ÿ p^ is large, but also less liquid; that is,
v2 ÿ c2 tends to be smaller. A may not be able to judge the quality of the good she is paid with if she has no experience with this good. Thus she risks being paid with low-quality goods, if at all.2 Since it is the combination of anonymity and liquidity that counts, it is not clear a priori whether or not to prefer homogeneous to differentiated goods as means of payment.3 However, A can take measures to reduce the quality problems in case of differentiated goods. Recall that the asymmetry of information is particularly acute in the absence of a double coincidence of wants. This is the case if A does not use good 2 herself and has no
The Export Pattern: Paying with Goods
63
experience with it. One possible remedy to overcome this informational gap is to make some investment in getting better acquainted with good 2. A can, for instance, invest in the future marketing of good 2 by taking an active role in the design of good 2. Such an investment, while being costly, offers two advantages. First of all, it allows A to learn about good 2 and to be a better judge of its quality. This in turn reduces B's leeway to cheat on quality. Second, by differentiating and designing good 2 for A's marketing, good 2 becomes even less anonymous and thus worsens B's outside option if he should try to cheat on A and sell good 2 to a third party. It is straightforward to show that condition (4.1) is affected by such an investment i as follows:
1 ÿ p^
i
v2
i ÿ c2
i a b c1 i:
4:2
On the one hand, i has a positive effect on the liquidity
v2 ÿ c2 , and it reduces the anonymity p^, both of which increases the left-hand side of the condition. On the other hand, the investment causes additional cost i to be borne by A, which increases the right-hand side of the condition. Thus the advantage of differentiated goods as a collateral is that they can be made more liquid and even less anonymous by investing in marketing the good. However, how preferable it is to choose differentiated goods as collaterals depends on how costly it is to make the necessary investment. 4.1
Ranking of Commodity Money
As we have argued above, goods that are used as medium of exchange in barter have to have certain attributes to qualify as commodity money. The challenge is to ®nd goods that are relatively liquid and exhibit a low degree of anonymity. High liquidity means that there is little uncertainty about the quality of the good offered as payment. Low anonymity means that the creditor is offered a valuable collateral because she can successfully label the collateral goods as belonging to her and the debtor can use the good only for the special purpose of repaying his debt. Homogeneous goods tend to be liquid but also anonymous, whereas differentiated goods tend to be less liquid but also less anonymous. We classify goods by whether they are homogeneous or differentiated. Furthermore we identify whether buyer A has undertaken an investment to make the differentiated good more liquid and thus
64
Barter Contracts in International Trade
more valuable and also less anonymous. For this purpose we classify the export goods into three categories: basic goods BASIC, consumer goods CONSUM, and investment and machinery goods INVEST. To classify the goods offered as commodity money as homogeneous or differentiated goods, we use the following proxies.4 MDIF indicates whether or not the good offered as medium of exchange in barter is differentiated. MINF indicates whether or not the particular export good is traded on an exchange. When the good is not differentiated and/or an organized market exists for the good, we assume that A is not faced with an informational asymmetry, since she can either judge the quality or she can readily obtain information about the physical and market characteristics of the good from the market. Even if goods are differentiated A may trust B to keep promises about quality out of reputational considerations if they have dealt with each other for some time. We capture this familiarity with each other and with the good used as medium of exchange by the variable REPEAT. The next set of variables capture whether or not A invested to make good 2 more liquid and less anonymous. A will have to invest in marketing good 2 if she does not intend to use it herself. MUSE is a dummy variable that takes the value of 1 if the industrialized country ®rm uses the barter good for its own production. Thus we can interpret the variable MUSE as an indicator for the absence of marketing investments. Furthermore the good will be more liquid and less anonymous when there is a technical relationship between the good 1 A sells to B and good 2 used as payment by B (e.g., good 1 is a machine and good 2 is output produced with this machine). This is measured by the variable RELATION. Additionally we use the variable SPECINV to capture whether or not A makes a relationship speci®c investment. SPECINV combines HIGHCASH with LOCAL. HIGHCASH is a dummy variable equal to 1 when the export good is sold at a higher price outside barter trade. LOCAL is a dummy variable equal to 1 if A sells the export good locally. SPECINV is a dummy variable equal to 1 if HIGHCASH is equal to 1 and LOCAL is equal to 0. When A pays a lower price for the good in barter than in monetary transactions this indicates either that A does not want the good as payment or that A has to be compensated for an investment she makes into the medium of exchange. In order to distinguish between these two cases we combine HIGHCASH with LOCAL. A has to undertake an investment in
The Export Pattern: Paying with Goods
65
marketing good 2 when she does not sell it locally (and the good is differentiated). Thus, when A pays a lower price in barter compared to monetary transactions and when she sells good 2 abroad, then we conclude that A has invested in the relationship. The empirical results are given in table 4.1. To characterize the properties of the three categories of export goods, we use BASIC, CONSUM, and INVEST as the dependent variables in the regressions. All the variables described above are used as independent variables in the estimation. Since BASIC, CONSUM, and INVEST are dummy variables which are bounded between 0 and 1, we report OLS as well as LOGIT estimates. Consider ®rst the properties of basic goods given in speci®cations 1 and 2 of the table. BASIC is a dummy variable that takes the value of 1 if the export good is a basic good or chemical product and 0 otherwise. The negative coef®cients on MDIF and the positive on MINF suggest that basic goods tend to be homogeneous goods that are traded on an organized market. These variables suggest that the moral hazard and adverse selection problem of low quality is of little concern when basic goods are used as medium of payment. Furthermore SPECINV indicates that A does not invest in the relationship when she is paid in basic goods (the coef®cient on SPECINV is negative and highly signi®cant). The negative coef®cient on SPECINV also indicates that basic goods are a desirable means of payment and collateral for A. This is additionally supported by the positive (but insigni®cant) coef®cients on MUSE and RELATION, which indicate that A need not invest in marketing good 2. Finally, the negative and signi®cant coef®cient on REPEAT, in turn, suggests that barter trade with basic goods as means of payment tend to take place in an anonymous setting, since the parties typically do not know each other from previous transactions. In sum, the data support that basic goods are relatively liquid goods that do not require a marketing investment on the buyer's side. Next consider the characteristics of consumer goods that are given in speci®cations 3 and 4 of table 4.1. CONSUM is a dummy variable equal to 1 if the good is a consumer good and equal to 0 if the good is a basic good or chemical product. Thus speci®cations 3 and 4 compare the properties of consumer goods relative to basic goods and chemicals. In contrast to basic goods, consumer goods tend to be differentiated goods in which the problem of information asymmetry arises (MINF has a negative signi®cant sign and MDIF a positive
Table 4.1 Ranking commodity money BASIC OLS (1)
CONSUM LOGIT (2)
INVEST
OLS (3)
LOGIT (4)
OLS (5)
LOGIT (6)
MINF
0.44 (0.000)
2.16 (0.000)
ÿ0.34 (0.000)
ÿ1.92 (0.001)
ÿ0.48 (0.000)
ÿ2.78 (0.001)
MDIF
ÿ0.12 (0.037)
ÿ0.75 (0.030)
0.19 (0.015)
1.20 (0.008)
0.15 (0.065)
0.72 (0.083)
MUSE
0.06 (0.383)
0.35 (0.373)
ÿ0.49 (0.000)
ÿ3.07 (0.000)
0.08 (0.350)
0.38 (0.385)
ÿ0.13 (0.027)
ÿ0.76 (0.031)
0.18 (0.028)
1.05 (0.021)
0.14 (0.084)
0.73 (0.087)
0.04 (0.628)
0.18 (0.671)
0.09 (0.310)
0.59 (0.255)
ÿ0.24 (0.038)
ÿ1.23 (0.039)
SPECINV
ÿ0.26 (0.031)
ÿ1.77 (0.047)
0.37 (0.005)
2.45 (0.009)
ÿ0.01 (0.983)
0.15 (0.927)
Intercept
0.34 (0.000)
ÿ0.63 (0.040)
0.46 (0.000)
ÿ0.41 (0.293)
0.49 (0.000)
ÿ0.10 (0.787)
F
7.4 (0.000)
9.0 (0.000)
7.0 (0.000)
0.15
0.27
0.22
REPEAT RELATION
Adjusted R 2 ÿ2 LL Percent correct Number of observations
211
215.9
135.9
143.0
75.8
73.9
72.5
211
134
134
131
131
Source: Sample of 230 countertrade contracts. Notes: Ordinary least square and logit regressions. Numbers in parentheses are p-values. BASIC Dummy variable equal to 1 when export good from EE/LDC country is basic good or chemical product CONSUM Dummy variable equal to 1 when export good from EE/LDC country is consumer good INVEST Dummy variable equal to 1 when export good from EE/LDC country is investment and machinery; investment goods include the SITC 7 (standard international trade classi®cation) MINF Dummy variable equal to 1 when export good from EE/LDC country is investment and machinery MDIF Dummy variable equal to 1 when export good from EE/LDC country is horizontally and/or vertically differentiated MUSE Dummy variable equal to 1 when export good from EE/LDC country is used by industrialized country ®rm in own production REPEAT Dummy variable equal to 1 when industrialized country ®rm and EE/LDC trade partner have interacted frequently RELATION Dummy variable equal to 1 when export good from EE/LDC country is technologically related to import good from industrialized country ®rm SPECINV Dummy variable equal to 1 when export good from EE/LDC country is purchased at lower price in barter by industrialized country ®rm and not sold locally
The Export Pattern: Paying with Goods
67
signi®cant sign). However, B's incentive to pay with low-quality consumer goods is controlled by the fact that A and B are familiar with each other from previous transactions (REPEAT has a positive and signi®cant sign). Furthermore there seems to be evidence that A reduces her quality uncertainty by investing in the relationship, becoming more active in the quality design of the good and establishing a market for it. Both MUSE and SPECINV have a negative and signi®cant coef®cient. Only RELATION is insigni®cant. This investment makes the consumer good a medium of exchange speci®cally designed as payment for A. In sum, consumer goods are less liquid and less anonymous goods than basic goods. The moral hazard problem of quality gets alleviated, however, because A invests in the relationship and because part of B's behavior is governed by trust due to A's and B's repeated interaction. Finally, consider the properties of investment goods given in columns 5 and 6 of table 4.1. They differ in one important dimension from consumer goods. A typically does not invest in making the medium of exchange more liquid and less anonymous when she is paid in investment goods. The coef®cient on MUSE and SPECINV is insigni®cant. RELATION is negative and signi®cant. Investment goods tend to be disposed by A on the local market. A typically accepts investment goods as payment when she knows B from previous transactions. In sum, investment goods are less liquid and more anonymous goods than consumer goods. Moreover investment goods seem to be less anonymous than basic goods (the coef®cient on SPECINV is positive, although insigni®cant, while negative in the BASIC equation).5 The properties of commodity money described in table 4.1 suggest a hierarchy with respect to goods' quality problems and therefore with respect to their usefulness as collateral goods. According to the table, basic goods are expected to be the most liquid medium of exchange, followed by consumer goods. The most illiquid medium of exchange with severe quality problems are expected to be investment goods. We examine this ®nding with data on the trader's judgment of the quality of export goods compared to an average quality of the same goods on the market. Table 4.2 indeed supports the ®nding that commodity money can be ranked by quality problems. 58 percent of all basic goods and chemical exports, 49 percent of all consumer good exports, but only 31 percent of all investment good exports in
68
Barter Contracts in International Trade
Table 4.2 Liquidity of commodity money Investment goods
Consumer goods
Basic goods
Excellent to good quality
(23.2) [31.0]
(38.9) [49.3]
(37.9) [58.1]
95 [45.7]
Others
(43.4) [69.0]
(33.6) [50.7]
(23.0) [41.9]
113 [54.3]
Column total
71 (34.1)
75 (36.1)
62 (29.8)
208 (100.0)
Row total
Source: Sample of 230 countertrade contracts. Notes: Numbers without fences are total numbers of cases; numbers in parentheses are row percentages; numbers in brackets are column percentages.
barter trade of our sample were ranked to be of excellent to good quality relative to a market standard. 4.2
Economic Incentives and Trade Pattern
In this section we analyze how a country's incentive problems of debt repayment affect the choice of collateral goods used as payment in barter contracts. For this purpose, consider once more condition (4.1):
1 ÿ p^
v2 ÿ c2 a b c1 :
4:3
From this condition it can be seen that the larger the country's creditworthiness problem (i.e., the smaller a), the larger the value of the collateral good has to be. The value of the collateral in turn will be the larger the more liquid the good, captured by v2 ÿ c2 , and the less anonymous the good, captured by p^, used as payment. The preceding section gives us no clear ranking of basic and consumer goods. Basic goods are in principle more liquid than consumer goods but buyers have to undertake an investment to ensure the quality of the consumer good and as table 4.2 suggests, this attempt tends to be successful. Moreover the investment makes the consumer good less anonymous compared to basic goods. The good becomes speci®cally designed to be used only by the creditor, thereby reducing its attractiveness for potential other parties. Investment goods instead seem to be clearly inferior as payment in kind. When paid with investment goods the buyer does not invest to market the good
The Export Pattern: Paying with Goods
69
and indeed the quality delivered is typically worse compared with consumer goods. Thus we expect that the lower is the debtor country's creditworthiness, the more likely it is that basic goods and consumer goods and the less likely it is that investment goods are used to collateralize future payments. To test this hypothesis, we use as a proxy for the creditworthiness of the debtor country the debt to GDP ratio, DEBT, as reported by the World Debt Tables of the World Bank. The idea is that the more B is indebted already, the fewer assets remain to be seized by A in case of default, and thus the lower B's creditworthiness.6 The results are given in table 4.3. We report LOGIT estimates because the dependent variables measuring the choice of collateral goods BASIC, CONSUM, and INVEST are bounded between 0 and 1. As expected, we ®nd in speci®cation 1 that DEBT has a signi®cant positive impact on the choice of basic goods as opposed to all other goods. The coef®cient on DEBT is insigni®cant in speci®cation 5, indicating that debt does not make consumer goods more or less preferable to basic goods. However, investment goods are clearly not used if outstanding debt is large as the signi®cant negative coef®cient in speci®cation 9 indicates. In speci®cations 2, 6, and 10 we introduce a regional dummy to control for the possibility that the product choice is driven by comparative advantage. REGION is a dummy variable equal to 1 if the debtor country is a developing country and equal to 0 if it is an Eastern European country. REGION is positive and signi®cant for basic goods, negative and signi®cant for consumer goods, and insigni®cant for investment goods. In accordance with comparative advantage, developing countries use basic goods while Eastern Europe uses consumer goods to collateralize future payments. Including this regional dummy reduces the signi®cance of the DEBT coef®cient in the equations for basic goods and for consumer goods. This suggests that the choice of basic goods relative to consumer goods is driven more by comparative advantage than by the incentive problems of debt repayment. The reverse picture emerges for investment goods. The regional dummy is insigni®cant. We therefore exclude REGION for the remaining speci®cations for investment goods. The signi®cant effect of debt in the investment choice equations suggests that investment goods are used as means of payment only when the country has a creditworthiness problem. Comparative advantage does not seem to affect the choice. The negative coef®cient of DEBT, however,
70
Table 4.3 Choosing collateral goods BASIC (1) lnDEBT
0.5 (0.02)
CONSUM (2)
(3)
(4)
0.2 (0.26)
0.3 (0.40)
0.2 (0.47)
1.8 (0.00)
2.0 (0.00)
1.9 (0.00)
REPEAT
ÿ0.4 (0.40)
lnTECHIMP
REGION
ÿ0.2 (0.36)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
ÿ0.6 (0.00)
ÿ0.4 (0.17)
ÿ0.7 (0.01)
ÿ0.7 (0.01)
ÿ0.8 (0.01)
ÿ0.8 (0.02)
0.0 (0.90)
0.2 (0.59)
0.2 (0.52)
ÿ1.8 (0.00)
ÿ2.0 (0.00)
ÿ2.0 (0.00)
ÿ0.7 (0.14)
0.1 (0.86)
ÿ0.0 (0.96)
1.5 (0.00)
1.5 (0.00)
1.5 (0.00)
1.6 (0.00)
ÿ1.6 (0.02)
ÿ1.4 (0.06)
1.3 (0.06)
1.2 (0.13)
1.0 (0.13)
0.9 (0.22)
0.5 (0.51)
0.6 (0.47)
0.3 (0.58)
0.4 (0.44)
ÿ0.1 (0.83)
ÿ0.2 (0.65)
ÿ9.6 (0.65)
0.2 (0.61)
ÿ0.2 (0.76)
RISK
0.3 (0.42)
ÿ0.3 (0.49)
0.4 (0.02)
0.4 (0.02)
COMPETE
0.3 (0.50)
ÿ1.5 (0.01)
ÿ1.3 (0.02)
ÿ1.3 (0.02)
1.0 (0.73)
1.1 (0.72)
Intercept
ÿ3.1 (0.00)
ÿ0.85 (0.35)
ÿ2 LL
206.6
190.4
166.5
157.9
187.1
82.3
84.9
86.2
86.4
58.8
Percent correct Number of observations
226
226
1.8 (0.41)
217
2.3 (0.41)
206
0.8 (0.23)
136
0.5 (0.53)
ÿ4.2 (0.08)
ÿ4.8 (0.10)
171.9
156.9
146.1
66.2
69.5
68.3
136
128
126
2.1 (0.00)
ÿ7.9 (0.71)
ÿ0.8 (0.78)
ÿ0.9 (0.76)
178.4
146.7
151.0
150.8
132.4
132.2
58.8
70.6
70.8
65.4
70.6
70.6
136
136
130
130
126
126
Barter Contracts in International Trade
lnEXPORT
(5)
INVEST
Notes: Logit regressions. Numbers in parentheses are p-values. BASIC Dummy variable equal to 1 when export good is a basic good or chemical product, and equal to 0 otherwise CONSUM Dummy variable equal to 1 when export good is consumer good, and equal to 0 if it is a basic good or chemicals INVEST Dummy variable equal to 1 when export good is investment good, and equal to 0 if it is a basic good or chemicals DEBT Debt to GDP ratio of LDC or EE country in 1987 REGION Dummy variable equal to 1 when trade partner is located in LDC country, and equal to 0 if he is located in EE country REPEAT Dummy variable equal to 1 when LDC or EE country imported from developed country ®rm on a regular basis, and 0 otherwise TECHIMP Share of technology imports in total imports of EE/LDC country in 1987, UN ®nancial statistics EXPORT Exports to GDP ratio of LDC/EE country in 1987 RISK Dummy variable equal to 1 when parties agreed on a ®xed price over contract duration, and equal to 0 otherwise COMPETE Dummy variable equal to 1 when there are only a small number of suppliers for the barter good worldwide, and equal to 0 otherwise
The Export Pattern: Paying with Goods
Source: Sample of 230 countertrade contracts.
71
72
Barter Contracts in International Trade
suggests that investment goods will be an effective collateral only when the creditworthiness problem is mild. In speci®cations 3, 7, and 11 we include three more variables that may affect the choice of collateral goods. To control for reputation effects, we include REPEAT as a proxy for whether or not the parties know each other from previous interactions. REPEAT measures the frequency of A's exporting activity to B. The underlying presumption is that if A has exported to B regularly in the past, she is more likely to continue to do so in the future than if she trades with B for the ®rst time. One might expect that problematic goods like investment goods are more likely to be chosen if B has an incentive to maintain a reputation for high product quality. This is con®rmed by table 4.3, where REPEAT has a positive and signi®cant coef®cient in speci®cations 11 to 14, while it appears insigni®cant in the other speci®cations. As a proxy for B's export opportunities in case of default we use the export ratio of B's country EXPORT. A high export ratio suggests that the country is well integrated into the world market indicating that B's outside option to barter is good; that is, B can sell easily to another party. This should make it more important to choose goods of high collateral value, like basic and consumer goods. However, EXPORT turns out to be not signi®cant in any of the speci®cations. Finally, we include TECHIMP as a proxy for B's bene®t from importing good 1. TECHIMP is the ratio of technology imports over total imports in B's country. A large share of technology imports indicates that B depends essentially on A's imports and that it will be particularly dif®cult for B to ®nd substitutes. B will be more reluctant to cheat on A, because A can punish by excluding B from future imports. Thus a high ratio of technology imports makes it less important to choose high-value collateral goods like basic or consumer goods. This is supported for basic goods, where the coef®cient is signi®cant and negative. For consumer goods and for investment goods, however, TECHIMP has either the wrong sign or is insigni®cant. An alternative interpretation is that the variable TECHIMP re¯ects comparative advantage rather than A's punishment potential in case of B's default on payment. Countries are more likely to export consumer goods rather than basic goods if they have access to more advanced technology. In the literature some complementary explanations for the export pattern of bartering countries have been proposed. In chapter 2 we reported Amann and Marin's (1994) argument that barter can serve
The Export Pattern: Paying with Goods
73
as an insurance mechanism if export revenues are stochastic. From this argument there follows that countries will pay in goods for which no organized futures markets exist; that is, they will choose in particular consumer and investment goods. Another potential explanation for the dominance of consumer and technology goods among barter exports is that barter offsets inef®ciencies due to imperfect competition. In order to control for these other possible explanations for the export pattern in barter we have included the variables RISK and COMPETE. RISK is a dummy variable of value 1 when the parties agreed on a ®xed price for the barter good over the duration of the contract. RISK captures the insurance motivation for barter. This variable has no explanatory power for the choice of basic or consumer goods. Its coef®cient is positive and signi®cant for investment goods. COMPETE is a dummy of value 1 when there are only a small number of suppliers for the barter good worldwide. The variable is supposed to capture market power reasons for the choice of barter goods. It is insigni®cant for the choice of basic goods, and it signi®cantly affects the choice of consumer and investment goods. However, consumer and investment goods tend to be used as means of payment when market power is low, as the negative sign of the coef®cient on COMPETE indicates. In chapter 5 we provide a further explanation for the use of nonstandardized export goods. In this chapter we explain the import pattern, where technology imports dominate, and show that barter trade can induce an ef®cient technology transfer, by using the barter good as a hostage for A's technology delivery. We show that this effect is easier to achieve if the hostage good is a differentiated good for which markets are dif®cult to establish. Accordingly this provides an additional explanation based on incentive problems for why differentiated goods like consumer and investment goods dominate among barter exports from Eastern European and developing countries when these countries have a large share of technology imports. This proves to be of some relevance, since TECHIMP positively affects the likelihood that consumer goods are used as means of payment and negatively the likelihood of the choice of basic goods.7 Let us turn again to the actual pattern of specialization of countertrade described in the introduction. As we have seen in table 2.2 among countertrade exports from developing countries and Eastern Europe consumer goods (32 percent) and investment goods (36
74
Barter Contracts in International Trade
Table 4.4 Creditworthiness and trade pattern Barter exports (in %)
Debt/GDP Mean
Standard deviation
Number of cases
Eastern Europe a Export patternb
86.2
33.8
26.4
194
Investment goods
38.7
28.8
24.1
75
Consumer goods
35.6
38.3
28.6
69
Basic goods
13.4
34.4
26.3
26
7.2 5.2
33.4 42.8
22.9 30.4
14 10 27
Chemicals Services Developing countries c Export patternd Consumer goods
13.8
76.6
72.1
22.2
51.4
33.4
6
Basic goods
51.9
97.1
81.2
14
Chemicals
25.9
57.4
72.6
7
39.1
37.7
221
Total Anova F 35:2 Marginal signi®cance 0.000
Source: Sample of 230 countertrade contracts. a. Former Soviet Union, former GDR, former Czechoslovakia, Hungary, Poland, former Yugoslavia, Romania, Bulgaria, and Albania. b. The percentages in the ®rst column refer to total barter exports from Eastern Europe. c. Brazil, Ecuador, Argentina, Nicaragua, Philippines, Indonesia, Malaysia, India, China, Israel, Iran, Egypt, Algeria, Syria, Cyprus, Togo, Zambia, and Zimbabwe. d. The percentages in the ®rst column refer to total barter exports from developing countries.
percent) dominate. How can we explain this pattern given the results of table 4.3? The table predicts that investment goods will be used to collateralize future payments only when the country's creditworthiness is not too bad. Consumer goods instead are equally good collaterals as basic goods and therefore can be used as substitutes for them. Our theory predicts that countries that differ in their creditworthiness will show a different pattern of barter trade. More speci®cally, countries with lower creditworthiness will use higher value collateral goods as means of payment in barter. Thus the large share of investment good exports in our sample let us expect that the
The Export Pattern: Paying with Goods
75
problem of creditworthiness, though present, is not too severe among the countries represented in the sample. This is indeed what we ®nd in the data. The debt to GDP ratios of the countries in the sample vary widely, ranging from 5 to 327 percent. However, because of the dominance of Eastern Europe in the sample, in particular the former Soviet Union and former Czechoslovakia, we have a relatively larger number of barter deals with countries that were not too severely indebted in 1987. Table 4.4 shows that 86.2 percent of the barter deals of the sample are with Eastern Europe and 13.8 percent with developing countries. Eastern Europe, with an average debt to GDP ratio of 33.8 in 1987, was substantially more creditworthy than the developing countries with an average debt to GDP ratio of 76.6. Eastern Europe has thus used investment goods as means of payment in barter because it could ``afford'' to use ``bad money'' with low collateral value. Investment goods provided collateral of suf®cient value compared to the gains from defaulting. Consumer goods turn out to be ``good money'' and qualify as collateral for countries with low creditworthiness. Thus developing countries with low creditworthiness use predominantly the most liquid goods, basic goods, as means of payment in barter. The large share of investment good exports among countertrade can be explained by the dominance of the former Soviet Union and the Czech Republic in the sample, whose creditworthiness problem has been comparatively less severe during the 1980s relative to that of the other countries in Eastern Europe and developing countries, in particular, Latin America.
This page intentionally left blank
5
The Import Pattern: Securing Technology Transfer
In this chapter we turn to an explanation of observation 2.4 that countertrade imports are predominantly technology goods.1 We reported in chapter 2 that the sale of technology is particularly prone to incentive problems. For example, the value of a plant or some equipment to the buyer is directly affected by actions of the seller that are dif®cult to specify in a contract, such as the level of care with which a plant is set up and workers are trained, or the supply of after-sale services and technological advances.2 Parsons (1985, 1987) proposes buyback contracts as a solution to this moral hazard problem. A buyback contract gives the industrialized country ®rm an incentive to spend more effort in setting up the plant properly in order to improve the quality of the output with which it is paid with. This argument has been modeled formally by Chan and Hoy (1991) and Choi and Maldoom (1992) as we have discussed in chapter 2. They show that the technological relation between the two goods is essential for this contract to work. However, as indicated in observation 2.3, buyback accounts for a surprisingly small fraction of all countertrade transactions. So some other mechanism must be at work that explains why the two trade ¯ows are tied. 5.1
Creating a Hostage
In this chapter we show that counterpurchase, the dominant form of countertrade, can solve the technology transfer problem even though the two trade ¯ows are not technologically related. Recall that in contrast to barter, counterpurchase is a form of countertrade where each trade ¯ow, the import and the export, is paid for in foreign exchange. We show, however, that like barter, counterpurchase can solve the creditworthiness problem, though monetary payments of
78
Barter Contracts in International Trade
both transactions take place. Interestingly we ®nd that the creditworthiness problem and the technology transfer problem are closely related. In particular, the existence of the creditworthiness problem is an important prerequisite for solving the technology transfer problem. We illustrate these arguments again by referring to our trade partners, A and B. Recall that B is a ®rm in Eastern Europe or in a developing country that now wants to import a whole factory, machinery, know-how, or other investment goods from an industrialized country. Since these are complex and sophisticated products, it is dif®cult to specify all aspects of these goods in a complete contingent contract. Only after the factory is in place B discovers after some time that he has been delivered a ``lemons'' technology. The products he can produce with this technology cannot compete with goods on the world market. As B ®nds it dif®cult to identify the quality of technology goods, he cannot legally condition his payment on A's quality choice. This means that A has little incentives to deliver highquality technology in the ®rst place. She may be tempted to undersupply ``quality'' (spare parts, technology advances, etc.) and blame adverse circumstances in the Eastern European or developing country for unsatisfactory performance. In the context of developing countries, this phenomenon has become known as the problem of technology transfer. In addition, as before, we take it as given that B, the Eastern European or developing country, lacks creditworthiness due to large debts and thus the industrialized country ®rm cannot be sure that the export is actually going to be paid for. How can the tying of an import ¯ow and an export ¯ow, in a counterpurchase contract mitigate these moral hazard problems? We argue that the export (the export of the Eastern European or developing country to the industrialized country) creates a ``hostage'' (in the language of Williamson 1983) that deters cheating on quality and defaulting on the payment in the original import. First of all, the export goods serve as deal-speci®c collateral that gives the Eastern European or developing country an incentive to ful®ll its ®nancial obligations. Second, the contract is designed such that the value of this collateral depends on the quality decision of the industrialized country ®rm in the ®rst transaction, though there is no technological relationship between import and export goods. This contractual arrangement makes the industrialized country ®rm internalize the externality its quality decision has on the Eastern European or developing country.
The Import Pattern: Securing Technology Transfer
79
To illustrate this, we suppose that A can deliver two different quality levels but B cannot verify this quality on delivery. Say that B agrees to deliver export goods to A that do not stand in any technical relation to the original technology sold to him. These export goods serve as collateral for the payment due for the import goods. The idea is now to make B's payment of the technology contingent on A's quality choice in the sense that if A delivers low-quality technology, B will not be able to produce the export good destined for A because he cannot generate enough revenues from the technology and thus has not enough funds to purchase the inputs required for the production of the export good. Thus, if A cheats on quality, she loses her collateral. Although import and export goods are not technologically related the countertrade contract establishes a ®nancial link. B can ®nance the production of the export good out of his revenues from using the imported technology and of liquidating his assets a. In order for B to hit a ®nancial constraint in case of bad quality, we need that B does not generate enough revenues and hence cannot ®nance the export good if low-quality technology is delivered, but can do so only if he receives high-quality technology. Our contractual solution to A's moral hazard problem uses the fact that B faces a ®nancial constraint. If B were able to produce the export good all by himself, independent of what A delivers, then A could not internalize the externality of her own quality decision on B. Thus the fact that barter is used by countries with a severe ®nancial constraint makes it particularly effective to solve the problem of technology transfer. Note that without the subsequent export the gains from trade of the technology import could not be realized. Thus, even if the second deal could be carried out more ef®ciently with a third party, such as a trading company specialized in marketing exports from developing or Eastern European countries, it may still be optimal to link the two transactions in a countertrade arrangement. It is possible, however, to renegotiate the subsequent export such that the technology exporting industrialized country ®rm markets the import through a specialized trading company without affecting the incentives in the ®rst transaction. A Formal Analysis Consider the two trade partners, A and B, and suppose that B wants to import a technology, like a turnkey factory, machinery, or some
80
Barter Contracts in International Trade
other investment good, from A. This technology, in the following called ``good 1`` or the ``import,'' is used by B to produce goods that are then sold on the world market. B's valuation of good 1 can be thought of as his pro®ts generated with good 1 (including spillover effects this technology may have on other industries). This valuation is higher, the more effort A spends in order to produce a high-quality product (after-sale service, spare parts, delivering technological upgrades, etc.). The problem is that quality is not observed by B when good 1 is delivered, nor is it veri®able to an outsider like a court; thus quality cannot be speci®ed unambiguously in a contract. For simplicity we assume that there are only two different quality levels, ``high'' and ``low'' quality.3 We normalize A's production costs for low quality and B's valuation for low quality to 0. Producing high quality increases A's costs by c1 and B's valuation by v1 .4 We assume that v1 ÿ c1 > 0 so that producing high quality is ef®cient. As in chapter 3 we assume that there is a second moral hazard problem related to B's payment for good 1. B does not have enough liquidity to pay for good 1 in cash at the date of delivery and thus needs to get a credit from A. Recall that a b 0 denotes the repayment that can be enforced by A where a represents assets that B's country holds abroad and that A can seize in case of default. However, the more severely indebted B's country, the more creditors have to share any assets that can be seized in case of repudiation, and thus the smaller will be a, the repayment that can be enforced by A. We assume that B is credit constrained on international capital markets, and that the payments A can enforce are not enough to cover A's cost to produce high quality, that is, a < c1 . Considered in isolation, the outcome of this double moral hazard problem must be inef®cient. It is neither possible to induce A to deliver high quality nor to get B to pay more than a < c1 . Thus, if at all, only the inferior-quality good can be traded. How can a countertrade arrangement help to overcome the double moral hazard problem? Suppose that the parties ®nd a second transaction to be carried out after B has realized his payoff from good 1 and has paid p1 . This time B is supposed to deliver one unit of another good, in the following called ``good 2`` or ``export,'' to ®rm A. Furthermore let us assume that A (a large company in a Western country) is neither liquidity nor credit constrained, and that the courts in A's country will enforce the payment for good 2. Thus, while B cannot be forced to carry out the second deal, there are no other contractual hazards.
The Import Pattern: Securing Technology Transfer
81
Let v2 denote A's valuation for good 2 and p2 denote the price that A has agreed to pay. In order to produce good 2, B has to incur investment and/or production cost c2 > 0 (in hard currency), for example, to buy machinery or inputs on the world market. In addition to the payoff p2 ÿ c2 , B may receive a nonveri®able bene®t b b 0 if the second trade takes place. The idea is that B learns how to market the good on the world market, that he acquires information about consumers' tastes or about the production technology, or that there is a goodwill effect: consumers learn about the quality of B's good that makes it more pro®table to sell other goods later on.5 The second transaction is interesting only if there are gains from trade, if v2 b ÿ c2 > 0. How does the design of the second transaction affect the incentives in the preceding technology trade? First, as we have already obtained in chapter 3, it generates an additional collateral for ®rm A. If by the time A has to pay p2 B has not fully paid p1 , A can withhold the difference from her own payment. Of course, this requires that B has an interest to deliver good 2 to A. The contract has to be designed such that B wants to deliver good 2 even if this means that his payment p1 becomes enforceable. Note that B cannot breach the contract and sell good 2 to a third party C in some other industrialized country. If this were to happen, A could refer to her countertrade contract, which promises good 2 to her, and use courts in C's country to seize good 2 as soon as it is delivered.6 In this chapter we assume that the probability of success of this legal action is suf®ciently high to make a separate deal between B and C unpro®table. It is straightforward to extend our analysis to the case where A may not succeed with probability 1, as we assumed in the preceding chapters. Second, the contract can be designed such that B's ability to ®nance the production cost c2 depends on the quality level of good 1. Suppose that A cheats on quality. Then B's pro®t is reduced by v1 . If p1 and p2 are chosen appropriately, this reduction in B's pro®ts will make it impossible for B to pay for c2 , so the second trade is lost.7 A might be willing to renegotiate and offer additional money to B in order to rescue the gains from the second transaction. However, cheating on quality is inef®cient. Therefore A will have to put in more money than she gained from cheating on quality initially. This way A can in fact be made residual claimant for her own quality decision. The sequence of events is summarized in ®gure 5.1. A and B negotiate prices p1 and p2 at date 0. At date 1, A chooses the quality level which determines B's pro®t from good 1. This decision is denoted by
82
Barter Contracts in International Trade
Figure 5.1 Time structure
q A f0; 1g, where q 1 denotes high quality and q 0 denotes low quality. At date 2, B chooses whether or not to pay p1 . If he does not pay p1 , A receives a. At the same time, B can invest c2 and produce good 2. This decision is denoted by d A f0; 1g. If d 1, then, at date 3, good 2 is delivered and A has to pay p2 ; otherwise, there is no trade and no payment. All payments are measured in date 3 money. The pro®t functions of A and B are given by UA p1 ÿ q c1 d v2 ÿ p2 ;
5:1
UB q v1 ÿ p1 d p2 ÿ c2 b;
5:2
respectively. Given our parameter assumptions, the ®rst-best transaction requires that A chooses q 1 at date 1 and B chooses d 1 at date 2. To achieve the ®rst-best transaction, the contract ( p1 ; p2 ) has to ensure that A chooses high quality, that B pays p1 , and that the second transaction takes place. To show under what conditions such a contract exists, we proceed in two steps. We ®rst consider the case where no renegotiation is possible. That is, if she delivered low quality, A cannot increase her payment p2 in order to rescue the second deal. This case applies when A is engaged in countertrade transactions repeatedly and has build up a reputation not to renegotiate. Then we discuss what happens if we allow for renegotiation, and if we consider the possibility of including third parties in the transaction, such as some company C that has a higher valuation for good 2 than A. We solve the game induced by a contract ( p1 ; p2 ) by backward induction, looking for conditions on prices and exogenously given parameters such that there is a unique subgame perfect equilibrium implementing the ®rst best. Note ®rst that without loss of generality, we can restrict attention to prices satisfying p1 < p2 a:
5:3
The Import Pattern: Securing Technology Transfer
83
In order to enforce p1 , A can seize a, and if the second deal takes place, A can retain the payment p2 . Thus the highest payment that can be enforced is a p2 , and it is a strictly dominated strategy for player B to pay more.8 Now consider B's decision at date 2 whether or not to carry out the second deal. If B decides to choose d 0, then it is clearly optimal to pay p1 if and only if p1 a a. On the other hand, if B chooses d 1, then he cannot avoid to pay p1 (given that equation 5.3 holds) because A can use p2 a as collateral. Thus B strictly prefers to carry out the second deal if and only if v1 ÿ p1 p2 b ÿ c2 > v1 ÿ minfp1 ; ag:
5:4
Note that B's incentive to carry out the second deal is independent of whether or not A delivered high quality in period 1. This is due to the fact that there is no technological relationship between goods 1 and 2. However, A's quality decision has an impact on B's ®nancial constraint. In order to ®nance the investment and production cost c2 , B can use his pro®ts from the ®rst transaction q v1 ÿ p1 and he can take up a credit up to the amount p2 a, which is secured by the payment p2 he is going to receive and his assets a.9 Thus, if we want B to be able to ®nance the second transaction if and only if A delivered high quality (q 1), then p1 and p2 must be chosen such that10 v1 ÿ p1 p2 a b c2 ;
5:5
ÿp1 p2 a < c2 :
5:6
Next, consider A's decision at date 1 whether or not to produce high quality. If (5.4), (5.5), and (5.6) hold, then the second transaction will take place if and only if q 1. Thus A strictly prefers to deliver high quality if and only if p1 ÿ c1 v2 ÿ p2 > minfp1 ; ag:
5:7
Finally, at date 0 both parties have to participate voluntarily in the countertrade transaction. A's and B's participation constraints are given by p1 ÿ c1 v2 ÿ p2 b 0;
5:8
v1 ÿ p1 p2 b ÿ c2 b 0;
5:9
respectively.
84
Barter Contracts in International Trade
To summarize, if for a given pair of prices
p1 ; p2 conditions (5.3) to (5.7) are satis®ed, then there is a unique subgame perfect equilibrium in which A chooses high quality, B pays p1 , and the second trade takes place. Furthermore, if (5.8) and (5.9) hold, then both parties bene®t from the countertrade transaction. The following proposition gives necessary and suf®cient conditions for the existence of an ef®cient countertrade contract: proposition 5.1 There exists a countertrade contract
p1 ; p2 which implements the ®rst best if and only if the following conditions hold: i.
v2 b ÿ c2 > c1 ,
ii. v2 ÿ c1 v1 a b c2 , iii. v2 > c1 ÿ a, iv. b > 0. Proof
See appendix I-A.
Note ®rst that the set of parameters satisfying i to iii is not empty. As an example, choose v2 large enough and c2 suf®ciently small such that v2 ÿ c2 > c1 , then i to iii hold. The proposition also requires that b, the nonveri®able bene®t of B from the second transaction, is strictly positive. If b 0 and i to iii hold, then there is still a countertrade contract ( p1 ; p2 ) and a subgame perfect equilibrium given ( p1 ; p2 ) in which the ®rst best is achieved. However, this equilibrium is no longer unique. In this case (5.4) and (5.6) hold with equality, so B is indifferent whether or not to carry out the second deal, and he could do so even if A delivered low quality. Thus, requiring that b > 0 is essentially a tie-breaking rule guaranteeing a unique subgame perfect equilibrium.11 Of course, countertrade is not always an ef®cient remedy to solve the underlying incentive problems. Considering i to iv again, it is easy to see that these conditions are less likely to be satis®ed if the incentive for A to cheat on quality, measured by c1 , is very big, and the minimum payment that can be enforced from B, a, is very small. Thus the incentive problems may not be too severe. On the other hand, countertrade is more likely to be an ef®cient institution, the higher are the gains from the second trade (i.e., the bigger v2 b and the smaller c2 ). All this is quite intuitive. Countertrade is supposed to solve the double moral hazard problem by introducing a hostage that is lost if either party cheats in the ®rst transaction. For this con-
The Import Pattern: Securing Technology Transfer
85
struction to work, it has to be the case that the hostage is suf®ciently valuable as compared to the gains from cheating. What can we say about the prices p1 and p2 that implement the ®rst best? If conditions i to iv are met, there exists a continuum of price pairs
p1 ; p2 satisfying (5.3) to (5.9). Note that for any such price pair with p1 b a, there exists another price pair
p10 ; p20 which also implements the ®rst best, where p10 a a, p20 p2 ÿ
p1 ÿ p10 and where c2 ÿ b a p20 a v2 ÿ c1 :
5:10
This price pair has a natural interpretation: B is asked to pay for the ®rst loan what can be enforced anyway and A is compensated for her costs in transaction 1 (mainly) through the second transaction. As condition (5.10) shows p20 has to be chosen such that the second deal is suf®ciently attractive to both A and B; that is, their incentive constraints (5.4) and (5.7) have to be satis®ed. An important question is whether a countertrade contract still achieves the ®rst best when A and B cannot commit not to renegotiate the initial contract. An opportunity for an ex post Pareto improvement arises off the equilibrium path. Suppose that A delivered low quality in the ®rst transaction. We have chosen p1 and p2 such that B is not able to ®nance the second deal in this case. However, A could offer to raise p2 , that is, put in additional money, in order to rescue the gains from trade from the second transaction. As we will show, there exists a renegotiation proof contract such that A prefers to deliver high quality rather than renegotiate to ensure that the export good is delivered if A's incentive to cheat on quality is not too big. Note that if A defaults on quality, she has to offer to increase p2 by at least D c2 p1 ÿ p2 ÿ a
5:11
to make sure that B will in fact be able to deliver good 2. If we want to make sure that A strictly prefers to deliver high quality and stick to the old contract, rather than cheating on quality in the ®rst transaction and putting in the minimum amount D, the following renegotiation constraint has to be satis®ed: p1 ÿ c1 v2 ÿ p2 > p1 v2 ÿ p2 ÿ D p1 v2 ÿ p2 ÿ
p1 ÿ p2 c2 ÿ a v2 a ÿ c2 ;
5:12
86
Barter Contracts in International Trade
or equivalently, p2 ÿ p1 < c2 ÿ c1 ÿ a:
5:13
Note that this condition implies (5.6). The contract
p1 ; p2 is ``renegotiation proof'' if the possibility of mutually bene®cial renegotiation does not upset the equilibrium, that is, if it satis®es condition (5.13). proposition 5.2 Suppose that A and B can renegotiate after A has chosen q but before B decides on d. A countertrade contract
p1 ; p2 that implements the ®rst best and that is renegotiation proof exists, if and only if in addition to i±iv the following conditions are met: v. c2 > c1 , vi. b > c1 . Proof
See appendix I-A.
Thus proposition 5.2 shows that the ®rst best can be implemented with a renegotiation proof countertrade contract if A's incentive to cheat on quality is not too big. Note that condition iv of proposition 5.1 is implied by vi. There may also be a possibility to renegotiate by including a third party in the contract. Suppose that there is another ®rm C in an industrialized country that is specialized in marketing good 2. Then there is scope for a Pareto improvement by including C into the countertrade contract. In Marin and Schnitzer (1995) we show that this form of renegotiation does not upset the ef®cient equilibrium and actually increases the scope for countertrade. Thus the main disadvantage usually associated with countertrade, namely that there has to be a double coincidence of needs, is not really a problem, since the parties can include a third party into the contract. In fact this is frequently done in reality. In our empirical investigation we found that most ®rms in industrialized countries that countertrade with Eastern European or developing countries have a division specialized in ®nding an appropriate buyer for the goods imported from these countries. 5.2
Empirical Evidence
In this section we discuss some testable predictions from our theoretical model, look for proxies for the variables we would like to
The Import Pattern: Securing Technology Transfer
87
measure, and estimate whether or not the derived predictions are consistent with our data on actual countertrade contracts.12 In chapter 3 we argued that barter can be used to collateralize trade credits. In this chapter we have seen that counterpurchase can serve the additional function to secure ef®cient technology transfer. If counterpurchase serves different functions than barter, then we expect this to be re¯ected by the variables that explain the choice between barter and counterpurchase. We expect that counterpurchase rather than barter will be chosen if the Eastern European or developing country ®rm wants to make sure that the industrialized country ®rm provides ef®cient quality.13 This leads to our ®rst hypothesis. hypothesis 5.1 (Choosing the contract) The contract will be counterpurchase rather than barter when there is a problem inducing the industrialized country ®rm to deliver high-quality technology goods. To estimate the effect of contract characteristics on the choice of contractual form, we use both OLS and logit regressions. Table 5.1 presents the results of testing hypothesis 5.1 using the choice between counterpurchase versus barter as the independent variable. A's incentive to cheat on quality is proxied by the variable XCLASS. XCLASS is a dummy variable that takes the value of 1 if the good exported by the industrialized country ®rm is a factory or machinery and 0 in all other deals. The underlying presumption is that in case of technology exports the incentive problems are more severe. The coef®cients for XCLASS prove to be consistently positive and significant. This suggests that the presence of technology increases the probability that the contract form will be counterpurchase rather than barter. We also include DEBT as a measure for B's creditworthiness in our regressions. DEBT stands for the ratio of debt over GDP of the country in which B is located. The coef®cients for DEBT have a negative sign (though statistically insigni®cant in speci®cations 1 and 3) indicating that the parties will sign a barter agreement rather than counterpurchase when the Eastern European or developing country's debt problems are severe. In this case B's lack of creditworthiness cannot be overcome and only barter contracts with simultaneous exchange of goods against goods are feasible where in fact no credit is given. In speci®cations 2 and 4, XCLASS is replaced by XCLASS lnDEBT to focus on those cases of technology exports that were accompanied by high indebtedness, since it is for those cases that our
88
Barter Contracts in International Trade
Table 5.1 Counterpurchase versus barter: Dependent variable CPURCH OLS (1) XCLASS lnDEBT
Adjusted R 2 F
(2)
0.14 (0.005) ÿ0.03 (0.20)
XCLASS lnDEBT Intercept
Logit (3)
(4)
1.23 (0.007) ÿ0.07 (0.013)
ÿ0.34 (0.195)
0.05 (0.002) 0.89 (0.000)
0.99 (0.000)
0.04
0.05
5.3 (0.006)
6.3 (0.002)
ÿ0.53 (0.043) 0.36 (0.006)
2.39 (0.000)
3.06 (0.000)
ÿ2LL
132.9
132.2
Percent correct
88.4
88.4
Source: Sample of 230 countertrade contracts. Notes: Ordinary least square and logit regressions of 202 observations. Numbers in parentheses are p-values. CPURCH Dummy variable equal to 1 when contract is counterpurchase and equal to 0 when contract is barter XCLASS Dummy variable equal to 1 when import to EE/LDC country is factory or machinery and equal to 0 for all other imports DEBT Debt in percent of GDP of EE/LDC country
theory predicts counterpurchase to provide optimal incentives. The coef®cients have the expected positive sign and are signi®cant. A comparison of speci®cations 1 and 2 with 3 and 4 shows that both OLS and logit regressions generate very similar results, which con®rms that OLS is a good linear approximation in this problem. Our model predicts furthermore that the prices p1 and p2 are chosen to provide optimal incentives to both parties involved in the countertrade deal. Thus an important test of our theory is to check how actual contracts in our data sample react to different incentive problems identi®ed by our model. In simple barter agreements the value of the export is chosen in order to provide suf®cient payment for the value of the import from the industrialized country ®rm. This is re¯ected by the compensation ratio which measures the value of the export as a percentage of the import value. For barter contracts this ratio is typically around 100 percent. If counterpurchase were just another form of
The Import Pattern: Securing Technology Transfer
89
barter and motivated only by ®nancial constraints we would expect counterpurchase contracts to show a similar pattern of compensation ratios. However, the two patterns differ signi®cantly, indicating that counterpurchase is driven also by other motivations, as our theory suggests.14 Note that the model does not allow for point predictions of p1 and p2 . However, counterpurchase implements an ef®cient allocation only if p2 ÿ p1 can be chosen in an interval the bounds of which are given by constraints (5.3) to (5.8). Suppose that the upper bound of this interval is reduced because of a change in the exogenous variables of the model. If p2 ÿ p1 was close enough to the upper bound, it will have to go down as well. Similarly, if a change in the exogenous variables leads to an increase of the lower bound, p2 ÿ p1 has to go up. This observation allows us to derive several testable predictions for p2 ÿ p1 . hypothesis 5.2 (Securing technology transfer) The greater is A's incentive to cheat on quality, the lower will be the compensation ratio. Recall that in our model we have normalized quantities to 1, whereas in our sample multiple items may have been traded. Thus the relevant variables to look at in our data set are import and export values. Since some of the deals in our sample are extremely large and others are rather small in value, the differences of import value and export value vary a lot over different contracts. To control for possible size effects, we normalize these differences by dividing each difference by its export value, which yields as a proxy the compensation ratio of
import value=
export value. An increase in c1 reduces the right-hand side of (5.7) and leaves all other constraints unaffected. Thus p2 ÿ p1 is restricted to be smaller. Intuitively, when A's incentive to cheat on quality increases, the contract needs to give A a larger pro®t in transaction 2 in order to make sure that A is interested in the second deal. This calls for a lower net transfer to B. As above, A's incentive to cheat on quality is captured by XCLASS, a dummy variable equal to 1 if the export good is a technology related product. hypothesis 5.3 (Restoring creditworthiness) The lower is B's creditworthiness, the larger will be the compensation ratio. As in the preceding chapters, a decrease in a increases the lefthand sides of (5.3) and (5.5) and the right-hand side of (5.6), and thus is associated with a higher p2 ÿ p1 . The variable a stands for the
90
Barter Contracts in International Trade
assets A can seize in case of B's default. Intuitively, the smaller that the contractable collateral is, the lower B's creditworthiness and thus the larger the collateral generated by the second deal needs to be. As a proxy for B's creditworthiness, we use again DEBT, the ratio of debt over GDP, assuming that B's creditworthiness decreases as DEBT increases. hypothesis 5.4 (Using export spillovers) The smaller is B's unveri®able bene®t from carrying out the second transaction, the higher will be the compensation ratio. A decrease in b increases the left-hand side of (5.4) and does not affect any other constraint. So p2 ÿ p1 has to be bigger. The reason is that as the spillover bene®t on B's future transactions becomes smaller, A's net transfer to B must become higher in order not to violate B's incentive to carry out the second deal. We use MDIF as a measure for the gain in goodwill, b, that B has from carrying out the second deal. MDIF is a dummy variable that is equal to 1 if the good imported from the Eastern European or developing country is horizontally and/or vertically differentiated, and 0 otherwise. The idea is that when B as a new entrant on A's market sells differentiated goods, he has to convince consumers about the quality of his products. Consumers learn the quality of these goods by using them. Therefore, when good 2 is differentiated, sales on A's market generate an informational externality about the quality of good 2, thereby contributing to B's future pro®ts. hypothesis 5.5 (Valuing the export good) The smaller is A's pro®t from marketing good 2, the smaller will be the compensation ratio. A reduction of v2 decreases the right-hand sides of (5.7) and (5.8) and forces p2 ÿ p1 to become smaller. A lower valuation of the second good makes it more dif®cult to satisfy A's incentive and participation constraints, requiring a lower net transfer from A to B. We proxy the pro®tability with which ®rm A can sell good 2 with the dummy variable MPOS. MPOS is a dummy variable set equal to 1 if the Eastern European or developing country ®rm is a leading producer of the good exported to the industrialized country, and 0 in all other cases. It is assumed here that if the market for good 2 is highly concentrated, then A's pro®t from marketing good 2 will be bigger. Table 5.2 presents the results of testing hypotheses 5.2 to 5.5 for counterpurchase contracts (excluding barter and buyback). The
The Import Pattern: Securing Technology Transfer
91
Table 5.2 Determining the degree of tying: Dependent variable lnCOMP (1)
(2)
(3)
(4)
lnDEBT
0.38a (4.88)
0.39a (4.96)
0.34a (4.18)
XCLASS
ÿ1.10a (6.70)
ÿ1.02a (6.31)
ÿ1.00a (6.19)
ÿ0.27b (1.93)
ÿ0.29b (2.08)
ÿ0.45a (3.05)
0.38 (1.81)
0.49b (2.18)
MDIF MPOS XCLASS lnDEBT Intercept F Adjusted R
2
0.59a (5.02)
ÿ0.01a (3.17) 3.31a (11.38)
3.37a (11.28)
3.47a (11.57)
2.22a (6.34)
37.9a
26.6a
21.1a
12.4a
0.30
0.31
0.33
0.22
Source: Sample of 230 countertrade contracts. Notes: Ordinary least square regressions of 176 observations. Numbers in parentheses are t-values. Levels of signi®cance: a 1 percent, b 5 percent. COMP Export value from EE/LDC in percent of import value to EE/LDC DEBT Debt in percentage of GDP of EE/LDC country XCLASS Dummy variable equal to 1 when import to EE/LDC is factory or machinery and equal to 0 for all other imports MDIF Dummy variable equal to 1 when export from EE/LDC is horizontally and/ or vertically differentiated product and equal to 0 when export is a homogenous product MPOS Dummy variable equal to 1 when EE/LDC exporter has leading market position and equal to 0 for all other export goods
coef®cients of XCLASS are negative and signi®cant for all speci®cations, con®rming hypothesis 5.2. When the industrialized country ®rm delivers technology, the contracts are designed with lower compensation ratios to increase the technology seller's stake in the second deal.15 The coef®cients of DEBT have the expected positive sign, as posed by hypothesis 5.3, and are signi®cant for all speci®cations. In speci®cation 2 we include MDIF in our regression to test hypothesis 5.4. The negative sign of the coef®cient con®rms hypothesis 5.4 and the inclusion of MDIF improves the overall statistical properties of the regression. Hypothesis 5.5 is tested in speci®cation 3. The estimated coef®cient of MPOS is consistent with the hypothesis that a higher pro®tability of marketing good 2 tends to increase the
92
Barter Contracts in International Trade
compensation ratio. Finally, in speci®cation 4, XCLASS is again replaced by XCLASS lnDEBT to focus on those cases of technology imports that were accompanied by high indebtedness. This leads to a loss in R 2 but at the same time improves the statistical properties of all other variables included in the regression. In this chapter we have seen that countertrade can solve the incentive problems on both the seller's and the buyer's side. Linking the export to the import creates a hostage, which induces both trading partners to ful®ll their contractual obligations. If the industrialized country ®rm delivers poor technology or if the Eastern European or developing country defaults on its payment this hostage is jeopardized. The countertrade contract establishes a ®nancial link such that the Eastern European or developing country can produce and deliver the export good only if the technology received is suf®ciently valuable. We have shown that the fear of losing the hostage, namely of losing a pro®table export deal, induces the technology ®rm in the industrialized country to internalize the externality its quality decision has on the Eastern European or developing country. Furthermore the export goods serve as a deal-speci®c collateral that gives the Eastern European or developing country an incentive to ful®ll its ®nancial obligations.
Appendixes to Part I
Appendix I-A Proof of Proposition 3.2 To see that (3.17) is necessary for restoring B's creditworthiness, note ®rst that (3.16) is equivalent to 1 0 a v2 ÿ c1 d
1
and then that (3.15) is equivalent to
1 ÿ p^
v2 ÿ c2
1 ÿ pv1 pc1 ÿ
1 ÿ da b v2 :
2
Thus both conditions can be ful®lled simultaneously only if 1
1 ÿ p^
v2 ÿ c2
1 ÿ pv1 pc1 ÿ
1 ÿ da b c1 ; d
3
which is equivalent to (3.17). To see that (3.17) is also suf®cient, recall ®rst that condition (1), which is equivalent to (3.16), is satis®ed by assumption. Thus, as long as (3.17) is ful®lled, the only possible problem that can arise is that condition (3.15) is violated, that is,
1 ÿ p^
v2 ÿ c2
1 ÿ pv1 pc1 ÿ
1 ÿ da a v2 :
4
If this were the case, then A could induce B to deliver good 2 by making a monetary side payment s conditional on B's delivery such that
1 ÿ p^
v2 ÿ c2
1 ÿ pv1 pc1 ÿ
1 ÿ da s b v2 :
5
94
Barter Contracts in International Trade
As long as (3.17) holds, it is possible to ®nd a side payment s such that (5) and (1) are both satis®ed:
1 ÿ p^
v2 ÿ c2
1 ÿ pv1 pc1 ÿ
1 ÿ da b v2 ÿ s b
1 c1 : d
6 r
Proof of Proposition 5.1 Note ®rst that given our assumptions that v1 > c1 and that a < c1 , condition (5.9) is implied by (5.4) and can be ignored. The ®rst part of the proof is to show that (5.3) to (5.8) imply i to iv:
. Conditions (5.4) and (5.7) imply that c2 ÿ b ÿ minfp1 ; ag < p2 ÿ p1 < v2 ÿ c1 ÿ minfp1 ; ag:
7
This in turn implies that
i
v2 b ÿ c2 > c1 :
That is, the gains from trade from the second transaction have to be larger than A's gain from cheating on quality in the technology trade.
. Condition (5.5) requires that v1 ÿ p1 p2 a b c2 , meaning that B
has to be able to ®nance the second deal if A delivered high quality. Together with A's participation constraint (5.8), which puts an upper bound on the net payment from A to B by imposing that p2 ÿ p1 a v2 ÿ c1 , this implies that
ii
v2 ÿ c1 v1 a b c2 ;
saying that A's maximum transfer to B (v2 ÿ c1 ) plus B's own funds v1 a must be suf®cient to ®nance c2 .
. Condition (5.3) requires that p1 ÿ p2 < a; otherwise, p1 cannot be
enforced. On the other hand, (5.8) requires that p1 ÿ p2 b c1 ÿ v2 to make it worthwhile for A to participate. This implies that
iii
v2 > c1 ÿ a;
meaning that A's valuation of good 2 must be big enough to compensate her for the cost of producing high quality that are not covered by B's minimal repayment a.
. The incentive constraint for B (5.4) requires that p2 ÿ p1 > c2 ÿ b ÿ
minfp1 ; ag, while (5.6), the condition ensuring that B cannot ®nance
Appendixes to Part I
95
the second deal if A delivered low quality, imposes p2 ÿ p1 < c2 ÿ a. Together these conditions imply that
iv
b > a ÿ minfp1 ; ag b 0:
Next we have to show that if (i) to (iv) hold, then there exist
p1 ; p2 satisfying (5.3) to (5.8). Note that if p1 b a, then only the difference p2 ÿ p1 matters in conditions (5.3) to (5.8), and we are free to ®x the absolute value of p1 as we like. For example, we can set p1 a. Thus, without loss of generality, we can restrict attention to prices p1 a a. Rewriting conditions (5.3) to (5.8) and substituting p1 minfp1 ; ag, we get A 1 ÿa < p2 ÿ p1 ;
8
B 1 c2 ÿ b ÿ p1 < p2 ÿ p1 ;
9
C 1 c2 ÿ v 1 ÿ a a p 2 ÿ p 1 ;
10
p2 ÿ p1 < c2 ÿ a 1 D;
11
p2 ÿ p1 < c2 ÿ c1 ÿ p1 1 E;
12
p2 ÿ p1 a v2 ÿ c1 1 F:
13
These conditions impose some restrictions on the choice of p1 . Lower bounds on p1 are imposed by (5.4) and (5.6), p1 > a ÿ b 1 G;
14
and by (5.4) and (5.8), p1 > c2 ÿ b ÿ v2 c1 1 H:
15
Upper bounds on p1 are imposed by (5.3) and (5.7), p1 < v2 ÿ c1 a 1 I;
16
by (5.5) and (5.7), p1 < v2 ÿ c1 ÿ c2 a 1 J;
17
and by the fact that we restrict attention to prices p1 such that p1 a a 1 K:
18
We can ®nd a p1 within these bounds if and only if G < I, G < J, G < K, H < I, H < J, and H < K. G < I is implied by (i), G < J is
96
Barter Contracts in International Trade
implied by our parameter assumptions, G < K is implied by (iv), H < I is implied by (i) and (iii), H < J follows from (i) and (ii), and H < K follows from (i) and the fact that a b 0. Pick any p1 a a satisfying (14) to (17). There exists a p2 such that
p1 ; p2 satisfy (5.3) to (5.8) if A < D, A < E, A < F, B < D, B < E, B < F, C < D, C < E, and C < F. By the choice of p1 , we know already that A < E, B < D, B < F, and C < E. A < D is trivially satis®ed. A < F is equivalent to (iii), and B < E is equivalent to (i). C < D folr lows from v1 > 0. Finally, C < F is implied by (ii). Proof of Proposition 5.2 Let us ®rst show that (v) and (vi) are implied by (5.3) to (5.8) and (5.13):
. The renegotiation constraint puts a lower bound on the net payment from A to B: p2 ÿ p1 < c2 ÿ c1 ÿ a. Together with (5.3) we have
ÿa < p2 ÿ p1 < c2 ÿ c1 ÿ a;
19
which implies that
v c2 > c1 ; meaning that A's gains from cheating in the ®rst transaction have to be smaller than the total cost to produce good 2.
. Condition (5.4), the incentive constraint of B, imposes an upper
bound on A's net payment to B: p2 ÿ p1 > c2 ÿ b ÿ minfp1 ; ag. Together with the renegotiation constraint (5.13), this implies that c2 ÿ b ÿ minfp1 ; ag < p2 ÿ p1 < c2 ÿ c1 ÿ a;
20
which in turn implies that
vi b > c1 : Again, A's incentive to cheat on quality may not be too big. Next we have to show that if (i) to (vi) hold, then there exist
p1 ; p2 satisfying (5.3) to (5.8) and (5.13). Recall that (5.13) implies (5.6). The proof follows the lines of the proof of proposition 5.1 where (5.6) is replaced by (5.13): p2 ÿ p1 < c2 ÿ c1 ÿ a 1 D 0 :
5:13
Appendixes to Part I
97
This tightens the lower bound imposed on p1 . Conditions (5.4) and (5.13) impose p1 > c1 a ÿ b 1 G 0 :
21
Note that G 0 < I is implied by (i), (v), and (vi), G 0 < J is implied by (i) and our parameter assumption, and G 0 < K follows from (vi). Thus we can ®nd a p1 a a satisfying (15) to (17) and (21). Pick any such p1 . There exists a p2 such that
p1 ; p2 satisfy A < D 0 , A < E, A < F, B < D 0 , B < E, B < F, C < D 0 , C < E, and C < F. By choice of p1 , B < D 0 . A < D 0 is equivalent to (v), while C < D 0 is implied by our parameter assumption. The other inequalities have been checked in the proof of proposition 5.1 already. r Appendix I-B De®nition of Variables and Sample Statistics
Minimum
Maximum
Standard deviation
38.9
4.5
326.6
37.5
Compensation ratio: export value in percent of import value
71.4
2.0
400.0
51.4
226
Export to GDP ratio in 1987 of EE/LDC country
27.9
4.2
63.9
11.8
TECHIMP**
222
Share of technology imports in total imports in 1987 of EE/LDC country
33.5
11.6
45.4
7.3
BASIC
230
Type of good exported from EE/LDC country: basic goods or chemical product
D 1, 63 observations
CONSUM
230
Type of good exported from EE/LDC country: consumer goods
D 1, 77 observations
INVEST
230
Type of good exported from EE/LDC country: investment and machinery
D 1, 75 observations
Variable
Observations
DEBT*
Description
Mean
226
Debt to GDP ratio in 1987 of EE/LDC country
COMP
230
EXPORT*
98
Barter Contracts in International Trade
Variable
Observations
MDIF
224
Characteristics of export good from EE/LDC country: horizontally and/or vertically differentiated
D 1, 116 observations
MINF
85
Characteristics of export good from EE/LDC country: good traded on organized market
D 1, 30 observations
Description
Mean
Minimum
Maximum
Standard deviation
REGION
230
Region of debtor country: LDC
D 1, 31 observations
REPEATX
227
Status of industrialized country ®rm: frequent imports to EE/LDC trade partner
D 1, 149 observations
REPEAT
212
Status of industrialized country ®rm and EE/ LDC trade partner: frequent trade partners
D 1, 97 observations
MUSE
230
Usage of export good from EE/LDC country by industrialized country ®rm: in own production
D 1, 55 observations
RELATION
230
Characteristics of export good from EE/LDC country: technologically related to import good from industrialized country ®rm
D 1, 41 observations
HIGHCASH
230
Characteristics of export good from EE/LDC country: sold at higher price outside barter
D 1, 20 observations
LOCAL
230
Usage of export good from EE/LDC country: sold at local market
D 1, 71 observations
SPECINV
230
Export good from EE/ LDC country purchased at lower price in barter by industrialized country ®rm and not sold locally
D 1, 14 observations
RISK
221
Fixed price contract
D 1, 124 observations
Appendixes to Part I
99
Variable
Observations
COMPETE
220
Market structure for barter good: small number of suppliers wordwide
D 1, 38 observations
XCLASS
229
Type of product exported: factory and machinery
D 1, 173 observations
MPOS
221
Market position of EE/ LDC ®rm: leading producer
D 1, 32 observations
Barter
230
Contract type: barter
D 1, 26 observations
Description
Mean
Minimum
Maximum
Standard deviation
Buyback
230
Contract type: buyback
D 1, 28 observations
CPURCH
230
Contract type: counterpurchase
D 1, 176 observations
Sources: *World Debt Tables, World Bank; **UN ®nancial statistics.
This page intentionally left blank
II
Contract Enforcement in Transition Economies
This page intentionally left blank
6
Stylized Facts and Competing Explanations
In this part of the book we turn to domestic barter in transition economies. Domestic barter started to grow in the former Soviet Union in 1994, after the fall of communism and after macroeconomic stabilization had been achieved. There are three dominant features that distinguish the development of the countries of the former Soviet Union from those of the early transition economies like Hungary, Poland, and the Czech Republic.1 observation 6.1 (Output decline) The decline in output has been much more pronounced in the former Soviet Union as compared to the early transition economies. Figure 6.1 compares the output decline of Central and Eastern European countries with the countries of the former Soviet Union (Borenzstein et al. 1999). In the ®gure it can be seen that the former Soviet Union stands at roughly half of its 1989 GDP level while Central Europe's output reaches 80 percent of its 1989 level. observation 6.2 (Growth of arrears) Inter-®rm arrears are much larger and growing much faster in the Republics of the former Soviet Union as compared to the early transition economies. Figure 6.2 shows that total arrears (®rm, tax, and wage arrears) in percent of GDP in Russia exploded from around 5 percent in 1994 to 40 percent in 1998. observation 6.3 (Dominance of barter) The cross-country pattern. Barter trade has become a dominant phenomenon in the domestic economies of the former Soviet Union, while being less important in Central and Eastern Europe.
104
Contract Enforcement in Transition Economies
Figure 6.1 Output decline in transition economies. Source: Borenzstein et al. (1999).
Figure 6.2 Total arrears as percentage of annualized GDP in Russia. Source: Russian Economic Trends, Russian European Centre for Economic Policy (RECEP), 1999.
Stylized Facts and Competing Explanations
105
Table 6.1 Barter in transition economies (% of ®rms with a barter share of more than 25%) Former Soviet Union Armenia
Percentage of ®rms
Central and Eastern Europe
Percentage of ®rms
8.0
Bulgaria
10.0
Azerbaijan
13.9
Croatia
72.4
Belarus
10.0
Czech Republic
10.7
Georgia Kazakhstan
17.1 38.8
Estonia Hungary
Kyrgyzstan
40.9
Lithuania
Moldova
55.4
Poland
13.8
Russia
50.4
Romania
16.8
Ukraine
46.6
Slovakia
30.2
Uzbekistan
23.8
Slovenia
45.6
Total
30.49
8.3 2.7 8.0
20.15
Source: World Business Environment Survey, World Bank-EBRD 1999.
The time pattern. Barter trade has exploded in Russia from 8 percent in 1994 to 53 percent in 1998.2 The cross-country pattern is documented by table 6.1, while the time pattern can be seen in ®gure 1.1 in the Introduction. These developments raise the question of whether there is a connection between these three observations: the output decline, the growth of arrears, and the explosion of barter. More speci®cally, what is the relationship between the output decline and the inter®rm arrears, on the one hand, and inter-®rm arrears and barter, on the other? Has the presence of barter in the former Soviet Union and its absence in Central Europe something to do with the fact that inter-®rm arrears are much larger and output declined much more sharply in the former Soviet Union? In this chapter we discuss competing explanations for these three observations. For this purpose we take a closer look at Ukraine, which with 46.6 percent of ®rms with a barter share of more than 25 percent clearly stands out to be a transition economy in which barter plays a dominant role. By looking at one individual country, we will try to understand the evolution of inter-®rm arrears and barter over time. 6.1
The Data
We interviewed 55 ®rms in Ukraine in 1997 to obtain information on 165 barter deals. The interviewed ®rms were selected from an
106
Contract Enforcement in Transition Economies
address list of ®rms of the local of®ce of the Harvard Institute of International Development in Kiev. These ®rms were contacted by phone in order to secure their cooperation. The 55 ®rms of our sample were ®nally those who agreed to participate in the survey. The unit of analysis of the survey was one particular barter deal. Each ®rm provided us with information on three barter deals. Each barter transaction involves two ®rms, the seller and the buyer. The questionnaire asked for information on about forty dimensions on the ``sale'' and the ``goods payment'' side of the barter transaction. The ``sale'' was de®ned by the transaction of the seller who initiated the deal. The ``goods payment'' was the buyer's payment for the ``sale.'' Because of the length of the questionnaire, we personally visited these ®rms to ®ll in the questionnaire. The data come from three cities: Kiev (50 percent), Zaporioshje (30 percent), and Dnipropetrovsk (20 percent). Many of the ®rms were well informed about the ®nancial and economic conditions of the ®rms they traded with because they served as ®nanciers. This is why we could obtain data on more than the 55 interviewed ®rms. Thus, depending on the variable, the ®rm information in our sample varies between 69 and 160 observations. The appendix II-B gives the summary statistics of the variables used. We take a ®rst look at some of the features of barter and the bartering ®rms in Ukraine based on our survey of 165 barter deals in Ukraine in 1997 and then proceed to explore the most common explanations of barter given in the literature. Table 6.2 shows that barter accounts for on average 45 percent of ®rm's sales with a minimum barter share of 1 percent and a maximum share of 100 percent. The barter deals are typically large in size, ranging between U.S.$10 and U.S.$5,000,000 with a mean size of U.S.$135,679. Firm arrears make on average 30 percent of ®rms' sales with a maximum of 626 percent. The ®rms of the survey show a somewhat similar variation with respect to tax arrears. On average, ®rms ®nanced 6.31 percent of output by bank debt with a maximum of over 100 percent. Firms of all sizes undertake barter deals. Firm size as measured by the number of employees ranges from 8 persons to 130,000 persons with a mean of 4,387 persons. Our sample ®rms grew on average like the total economy of Ukraine between 1994 and 1996. Table 6.3 describes the commodity composition of barter deals. It shows that barter occurs especially in the machinery and vehicle
Stylized Facts and Competing Explanations
107
Table 6.2 Barter and ®rm characteristics Mean Relative ®rm growthb
Minimum
Maximum
Casesa
0.01
ÿ2.03
25.43
153
Total arrears in percent of output
41.40
1.00
687.90
138
Firm arrears in percent of output
30.15
0.70
626.0
Tax arrears in percent of output
7.15
121.5
0.00
138 150
Wage arrears in percent of output
3.38
0.00
38.6
150
Bank debt in percent of output Barter in percent of outputc
6.31
0.00
104.2
150
100
165
Size of barter deals in U.S. dollars Firm size by number of employees
45.21 135,679 4,386.6
1.00 10
5,000,000
150
8
130,000
160
Source: Survey of 165 barter deals in Ukraine in 1997. a. The number of ®rms exceeds the number of interviewed ®rms because each deal involves two ®rms (a seller and a buyer). The percentage given in the table is the mean over the total of selling as well as buying ®rms. b. Percentage deviation of ®rm's output growth relative to the growth rate of GDP between 1994 and 1996. c. The estimate of barter in percentage of output for Ukraine with 45.21 percent differs from the number given in World Business Environment Survey (see table 1.1) because our survey includes only bartering ®rms. Table 6.3 Commodity composition of barter deals (in %) Sale
Goods payment
Basic industrya Textile and leather
28.0 7.7
31.5 5.4
Wood and paper Machinery and vehiclesb Food and beverages
2.4 16.7
4.2 21.0
16.1
22.1
Chemicals
13.1
8.9
Services
14.3
0.6
Others
0.0
4.8
Missings
1.8
1.8
Total
100
100
Source: Survey of 165 barter deals in Ukraine in 1997. a. Metal ores, coke and petroleum, nonmetallic mineral products, basic metals, electricity, and gas and steam. b. Metal products, machinery and equipment, electrical machinery, electronics, precision instruments, and motor vehicles.
108
Contract Enforcement in Transition Economies
sector, in the food and beverages sector, and in the basic sector consisting of metal ores, coke and petroleum, nonmetallic mineral products, basic metals, electricity, gas, and steam. 6.2
Output Decline
In the ®rst decade after the end of communism, the output collapse in transition economies was effected in varying ways. The Transition Report 1999 traces these differences in its manifestation to four factors. Countries differed in their initial conditions, in the timing and scale of the introduction of comprehensive stabilization programs, in the extent of their structural reforms, and in the scale and scope of their institutional reforms. In a recent paper Blanchard and Kremer (1997) focus on the last of these four factors, the institutional environment. They argue that the large decline in output in the former Soviet Union has been caused by ``disorganization'' and holdup problems. Disorganization arises when old relationships break down before new ones can be established. In such a ``no future'' environment a typical mechanism to constrain opportunistic behavior such as reputation does not work. Speci®city in the relations between ®rms, together with incompleteness of contracts, results in disorganization in which intermediate producers in a chain of production refuse to deliver inputs and that in turn leads to the collapse of output. Output collapses due to an input shortage. A similar argument is made by Roland and Verdier (1999). In their model, output may fall because of market imperfections due to search frictions and Williamsonian relationship speci®c investments. Konings and Walsh (1999) examine empirically the impact of disorganization on growth. They ®nd that disorganization did not constrain employment and productivity growth in newly established private ®rms in Ukraine. Calvo and Coricelli (1995a, b), in turn, have argued that the output losses in the early transition economies like Central Europe have been caused by a lack of credit. They suggest that inter-®rm arrears in the early transition economies have been a response to the liquidity squeeze in the economy. Because of the lack of bank credits, ®rms turn to trade credits from other ®rms to alleviate the ®nancial squeeze. According to Calvo and Coricelli output collapses due to a ®nancial shortage. Their argument suggests that observation 1 is related to observation 2.
Stylized Facts and Competing Explanations
109
A number of recent empirical studies suggest that the input shortage explanation of the output decline based on the lack of trust and nonfunctioning legal institutions has some validity. Hay, Shleifer, and Vishny (1996) and Aslund (1995) document in their studies the lack of functioning legal institutions in the former Soviet Union. McMillan and Woodruff (2000) carried out a survey on contract enforcement among managers in Poland, Slovakia, Romania, Russia, and Ukraine. They obtain that just over two-thirds (68 percent) of those surveyed con®rmed that courts can enforce contracts with customers. Moreover only about 55 percent of Russian and Ukrainian managers con®rmed that the courts were effective in enforcing contracts. Their results indicate that whenever courts and trade associations are functioning they have a signi®cant impact on business trust.3 McMillan and Woodruff ®nd furthermore that ongoing customer-supplier relationships helped to enforce credit contracts. 4 In contrast to Hay, Shleifer, and Vishny (1996) and Aslund (1995), Hendley, Murrell, and Ryterman (2000) ®nd that courts are effectively used and that business people do have some con®dence in their legal system. However, some aspects of the legal systemÐnotably the lawmakers and the policeÐare poorly regarded by the business community and the public at large. 6.3
Inter-®rm Arrears
The literature on inter-®rm debt in transition economies asks the following question: Why are ®rms giving loans to other ®rms when the same ®rms are not considered creditworthy enough by the banks and therefore do not get loans from banks? The answer that is most commonly given is the absence of market discipline. State-owned ®rms that show the highest inter-®rm debt are seen to be able to get loans from other ®rms because of the soft budget constraint. If the state-owned ®rms are creditworthy because of the backing of the government, they are expected to be the least credit constrained and thus to show the highest bank debt. As table 6.4 shows, in our sample of bartering ®rms this is indeed the case. State-owned ®rms appear to have on average higher bank debt (7.5 percent of output), while private ®rms have negligible bank debt outstanding (0.1 percent of output). State-owned ®rms show also the highest inter-®rm arrears compared to private ®rms (68 percent and 24 percent of output, respectively). This suggests that state ®rms used their privileged
110
Contract Enforcement in Transition Economies
Table 6.4 Ownership, debt, and barter Percentage of output Bank debt
Firm debt
Tax arrears
Barter share
Domestic state enterprise
7.5
68.0
4.8
Domestic private ®rm
0.1
23.8
1.6
58.3
Foreign or GUS ®rm
Ð
Ð
0.0
48.0
Cooperative or collective ®rm
6.3
16.9
9.6
44.8
Worker Government
Ð Ð
Ð Ð
Ð Ð
50.4 10.8
Joint venture
3.0
13.7
0.0
34.6
Total
5.9
32.0
6.5
51.0
Mean value of respective variables
F-test Signi®cance level
0.5 (0.789)
1.5 (0.180)
1.2 (0.315)
56.6
4.2 (0.000)
Source: Survey of 165 barter deals in Ukraine in 1997.
status of creditworthiness to get cash credit from banks as well as trade credit from other ®rms. An explanation for the phenomenon of inter-®rm debt cannot, however, rest exclusively on the argument of soft budget constraint. Inter-®rm arrears are not a phenomenon of state ®rms alone. In our sample of bartering ®rms, only 29 percent are state controlled. Therefore there are additional forces at work that go beyond the lack of market discipline. We see this force in the problem of creditworthiness.5 The capital and credit market does not function well in transition economies for a variety of reasons. Creditors are inexperienced with credit evaluation. Banks have dif®culties in distinguishing bad from good debtors. There is no history to allow them to judge credit risk because of the drastic changes in the environment. In some of the transforming economies, a bankruptcy law has not yet been introduced. Defaulting on debt repayment remains without consequences, and therefore ®rms have little incentive to repay their loans from banks. Many experts have suggested that one of the solutions to inter®rm debt is to restore the creditworthiness of ®rms by introducing a bankruptcy procedure. But in many countries like Hungary and Ukraine, for example, a bankruptcy law has been introduced without any impact on levels of inter-®rm debt. Furthermore a study by
Stylized Facts and Competing Explanations
111
Mitchell (1993) suggests that the introduction of a bankruptcy law by itself will not improve debt repayment because creditors do not use the bankruptcy procedure to get to their money. Among other factors of Mitchell's explanation for this creditor's passivity is the low expected value of their claims net of bankruptcy costs. This is due to the poor state or vintage of the capital stock of a debtor ®rm, the absence of a market for capital, and the priority assigned to a creditor in bankruptcy relative to the ordering of other creditors. Mitchell's explanation suggests that inter-®rm debt is not going to go away with the introduction of a bankruptcy law and that creditor's passivity prevents bankruptcy from restoring the creditworthiness of ®rms. The question remains: Why are ®rms able to give loans to other ®rms when the banking sector is reluctant to provide capital despite the availability of a bankruptcy procedure to pursue nonpaying debtors? In the following chapter we set out to give an answer to this question, and we show that barter becomes important in this context. We now turn to explanations given for the rise of barter. 6.4
Barter
A number of different explanations for barter have been put forth in the recent debate on this phenomenon. Among these are soft budget constraints and delay in restructuring, the virtual economy and tax avoidance. Soft Budget Constraints and Restructuring One of the most popular explanations of barter in transition economies are soft budget constraints and the lack of market discipline.6 The absence of hard budget constraints leads managers and workers to avoid the costs arising from restructuring by maintaining production in inef®cient activities. Barter is seen as a possibility to conceal the true market value of output.7 Table 6.4 looks at the question whether barter can be explained by problems of corporate governance and/or mode of ownership. Are state-®rms using barter more often than private ®rms? Do they try to avoid restructuring by using barter to conceal the true value of output as has been claimed? The table demonstrates that barter does not seem to be a phenomenon of state-owned enterprises. Newly established private ®rms show the same or higher barter exposure as state-owned ®rms or cooperatives.
112
Contract Enforcement in Transition Economies
The average barter share of state enterprises is 56.6 percent and that of private ®rms 58.3 percent. In addition table 6.5 shows that there is no relationship between the barter intensity of the ®rm and the productivity of the ®rm, if at all the relationship is positive (the correlation coef®cient is 0.05). This suggest that avoiding restructuring by inef®cient and loss-making ®rms is not the prime reason for barter. Table 6.5 looks also at the relationship between arrears and the ef®ciency level of the ®rm. The different types of arrears in percent of output do appear to be declining with the productivity level of the ®rm. However, the correlation coef®cients between arrears and ef®ciency (given at the bottom of table 6.5) are near zero and not signi®cant at conventional levels except for the correlation between wage arrears and productivity. This evidence suggests that neither soft budget constraints nor a reluctance to move into ef®cient activities seem to be the driving force behind barter. The data do suggest, however, that very large arrears (®rm arrears of more than 50 percent of ®rm's output) tend to be a phenomenon of less ef®cient ®rms. In order to control for size effects in the relationship between productivity and barter, on the one hand, and productivity and arrears, on the other, table 6.6 calculates the relevant correlation coef®cients for different ®rm sizes. The correlation between barter and ef®ciency is near zero and insigni®cant. However, the correlation between the different types of arrears and ef®ciency becomes more negative and more signi®cant for all ®rm sizes except for medium-sized ®rms. The Virtual Economy8 We now turn to the ``virtual economy'' argument of Gaddy and Ickes (1998) which has been one of the most in¯uential explanations of barter in Russia. The virtual economy argument claims that barter helps to create the image that the manufacturing sector in Russia is producing value while in fact it is not. This argument rests on the assumption that the manufacturing sector is value-subtracting and most participants in the economy have an interest to pretend that it is not. Barter allows the parties to pretend by allowing the manufacturing sector to sell its output at a higher price than its market value and the value-adding natural resource sector (Gazprom) to accept this high price because of a lack of other sources. This way the manufacturing sector survives by drawing resources from the
Stylized Facts and Competing Explanations
113
Table 6.5 Barter, arrears, and ef®ciency Ef®ciencya Barter share in percentage of output
Mean
Standard deviation
Cases
45.61
28.41
1,500±7,000
48.18
29.91
153 57
7,100±15,000
44.05
27.59
60
15,100±140,000 F 0:366
44.17
27.82
36
41.42 69.60
101.07 150.16
138 57
Signi®cance level 0.694 Total arrears in percentage of output 1,500±7,000 7,100±15,000
17.89
15.13
48
15,100±140,000
26.97
38.44
33
F 4:024 Signi®cance level 0.020 Firm arrears in percentage of output
30.15
90.89
138
1,500±7,000
53.88
137.82
57
7,100±15,000
13.53
14.56
48
15,100±140,000
13.32
10.29
33
3.38 6.78
6.00 8.45
150 57
F 3:429 Signi®cance level 0.035 Wage arrears in percentage of output 1,500±7,000 7,100±15,000
1.71
1.80
57
15,100±140,000
0.63
1.67
36
F 18:701 Signi®cance level 0.000 Tax arrears in percentage of output
7.15
19.31
150
1,500±7,000
8.94
13.79
57
7,100±15,000
2.38
6.25
57
15,100±140,000
11.88
33.98
36
F 3:158 Signi®cance level 0.045 Source: Survey of 165 barter deals in Ukraine in 1997. Notes: The Pearson correlation coef®cient between the barter share and ef®ciency is 0.05, between total arrears and ef®ciency ÿ0.13, between ®rm arrears and ef®ciency ÿ0.12, between wage arrears and ef®ciency ÿ0.25, and between tax arrears and ef®ciency ÿ0.33. Except for wage arrears none of the correlations are signi®cant at conventional levels. a. Output in U.S.$ per employee
114
Contract Enforcement in Transition Economies
Table 6.6 Barter, arrears, and ef®ciency by ®rm size Ef®ciencya Small ®rmsb Medium ®rmsc Large ®rmsd
Barter share
Total arrears
Firm arrears
Wage arrears
Tax arrears
0.09 (0.43)
ÿ0.27 (0.04)
ÿ0.22 (0.10)
ÿ0.33 (0.00)
ÿ0.20 (0.10)
ÿ0.19 (0.22)
0.28 (0.08)
ÿ0.19 (0.22)
0.00 (1.00)
0.39 (0.01)
0.01 (0.98)
ÿ0.31 (0.08)
ÿ0.31 (0.08)
ÿ0.35 (0.05)
ÿ0.35 (0.04)
Source: Survey of 165 barter deals in Ukraine in 1997. Notes: The numbers are Pearson correlation coef®cients, and the number in parentheses give the signi®cance levels. a. Output in U.S.$ per employee. b. Output level between 0 and 4 billion U.S.$. c. Output level between 4 and 20 billion U.S.$. d. Output level between 20 and 500 billion U.S.$.
natural resource sector. According to the argument, keeping up the illusion of a value-adding manufacturing sector is highly costly for the Russian economy at large because this cross-subsidizing from the value-adding natural resource sector to the value-subtracting manufacturing sector prevents the manufacturing sector from moving into valuable activity. But, if the natural resource sector is producing valuable output, why does the sector not have other opportunities than to subsidize the manufacturing sector? In fact the natural resource sector is supposed to have signi®cant bargaining power in the interaction with other sectors when it is producing goods that the market values highly. Why then does the sector end up subsidizing the rest of the economy? The argument does not seem to be a tight one. The argument appeals to experts of central planning and policy observers in transition economies, because the practice of cross-subsidizing across different activities in the economy was a widespread feature of central planning. Therefore let us pretend for a moment that the virtual economy argument is a tight one, and let us see whether it is actually true. We can answer this question from our survey data, since we have information on the percentage price difference between the barter price and the cash price for each of the 165 barter deals in the sample. We have this information for both sides (the ``sale'' and the ``goods
Stylized Facts and Competing Explanations
115
Table 6.7 Terms of trade of the non-cash economy Differential between barter and cash pricea Sale ÿ17±0%
Goods payment ÿ50±0%
25.93
ÿ200±0%
10.49
0%
73.62
0%
62.96
0%
45.06
0±50%
23.31
0±200%
11.11
0±50%
44.44
Total
3.07
Terms of trade
100.00
100.00
100.00
Source: Survey of 165 barter deals in Ukraine in 1997. a. In percentage of cash price.
payment'') of each barter deal so that we can calculate the net terms of trade effect of barter. SCASH is the percentage price difference between the barter price and the cash price on the ``sale'' side of the barter deal. PCASH is the percentage difference between the barter price and the cash price on the ``goods payment'' of the barter deal. TOT measures the net terms of trade of barter and is calculated by TOT SCASH ÿ PCASH.9 Table 6.7 gives the terms of trade effect of the non-cash economy. It appears from the table that on the ``sale'' side of the deal the prices charged in barter are in¯ated by up to 50 percent compared to cash deals. This happened in 23.3 percent of the cases while in 73.6 percent of the deals there was no difference between the two prices charged. In 3.1 percent of the cases, the ®rms involved discounted the price on the ``sale'' by up to 17 percent. In order to calculate who bene®ts from demonetization, one has to look also at the pricing behavior on the ``goods payment'' side of the deal. Here it appears that in 25.9 percent of the cases the ®rms discounted the price for the barter good compared to what they typically charge in cash deals by as much as 50 percent. In 62.9 percent of the deals, there was no discounting or in¯ating on the barter prices for the goods payment. In 11.1 percent of the deals, the barter prices were in¯ated by as much as 200 percent. Because of these differences in the pricing behavior between barter and cash deals, the net terms of trade effect of barter appears to be quite substantial ranging between ÿ200 percent and 50 percent. As a result the non-cash economy appears to lead to a substantial shift in the terms of trade compared to the casheconomy. In almost 45 percent of the deals barter shifts the terms of
116
Table 6.8 Pricing behavior of sectors Selling sector
Electricity and gas
scasha
pcashb
Mean
0.00
ÿ4.12
4.12
3.78
7.42
ÿ3.64
Standard deviation
0.00
8.52
8.52
8.80
45.06
42.11
18
18
N Coke and petroleum
Food and beverages
Mean
17
18
pcashb
5.48
1.45
4.03
1.13
ÿ1.31
14.45
10.28
8.37
5.00
6.54
N
13
13
Mean
13
16
16
tot c
2.44 6.57 16
5.00
ÿ1.29
6.29
2.50
0.58
1.92
Standard deviation
10.16
6.05
10.05
8.09
17.02
18.29
N
17
18
18
17
17
18
Mean
2.64
1.00
1.64
3.03
ÿ2.47
5.51
Standard deviation
6.53
38.45
35.75
9.45
15.38
14.09
27
27
36
36
27
36
Mean
1.86
0.26
1.61
5.21
ÿ4.17
Standard deviation
8.46
6.86
9.99
7.11
7.93
N Machinery and vehicles
17
scasha
Standard deviation
N Textiles and leather
17
tot c
16
16
16
12
12
9.38 9.66 12
Mean
3.66
0.91
2.75
3.46
ÿ5.06
8.52
Standard deviation
7.41
10.64
13.51
7.67
9.96
11.40
28
28
N
28
30
30
30
Contract Enforcement in Transition Economies
Metal ores and other nonmetallic minerals
Buying sector
Chemicals
Mean
6.08
ÿ3.60
9.68
7.19
0.07
7.12
Standard deviation
9.49
12.50
11.18
9.47
12.01
8.59
22
22
N Services
15
15
2.83
ÿ4.04
6.86
0.00
ÿ10.00
Standard deviation
6.54
17.72
16.27
0.00
0.00
0.00
23
23
1
1
1
23
10.00
Mean
3.43
ÿ1.16
4.59
3.52
ÿ1.21
4.73
Standard deviation
8.30
18.63
18.07
8.26
19.66
18.86
N Anova
15
Mean N
Total
22
F-test Signi®cance level
163
163
163
146
146
146
1.08
0.33
0.52
0.76
0.77
0.93
(0.382)
(0.937)
(0.817)
(0.619)
(0.614)
(0.489)
Source: Survey of 165 barter deals in Ukraine in 1997. a. Difference between the barter price and the cash price in percentage of the cash price on the ``sale'' side of the barter deal. b. Difference between the barter price and the cash price in percent of the cash price on the ``goods payment'' of the barter deal. c. Terms of trade tot scash ÿ pcash.
Stylized Facts and Competing Explanations
Table 6.8 (continued)
117
118
Contract Enforcement in Transition Economies
trade toward the ``sale'' side of the transaction. In those cases the ``real'' barter price of the ''sale'' is in¯ated by up to 50 percent compared to the cash price for the same goods. As table 6.7 illustrates the terms of trade shifts quite substantially in non-cash transactions. Thus the virtual economy argument has the potential of explaining some of the variation in the terms of trade of barter. If the virtual economy argument is valid, we expect that the manufacturing sector (e.g., textiles, leather, machinery, and vehicles) is overpricing its output in barter compared to cash deals for the same product and pays less than the market value for natural resources (e.g., gas and electricity). Furthermore we expect this pricing distortion to be more pronounced for less ef®cient sectors. In order to test these hypotheses, we have to distinguish whether the sector is on the buying or selling end of the barter transaction. The reason is that overpricing the ``sale'' will bene®t the sector that is on the selling end of the barter transaction and hurt the sector that is on the buying end of the same transaction. Similarly discounting the price for the ``goods payment'' will bene®t the sector that is on the buying end and hurt the sector that is on the selling end. TOT measures the net terms of trade and is calculated by TOT SCASH ÿ PCASH. Thus, when the sector is on the selling end and TOT takes a positive value, barter bene®ts this sector by shifting the terms of trade in its favor. Similarly, when the sector is on the buying end and TOT takes a positive value, then barter hurts this sector by shifting the terms of trade in its disfavor. We are now ready to put the virtual economy argument to an empirical test. Table 6.8 examines whether differences in the pricing behavior across sectors can be identi®ed. The table aggregates the 165 barter deals into 8 sectors and looks at their pricing behavior in non-cash deals compared to cash deals. The table distinguishes whether the sector is on the selling or buying end of the transaction. From table 6.8 it appears that there is no systematic difference in the pricing behavior across sectors in non-cash transactions (the F-test of the analysis of variance [Anova] is not statistically signi®cant at conventional levels). Take the example of machinery and vehicles. When this sector is on the selling end of the transaction, it overprices its output on average by 3.66 percent compared to cash deals, and it is overpriced on the goods payment by 0.91 percent on average. So the sector's net bene®t from barter is 2.75 percent (in terms of its cash
Stylized Facts and Competing Explanations
119
price).10 So far so good. But the same appears to be true for the natural resource sector like electricity and gas. This sector's net bene®t from barter is 4.12 percent (in terms of its cash price). What seems to matter for the pricing behavior in non-cash transactions is not the sector but whether the sector is on the selling or buying end of the transaction. Take again the example of machinery and vehicles. When this sector is on the buying end of the barter deal, it pays more for the ``sale'' by 3.46 percent on average and sells its ``goods payment'' at a 5.06 percent discount compared to cash deals, so the sector's net loss from barter is 8.52 percent on average. This net loss from non-cash transactions appears to occur in all the other sectors as well except for electricity and gas, when the sector is a buyer rather than a seller. It appears then that the sectors gain from barter when they sell and they lose from barter when they buy. The only sector that seems to gain from barter independent from its buying or selling status appears to be the natural resource sector electricity and gas. This is just the opposite from what we would have expected if we believed in the virtual economy argument of Russia's non-cash economy. We now examine whether this result depends on the level of aggregation of sectors. In table 6.9 we aggregate the sectors to a natural resource sector (e.g., electricity and gas, coke and petroleum, metal ores, and other nonmetallic minerals) and to a manufacturing sector (e.g., textiles, leather, machinery, vehicles, and chemicals). We construct a variable that we call virtual economy 1, which includes all deals in which the natural resource sector was on the selling end of the transaction and the manufacturing sector on the buying end. If the virtual economy argument is valid, then we expect to see a discount on the ``sale'' price and a markup over the cash price on the ``goods payment'' leading to a net terms of trade shift in favor of the manufacturing sector. A look at table 6.9 reveals that the opposite is the case. The natural resource sector lives at the expense of the manufacturing sector who suffers a loss in the terms of trade of 6.88 percent on average when the natural resource sector is the seller and manufacturing the buyer in the transaction. Can the manufacturing sector draw on the resources of the natural resource sector when he is selling to this sector rather than buying from it? This case is captured by the variable virtual economy 2, which includes all deals in which manufacturing is the seller and the natural resources sector the buyer in the transaction. The table reveals that in this case both
120
Table 6.9 Is Russia's economy virtual? Percent price differential on Goods payment b
Salea Mean
Standard deviation
Mean
Terms of trade c Standard deviation
Mean
Standard deviation
N
Virtual economy 1 Seller: natural resources d 1.47
4.24
ÿ5.41
9.24
6.88
9.30
17
Other
3.66
8.63
ÿ0.66
19.39
4.32
18.84
146
Total
3.43
8.30
ÿ1.16
18.63
4.59
18.07
163
Anova F-test Signi®cance level
1.06
0.99
0.31
(0.306)
(0.321)
(0.581)
Virtual economy 2 Seller: manufacturing e Buyer: natural resources d
1.43
5.85
0.30
12.91
1.13
14.20
23
Other
3.76
8.61
ÿ1.40
19.44
5.15
18.62
140
Total
3.43
8.30
ÿ1.16
18.63
4.59
18.07
163
Anova F-test Signi®cance level
1.55
0.16
0.98
(0.215)
(0.686)
(0.324)
Contract Enforcement in Transition Economies
Buyer: manufacturing e
Firm's ef®ciency f Low
3.13
7.78
ÿ0.13
11.21
3.27
12.26
57
Medium
2.76
7.07
ÿ4.88
10.60
7.64
12.24
60
High
2.91
10.11
1.16
33.96
1.75
31.29
36
Total
2.93
8.08
ÿ1.69
19.03
4.62
18.58
153
Anova F-test Signi®cance level
0.03
1.45
1.38
(0.969)
(0.239)
(0.255)
Source: Survey of 165 barter deals in Ukraine in 1997. a. Difference between the barter price and the cash price in percentage of the cash price in the ``sale'' side of the barter deal. b. Difference between the barter price and the cash price in percentage of the cash price in the ``goods payment'' of the barter deal. c. Terms of trade tot scash ÿ pcash. d. Includes electricity and gas, coke and petroleum, and metal ores and other nonmetallic minerals. e. Includes textiles and leather, machinery and vehicles, and chemicals. f. Output per employee. Low: 1,000 to 7,000 U.S.$; medium: 7,100 to 15,000 U.S.$; high: 15,100 to 140,000 U.S.$.
Stylized Facts and Competing Explanations
Table 6.9 (continued)
121
122
Contract Enforcement in Transition Economies
sectors are overpricing their output in non-cash transactions compared to cash leading to a slight terms of trade gain for the manufacturing sector of 1.13 percent. However, the constellation of the manufacturing sector as the seller and the natural resource sector as the buyer in barter has been taking place in 23 deals only out of a total of 165 deals. These numbers are much too small to plausibly explain the enormous shift towards non-cash transactions in Russia. Furthermore the F-tests reject the hypothesis that there is any difference in the pricing behavior for both constellations. We turn now to the second prediction of the virtual economy argument. The bottom part of table 6.9 examines whether the price distortions between non-cash and cash deals are more pronounced for less ef®cient ®rms. If the virtual economy argument is valid we expect this to be the case, because ®rms with lower productivity will need to in¯ate their prices by more or get bigger discounts for the barter goods in order to pretend to produce value added. From the table it appears that there is no statistically signi®cant relation between the price distortions and the ef®ciency of the ®rm. If at all, it appears to be the ®rms with productivity levels in the middle range that show the largest price differentials between non-cash and cash transactions.11 To conclude, the virtual economy argument is virtual and has no basis in the data. Who bene®ts from the non-cash economy does not depend on the sector, as the argument claims, but is exclusively driven by the selling or buying status of ®rms. But why would the selling or buying status of ®rms determine in who's favor the terms of trade shifts in barter? Or to put it differently, why are frequently prices for the ``sale'' in¯ated and prices for the ``goods payment'' discounted in barter transactions? We turn to an answer to this question in chapter 8. Taxes Finally, we turn to the tax incentives for barter. Barter is seen by many experts as a possibility to avoid paying taxes. First of all, it allows a distortion of the true value of pro®ts, and thus reduces the tax payment that is due. Second, since the banking sector acts as a tax collection agency that transfers ®rm's incoming cash on bank accounts to the state to pay for outstanding tax arrears, barter allows ®rms to circumvent paying taxes because it avoids payments in cash.
Stylized Facts and Competing Explanations
123
Table 6.10 Motives for barter 1997 responses (in %) No cash
87.5
No bank loan
29.1
No trust in the value of money Faster payment compared to cash payment No struggle with other creditors
6.0 72.1 7.8
No courts to enforce rights To maintain production
6.0 12.5
Goods in stock could be used
66.1
Liquid good Better deal on the price in barter Reducing the tax burden
1.8 20.8 9.5
Avoiding the controls on foreign trade
1.8
Reducing regulation
6.8
Capital ¯ight to the West State pressure
0.0 1.8
Others
1.2
Source: Survey of 165 barter deals in Ukraine in 1997. Notes: Answers have been ranked from very important to irrelevant. The percentages give the ``very important'' responses.
We have asked the ®rms whether there was a tax advantage reason for using barter. Only for 9.5 percent of the barter deals ®rms answered that taxes were a very important or important reason to engage in this form of exchange (see table 6.10). Even if one takes into account that the data have some noise, this low number suggests that tax reasons are not the major motivation behind barter. 12 6.5
New Interpretation of the Evidence
The empirical evidence so far indicates that barter is a dominant phenomenon of the former Soviet Union.13 Furthermore the tables suggest that the most common explanations of barterÐthe lack of market discipline, the lack of restructuring, the virtual economy, and tax avoidanceÐare not supported by the data. An explanation of barter has therefore to be found somewhere else. More speci®cally, any explanation of barter has to address the following two questions: First, why would parties want to tie two deals? Second, why would parties want to pay in goods rather than money? We turn to an answer for this question in the next chapters.
124
Contract Enforcement in Transition Economies
Before we come to a speci®c answer to these questions we turn to the answers given by the ®rms themselves. Table 6.10 indicates that barter is predominantly motivated by ®nancial considerations. In 87.5 percent of the deals a key reason for using barter is the lack of cash. In 29 percent of the barter deals the ®rm could not get a loan even if it were ready to pay a high interest rate. In 72 percent of the cases the party uses barter, because she expects to be paid faster in this form of exchange. Another important reason for barter seems to be to smooth production. In 66 percent of the cases the ®rm can use goods stored as inventories as means of payment in barter deals, and in 12.5 percent of the cases the ®rm uses barter, because it is the only way to maintain production. Additionally barter is used as a way to change the relative price for the good in question in 20.8 percent of the deals. In the following three chapters we offer a theoretical explanation for and empirical evidence on how the output decline, inter-®rm arrears, and barter in transition economies are connected. We suggest an alternative explanation of barter that takes into account that barter is primarily driven by ®nancial considerations and the nonavailability of bank loans as can be seen from table 6.10. Furthermore our theory of barter will be able to explain why ®rms can get better deals in barter as compared to cash transactions. In chapter 7 we explore the relationship between arrears and barter. More speci®cally, we explain why ®rms are able to give loans to other ®rms in form of arrears when the banking sector is reluctant to provide capital. We will argue that barter becomes important in this context. In a barter trade one ®rm gives a trade credit to another ®rm that is repaid in goods rather than money. The reason ®rms are willing to do so is that barter creates a deal-speci®c collateral that allows them to deal with the problem of creditworthiness of ®rms. Our analysis points out the potential importance of the institution of barter in the context of transition. Barter can be used to collateralize a trade credit when ®rms' creditworthiness problem is severe, and it allows ®nancing of business activities that otherwise would not take place. Through this credit channel barter helps to smooth the output decline and thus prevents output from falling even further. In chapter 8 we focus on explaining the output decline in transition economies and study the impact of arrears and barter on ®rms' growth. For this purpose we combine the input shortage explanation of the output decline of Blanchard and Kremer with the ®nancial
Stylized Facts and Competing Explanations
125
shortage explanation of the output decline of Calvo and Coricelli. Surprisingly we ®nd that when both problems are presentÐthe input shortage as well as the ®nancial shortageÐthe picture changes drastically. In particular, we ®nd that one problem can help with the other, rather than making things worse. More speci®cally, we show how the credit constraint can alleviate the holdup problem. The fact that the input seller has to make sure to get paid when the input purchaser is short of cash to pay for these inputs gives the input purchaser bargaining power. This bargaining power in turn reduces the possibility that the input supplier can exploit the input purchaser's need for the input. In chapter 9 we test the predictions of the model with ®rm and deal speci®c data of 165 barter deals among ®rms in Ukraine in 1997. We ®nd that the input shortage explanation of the output decline that Blanchard and Kremer established for industry data holds also for ®rm level data. Furthermore we ®nd that in addition to the input shortage, the ®nancial shortage and barter each have an important effect on output growth. Finally, we establish with deal-speci®c data how disorganization and the ®nancial shortage affect the terms of trade in barter deals.
This page intentionally left blank
7
Arrears and Barter: Extending Trade Credits
In the preceding chapter we raised the question why ®rms are willing to give loans to other ®rms in form of arrears when the banking sector is reluctant to provide capital. In this chapter we take up this question and show that ®rms can do what banks cannot do because they can engage in barter trade.1 In a barter trade one ®rm gives a trade credit to another ®rm which is repaid in goods rather than money. In part I we have argued that barter trade is a possibility to collateralize trade credits. We will see that this is also the main function of barter in transition economies. Barter offers a number of advantages. First, barter does not attempt to improve the overall creditworthiness of ®rms (as in bankruptcy) but rather restores the creditworthiness of the ®rm for one speci®c deal. In a barter trade a deal-speci®c collateral is created in the form of the future goods payment. Depending on the degree of the credit problem of the debtor, the creditor can choose the value of the collateral relative to the trade credit that he gives to the debtor. This way the debtor's creditworthiness is restored for one speci®c deal. Giving a trade credit in the form of a barter deal is available to ®rms only, since banks are not allowed to engage in the trading business. The option of improving a debtor's creditworthiness by doing a barter deal is therefore not available to banks, which explains why ®rms are able to give loans when banks are reluctant to do so. Second, in the early stage of transition barter trade can compensate for creditor's passivity with respect to using a bankruptcy procedure to pursue defaulting debtors. Instead of relying on the low and unknown liquidation value of the ®rm (as in bankruptcy), the creditor and debtor create a deal-speci®c collateral of positive and known value. Furthermore in a barter trade the creditor does not need to
128
Contract Enforcement in Transition Economies
Table 7.1 Barter as credit (in %) Ex ante
Actual
Prepurchase
14.9
No termination point
20.2
0 month
46.4
46.4
1 month
9.5
16.1
1±3 months
3.6
12.5
3±7 months Missings
3.6 1.8
8.3 2.4
100.0
100.0
Total
14.3 Ð
Source: Survey of 165 barter deals in Ukraine in 1997.
share the bene®ts from her legal actions against a defaulting debtor with other creditors. In a barter deal the creditor obtains property rights on goods, and this effectively means that she does not need to queue with other creditors for the money. So, in contrast to a bankruptcy procedure in a barter trade, there is no priority ordering of creditors. This makes payment in goods a superior credit enforcement mechanism compared to payment in money. 2 In addition the known and positive value of the deal speci®c collateral in barter with no priority ordering of creditors gives the creditor a greater incentive to pursue legally a nonpaying debtor as compared to a bankruptcy procedure. This way barter trade can be a remedy against creditor's passivity in the early stage of transition. Third, barter is a more information-intensive form of ®nancing. Typically the ®rms who extend trade credits know each other from previous transactions (one ®rm is a producer, and the other ®rm is an input supplier). This way the problem of credit evaluation, which the bank has, might not arise (see also table 9.2, which illustrates this point). Table 7.1 looks at whether within barter a trade credit is in fact given. In 36.9 percent of the deals, there was a trade credit. The time period between the sale and the goods payment varied between 1 month and 7 month. In 20 percent of the deals the parties did not make an agreement on the termination of the credit. Table 7.2 shows that when a trade credit was given, the parties agreed on it ex ante in 16 percent of the barter deals only. In 17 percent of the cases a trade credit was given ex post by the selling ®rm because the buyer was unable or unwilling to pay.
Arrears and Barter: Extending Trade Credits
129
Table 7.2 Trade credit Percentage Ex ante No
81.5
Yes
16.1
Not applicable Missings Total
0.6 1.8 100.0
Ex post Buyer was unable or unwilling to pay
16.7
Seller wanted to be paid later Not applicable
3.6 76.2
Other Missings Total
1.8 1.8 100.0
Source: Survey of 165 barter deals in Ukraine in 1997.
Table 7.3 reports on the outstanding debt of bartering ®rms and examines whether there is a relationship between the size of the ®rm's outstanding debt and the extent to which the ®rm engages in barter. Firms who barter tend to have large outstanding bank debt, ®rm debt and outstanding tax arrears (exceeding 100 percent of ®rm sales in 1996). This suggests that these ®rms had little creditworthiness to obtain further credit. If our explanation is correct, we expect a positive association between the barter share of the ®rm and its outstanding debt and a negative association between the barter share and bank debt. Barter can help ®rms with weak overall creditworthiness when they cannot get a bank loan by restoring their creditworthiness for one particular deal. The table shows that the barter share of the ®rms indeed tends to increase with outstanding ®rm and wage arrears. At the same time barter tends to be lower for those ®rms which have access to bank loans. Furthermore a simple correlation between the ®rm's bank debt with its ®rm arrears reveals a weak negative correlation between the two (the correlation coef®cient is ÿ0:185) once the state ®rms are excluded suggesting that ®rm debt helped to compensate the liquidity squeeze induced by low bank debt for those ®rms in the economy with restricted access to bank loans.
130
Contract Enforcement in Transition Economies
Table 7.3 Barter and creditworthiness Barter share in percentage of output Debt in percentage of output
Mean
Standard deviation
Cases 138
Total 0±10%
43.67
27.7
22.04
17.5
42
10±20%
56.28
23.5
42
20±690%
51.70
27.1
66
debta
F 25:374 Signi®cance level 0.000 Bank debt
45.12
28.5
150
0% 1±5%
48.63 37.57
32.0 25.9
63 42
5±105%
47.27
24.6
45
138
F 2:114 Signi®cance level 0.124 Firm arrears
43.67
27.7
0±10%
32.13
25.4
60
10±20%
48.44
21.5
33
20±626%
55.57
28.9
45
150
F 11:350 Signi®cance level 0.000 Wage arrears
45.13
28.5
0%
37.63
29.8
75
1±10% 10±40%
49.25 77.33
24.2 17.8
66 9
150
F 10:151 Signi®cance level 0.000 Tax arrears
45.13
28.5
0%
39.84
21.1
87
1±10%
52.15
27.6
39
10±125%
52.88
28.5
24
F 3:704 Signi®cance level 0.027 Source: Survey of 165 barter deals in Ukraine in 1997. a. Except bank debt.
Arrears and Barter: Extending Trade Credits
131
In order to examine the role of barter to alleviate the liquidity squeeze of ®rms, we explore next whether ®rms in our sample have in fact faced a liquidity shortage. Blanchard and Kremer (1997) report evidence based on a survey among 500 ®rms in Russia that suggests that the ®nancial constraint was the most important constraint experienced by enterprises (see their table IV). Between 1993 and 1995 over 60 percent of the ®rms experienced a shortage of ®nancial resources compared with only over 20 percent of the ®rms experiencing shortages of materials. As we have reported above, Calvo and Coricelli (1995a, b) have argued that credit contraction and the associated liquidity shortage have caused the output decline in Eastern Europe.3 We now turn to assess whether inter-enterprise arrears can be seen as a response to the liquidity crunch in the economy.4 If this is the case, we expect inter-enterprise credit to be negatively associated with bank credit. Firms that cannot get bank credit turn to other ®rms for trade credit. To examine the relationship between these two types of credit in more detail, we regress the share of inter-®rm arrears in percent of the ®rm's output on the share of the ®rm's bank debt and a set of other variables that we consider play a role for the size of arrears. The results are reported in table 7.4. In columns 1 to 3 the results for the total sample are given. The coef®cient on bank debt is positive and highly signi®cant, suggesting that ®rms with access to bank credit were also successful in getting inter-enterprise credit. One of the reasons why the two types of credit move in the same direction is the ownership status of the ®rm. Rostowski (1993) and others have argued that arrears are simply a manifestation of soft budget constraints. To control for this possibility we divide the sample into private and state ®rms and rerun the regressions. It turns out that for private ®rms inter-®rm credit cushioned the liquidity contraction induced by lower bank credit (see columns 4 and 5 of table 7.4). State-®rms in contrast appear to be able to use their privileged status of creditworthiness to get cash credits from banks as well as trade credits from other ®rms (columns 6 and 7). We turn now to the relationship between arrears and barter. Barter creates a deal-speci®c collateral that can be used to collateralize a trade credit when credit enforcement becomes prohibitively costly. As we model formally in the next chapter, this collateral function of barter explains why ®rms are able to give loans to each other when
132
Table 7.4 Barter, arrears, and liquidity squeeze Firm arrears
Total arrears
All ®rms (1)
(2)
(3)
Private ®rms
State ®rms
All ®rms
(4)
(5)
(6)
(8)
(7)
(9)
(10)
Private ®rms
State ®rms
(11)
(12)
(13)
(14)
2.61 (0.00)
2.40 (0.00)
2.61 (0.00)
ÿ1.77 (0.13)
ÿ1.77 (0.28)
2.27 (0.01)
3.64 (0.00)
3.72 (0.00)
3.54 (0.00)
3.72 (0.00)
ÿ4.17 (0.17)
ÿ9.02 (0.02)
3.42 (0.00)
4.80 (0.00)
Barter
0.92 (0.00)
0.90 (0.00)
0.92 (0.00)
0.49 (0.00)
0.48 (0.00)
2.57 (0.00)
2.22 (0.00)
1.07 (0.00)
1.06 (0.00)
1.07 (0.00)
0.40 (0.03)
0.67 (0.00)
2.67 (0.00)
2.32 (0.00)
117.59 (0.05)
120.36 (0.04)
117.64 (0.05)
748.54 (0.00)
135.87 (0.02)
138.20 (0.00)
135.85 (0.02)
Complexity State
23.57 (0.15)
Private R 2 adjusted N
7.70 (0.85)
0.22
0.23 149
755.12 (0.00)
19.78 (0.23) 0.53 (0.98)
149
195.70 (0.04)
0.22 149
ÿ0.19 (0.99) 0.87 13
0.83 12
0.24 49
0.40 49
0.34 151
0.35 151
0.34 151
0.47 14
0.61 13
0.35 49
0.48 49
Source: Survey of 165 barter deals in Ukraine in 1997. Notes: Numbers in parentheses are p-values. Bank debt Firm's bank debt in percent of output Barter Share of ®rm's barter in percent of output Complexity Complexity index of the industrial sector of the ®rm; the index is equal to zero if there is only one input and tends to one if the sector uses many inputs State Dummy variable equal to one if ®rm is state owned Private Dummy variable equal to one if ®rm is private ®rm
Contract Enforcement in Transition Economies
Bank debt
Arrears and Barter: Extending Trade Credits
133
Table 7.5 Solving creditworthiness in cash deals 1997 responses (in %) Advanced payment
92.9
Close relationship
59.5
Valuable collateral available
12.5
Threat of bankruptcy
64.3
Protection ®rms as collectors
12.5
Getting rid of inventories Maintaining production
19.0 5.4
Administrative enforcement
12.5
Refusal of service
1.8
Trust in contract
0.6
Others
1.8
Missings
1.8
Source: Survey of 165 barter deals in Ukraine in 1997. Notes: Because of multiple responses, the ®gures do not add up to 100.
the banking sector is reluctant to provide capital. Barter can help ®rms with weak overall creditworthiness by restoring the creditworthiness for a speci®c deal. When arrears reach a critical level at which credit enforcement becomes prohibitively costly, the only way ®rms will still extend trade credit to other ®rms is when they do so in the form of barter trade. Thus, at some critical level of arrears, the only way ®rm's arrears can further grow is when they undertake barter. If our explanation of barter is correct, we expect a positive coef®cient on the barter share. This is indeed the case for the overall sample and the two subsamples of ®rms (columns 2 and 3, 4 and 5, and 6 and 7 of table 7.4). Finally, we rerun the regressions with total arrears (e.g., wage and tax arrears) rather than ®rms' arrears alone as the dependent variable with very similar results. The data seem to suggest then that the Ukrainian ®rms in fact experienced a liquidity squeeze that barter has helped to alleviate. Table 7.5 illustrates how ®rms have dealt with the lack of creditworthiness of their business partners in cash deals. 93 percent of the ®rms answered that they asked their business partners to pay in advance before the delivery of goods. 60 percent of the ®rms extended trade credits only when they had a close relationship to their business partner. Surprisingly, 64 percent of the ®rms answered that the threat of bankruptcy helped to get paid in cash deals.5
134
Table 7.6 Inverse U-curve: Output growth, arrears, and barter Mean
Standard deviation
<25%
ÿ0.48
1.9
81
>25%
1.05
6.4
57
ÿ1.12
0.7
63
Mean
Casesa
Standard deviation
Cases a
Firms with Total arrears Bank debt
0
Tax arrears
<5%
>0
0.87
5.4
87
ÿ0.17
2.5
111
<50%
ÿ0.32
2.4
>50%
3.24
10.2
18
Bank debt
<10%
ÿ0.42
2.3
135
>10%
4.08
11.1
15
Tax arrears
<20%
ÿ0.40
2.3
135
Total arrears
0.61
7.3
39
0 >0%
ÿ0.05 0.11
2.5 5.5
75 75
Wage arrears
<9% >9%
Firm arrears
<10%
ÿ0.58
1.7
58
Firm arrears
>10%
0.67
5.6
80
<30%
ÿ0.79
2.0
98
1.44
6.3
Barter share
>30% Total sample
0.01
55
3.88
11.2
15
0.13 ÿ1.49
4.4 0.6
141 9
<50%
0.26
4.6
126
>50%
ÿ1.00
0.3
12
<70%
0.27
4.6
128
>70%
ÿ1.34
0.4
25
>20%
Barter share
153
Source: Survey of 165 barter deals in Ukraine in 1997. Notes: Output growth is the percentage deviation of ®rm's output growth relative to growth rate of GDP between 1994 and 1996. a. The number of cases may exceed the number of interviewed ®rms because each barter deal involves a selling as well as a buying ®rm.
Contract Enforcement in Transition Economies
>5% Wage arrears
120
Arrears and Barter: Extending Trade Credits
135
Some of the ®rms had valuable collateral to offer to guarantee payment (12.5 percent of ®rms). 12.5 percent of the ®rms used protection ®rms or the Ma®a as cash collectors. If the interpretation of barter as a means to soften the liquidity squeeze is correct, we expect a positive relationship between barter and output growth. Barter allows to furnish ®rms with the required liquidity to produce. Otherwise, production would be hindered by the lack of credit. In table 7.6 we take a ®rst look at whether this positive relationship can be found in the data. The table reports the relationships among the output growth of the ®rm, the liquidity squeeze, and barter. We ask the question whether ®rms with large ®rm arrears, total arrears, and a big exposure to barter did relatively better in terms of output growth compared to the economy as a whole. We take arrears as evidence that the ®rm faced a liquidity constraint and therefore turned to other ®rms for credit. We measure the relative growth performance of the ®rm by the mean percentage deviation of the output growth of the ®rm between 1994 and 1996 relative to GDP growth in Ukraine in the same period. The table shows that the ®rms of our sample experienced the same growth rate as the GDP of the Ukraine economy. However, ®rms with total arrears of more than 25 percent of output did substantially better in terms of output than ®rms with total arrears of less than 25 percent. When total arrears are decomposed into tax, wage, and ®rm arrears, a slightly different picture emerges for wage and ®rm arrears. When wage and ®rm arrears become very large (over 9 percent and over 50 percent of output, respectively), then the ®rm's output performance becomes worse than that of the economy as a whole. A similar picture emerges for the ®rm's barter exposure. Firms with a barter share of output over 70 percent did less well, and those with a barter share of over 30 percent performed better compared to the economy as a whole. Thus the relationship between barter and output growth is not monotone, as expected, but inversely U-shaped. This suggests that the trade credit story given in this chapter is incomplete. It cannot account for the fact that when the barter exposure of the ®rm becomes very large, barter stops to further production. We address this puzzle in the next chapter.
This page intentionally left blank
8
The Output Decline: The Lack of Trust and Liquidity
In chapter 7 we reported on a number of explanations that have been put forward for the unexpected output decline in the former Soviet Union. In this chapter we develop a theory that establishes a link between the output decline, inter-®rm arrears, and barter.1 In chapter 9 we will confront this theory with the data and show that the theory is able to explain the time and cross-country pattern of the output decline, arrears, and barter in transition economies. 8.1
Buyer-Supplier Relationships
Blanchard and Kremer (1997) explain the rapid output decline in the former Soviet Union by disorganization and holdup problems. Central planning was characterized by a complex set of speci®c relations between ®rms. Many ®rms had only one supplier from which to buy and knew of only one or a few buyers to whom to sell. This picture of little outside opportunities is still observed in Ukraine in 1997, as table 8.1 shows. In 20 percent of the sales within barter deals, the parties had no alternative partner, and in 37 percent of the sales only a few alternative partners to carry out the business. The Ukraine economy in 1997 was also characterized by a high degree of personalized exchange. Table 8.2 illustrates how close the interaction between the trade partners were by looking at the history and the quality of the relationship between business partners. This large degree of personal connections and close business ties is another indicator for an environment with little outside opportunities. Almost 50 percent had frequent business ties before. Over 50 percent of the ®rms had already previously been input, raw material, or energy suppliers. Over 10 percent have been distributors. Only in 32
138
Contract Enforcement in Transition Economies
Table 8.1 Lock-in: Business alternatives (in %) Sale
Goods payment
No alternative
20.2
6.0
A few alternatives
36.9
32.1
Many alternatives
41.1
60.1
Source: Survey of 165 barter deals in Ukraine in 1997. Table 8.2 Close business ties 1997 responses (in %) History of interaction First time
23.2
Occasionally
28.6
Frequently
46.4
Missings Total
1.8 100.0
Quality of interaction Supplier of inputs
28.6
Supplier of raw materials
11.9
Energy provider
11.9
Product distributor Input supplier and distributor Partner in other business activities No relationship Missings Total
3.0 7.7 3.0 31.5 2.4 100.0
Source: Survey of 165 barter deals in Ukraine in 1997.
percent, the two trade partners had no special business relationship before. An environment with little outside opportunities typically creates holdup problems and opens room for bargaining. Under central planning the main instrument to enforce production and delivery of goods was the coercive power of the state. Transition eliminated the central planner and thus the instrument to limit the adverse effects of speci®city, without having created yet the institutions to deal with speci®city such as vertical integration and contracts that exist in the West.2 Furthermore, in times of transition, the anticipation of
The Output Decline: The Lack of Trust and Liquidity
139
changing business partners and the disappearance of ®rms shortens horizons and reduces the scope for long term relationships. Thus in such a ``no future'' environment a typical mechanism to constrain opportunistic behavior such as reputation does not work. Blanchard and Kremer argue that speci®city in the relations between ®rms together with incompleteness of contracts results in disorganizationÐthe breakdown of many economic relations before new ones can be establishedÐwhich in turn explains the large output losses. In the Blanchard and Kremer model speci®city arises in a chain of production in which the primary input supplier stands in the beginning of a chain of production and after n steps the ®nal good is produced. Each buyer along the chain knows only the supplier it was paired with under central planning. The primary input supplier has an alternative use for the input, while all intermediate producers along the chain of production are assumed to be able to sell to the following buyer only. Blanchard and Kremer formulate the holdup problem by assuming that it is impossible for each ®rm in the chain to sign a contract with the buyer (the next ®rm in the chain) before it has produced the good. Each ®rm must ®rst buy inputs and produce, and only thenÐonce the cost of producing is sunkÐcan strike a bargain with the next producer in the chain. At this stage, however, each intermediate producer's reservation value is zero, and thus the next producer in the chain can ``hold him up'' and exploit his dependency by offering to purchase the good only at a price that does not cover each intermediate producer's costs. In anticipation of being ``held up'' by the following ®rm each intermediate producer will stop to deliver inputs to the next producer and thus the chain of production breaks down. Output collapses because of a shortage of inputs. In the Blanchard and Kremer model output collapses because ®rm relations are speci®c (the intermediate producers cannot sell the good to someone else) and because contracts are incomplete (each intermediate producer must produce its intermediate good before bargaining over the price for the input with the next producer along the chain). If the government retained its coercive power, it could force suppliers to deliver, and thus output would not decline. In this chapter we offer a mechanism other than the coercive power of the state by which intermediate producers can be induced to trade inputs and producers next in the chain can be prevented from reneging and renegotiating the price. For this purpose we will develop a model that introduces liquidity and credit constraints into
140
Contract Enforcement in Transition Economies
a Blanchard and Kremer type of production chain. Analogous to Blanchard and Kremer we consider a good that requires n steps of production. In contrast to Blanchard and Kremer, however, we consider a situation where the supplier holds up the buyer rather than the other way around. In our formulation the buyer makes a ®rmspeci®c investment in order to ®nd an adequate input supplier. This investment takes place before the delivery of the input. When the supplier delivers the input, they negotiate over the input price. At this stage the buyer's investment is already sunk and not taken into account in the bargaining over the input price. This is what constitutes the holdup problem on the buyer's side. Thus the buyer might not invest in ®nding a supplier relationship. This formulation of the holdup problem on the buyer's rather than the supplier's side seems to be more plausible in the context of the former Soviet Union, since input suppliers are on the short side of the market and thus input buyers have to spend time and money in order to ®nd adequate suppliers and to establish a business relationship.3 We assume further that the intermediate producer has no cash to pay for the input at the date of the delivery of the input. He requires a trade credit from the input supplier that he can repay when he sells the input to the next ®rm. In order to make sure that his trade credit is repaid, the input supplier has to incur credit enforcement costs (he has to involve legal ®rms or the Ma®a). We show that the fact that the intermediate producer has no cash to pay for the input gives the input purchaser some countervailing bargaining power to deal with the holdup problem just described. This bargaining power in turn reduces the possibility that the input supplier can hold up the buyer. More speci®cally, since the buyer has no cash to pay for the input upon delivery, once the input supplier delivers the input, the bargaining power reverses and shifts to the buyer. Now the input supplier has to worry of being paid. When it comes to paying for the input after the input buyer has realized his pro®ts from selling the input to the next ®rm (which the input seller is assumed to be observing), the input buyer will hold up now the input seller and renegotiate the price for the input. He will try to lower the input price by the enforcement costs that the seller has to incur if the buyer does not pay voluntarily. In equilibrium the input seller will accept this lower price, since his alternative is to insist on the original price and to involve the Ma®a.
The Output Decline: The Lack of Trust and Liquidity
141
However, the input supplier anticipates this future price renegotiation by the buyer and tries to in¯ate the input price in the ®rst place when he delivers the input to cover the enforcement costs. The problem is that marking up the input price in anticipation of the future price renegotiations is possible only at low credit enforcement costs. When these costs are suf®ciently large, the buyer's liquidity constraint makes it impossible for the supplier to pass on these costs to him. The reason is that the most the buyer can pay for the input is the cash he himself realizes from selling the intermediate good to the next buyer. If enforcement costs are suf®ciently large, the input buyer's cash from the sale to the next ®rm will not be enough to cover these costs. This is the circumstance when the input buyer can exploit the fact that he is liquidity constrained to shift the surplus in his favor and thus prevent being held up by the input supplier. When credit enforcement costs are too large, however, it stops to work. The input supplier refuses to participate in the deal, since at large credit enforcement cost he cannot expect a positive pro®t. Our result that the lack of cash can alleviate the holdup problem stands in contrast to Blanchard and Kremer's model in which, if at all, the role of cash is seen as a positive commitment device that helps solve the coordination problem of ®rms. If the intermediate producer's liquidity constraint is alleviating the holdup problem, what is then the role of barter? Barter becomes important when credit enforcement becomes so costly that input suppliers refuse to participate in the deal. Thus, if the input buyer has no cash, and requires a trade credit from the input supplier, and the legal system to enforce payment is poorly developed, a potentially valuable transaction does not take place. Under these circumstances barter can help to maintain production. Barter introduces a hostage, a commitment devise that prevents the buyer from fully exploiting his bargaining power.4 More speci®cally, when enforcement costs are prohibitively large for the input supplier to participate in the deal, introducing a second pro®table deal in the form of the goods payment allows the input buyer to commit not to exploit his bargaining power and to shift some of the pro®t back to the input supplier to make him participate in the deal. Barter is a self-enforcing arrangement that makes the intermediate producers along the chain of production to lose from reneging the contract. This way barter helps to cope with speci®city when legal institutions are not fully reliable.5
142
Contract Enforcement in Transition Economies
Figure 8.1 Time sequence at production step 1
8.2
The Problem of Contract Enforcement
In this section we develop a chain of production model that captures the problem of disorganization and credit constraint. Similar to Blanchard and Kremer (1997), we consider a good that requires n steps of production. Each production step is carried out by a different ®rm. One unit of the input good gives, after n steps of production, one unit of the ®nal good. Each buyer along the chain can negotiate only with his supplier. This leads to n bargaining problems along the chain. We assume Nash bargaining at each production step with both parties equally sharing the joint surplus, whenever possible. The value of the ®nal good is denoted by v > 0. Intermediate goods j produced at production step j 1; . . . ; n ÿ 1 have a value of vj a v if sold as input good for the next production step, but they have a value of zero if sold to someone outside the production chain. Within the production chain, the value of intermediate good j, vj is determined by the payment its producer Bj receives when selling it to the next producer along the chain of production, Bj1 . A Holdup Problem with Liquidity Constraints We now look at the production steps in more detail. Figure 8.1 illustrates the time sequence at production step 1. Consider the parties involved in this ®rst step of production, the supplier of the original input good, S1 , and the buyer, B1 . Blanchard and Kremer formulate the holdup problem by assuming that S1 cannot sign a contract with B1 before he has produced the input. S1 must ®rst produce and only thenÐwhen the cost of producing is sunkÐcan bargain over the input price. In contrast to Blanchard and Kremer, we consider a situation where the supplier holds up the buyer rather than the other way around. We assume that B1 needs to make a relationship-speci®c investment i at date 0.9. This investment could be thought of as the
The Output Decline: The Lack of Trust and Liquidity
143
time and money B1 spends in order to ®nd an adequate supplier and establish a business relationship. As we have argued above, this formulation of the holdup problem on the buyer's rather than the supplier's side seems to us to be more plausible in the context of the former Soviet Union, since input suppliers are on the short side of the market. Thus it is the input buyer who has to spend time in order to ®nd an adequate supplier. The problem is that at the time of this investment, the two ®rms cannot write a contract that commits S1 to deliver the input good for a particular price in the future. This leads to a holdup problem in the bargaining of the price when the input good is actually delivered. At date 1, the two parties can negotiate about the delivery of S1 's input good and about the price. To save on notation, we normalize S1 's opportunity cost of delivering the input to zero, and we assume that S1 delivers the input only if he expects a strictly positive surplus from the transaction. As speci®ed above, v1 denotes the value of the input good to B1 . This value is determined by the future bargainings and solved recursively below. In contrast to Blanchard and Kremer, we assume that B1 cannot pay cash at the time of delivery because he is liquidity constrained. This assumption re¯ects a common problem in transition economies, as we have seen in the previous two chapters. Thus S1 has to deliver the input good on a credit basis, if at all. B1 will be able to pay when he is paid v1 by the next buyer in the second production step. But, of course, enforcing credit repayment in transition economies is notoriously dif®cult. We capture this notion by assuming that S1 has to incur some (arbitrarily high) cost x
p to enforce repayment of p. This cost could be thought of as the cost of using the legal system, including lawyer fees and potentially bribes for judges or public authorities, or the cost of private enforcement, including the use of Ma®a and the like. These costs are higher, the less developed is the legal system and the more indebted is B1 . We assume for simplicity that the ®xed enforcement cost x
p1 1 x.6 If S1 has decided to deliver the input good at date 1, B1 can try to exploit the fact that credit enforcement is costly and default on some of his payment when he has been paid by buyer B2 and thus has the resources to pay. This will be the case when the production of the ®nal good is complete and all sellers along the production chain have been paid by their buyers, at date n:1. Let p~1 denote the price paid by B1 at this date.
144
Contract Enforcement in Transition Economies
This ®rst production step is repeated at steps 2 to n, with good 1 being used as an input good sold by B1 (now called S2 ) to B2 , and so on. Note that buyers B2 up to Bn may have to undertake a similar relationship-speci®c investment i j , j 2; . . . ; n, and may be similarly credit constrained as B1 . For notational convenience we restrict attention to investment levels i j , j 2; . . . ; n, such that i j a i for all subsequent buyers. Similarly the enforcement costs x are the same in all production steps (see note 5 above). When production is ®nished after n steps and the value of the ®nal good is realized, Bn can use the revenues from selling this good to pay p~nÿ1 , the price he actually pays after delivery of the intermediate good n ÿ 1. Similarly, when Sn is paid, he can use his revenues for paying Snÿ1 , and so on. We assume that the maximum payment that can be enforced at each production step j is equal to the revenues vj generated from selling the good to the next production step. 7 Let us now solve production step 1 recursively, taking as given the value of the good to be produced at this step, v1 . Recall that at date 1, when S1 delivers the input good, B1 has no cash to pay for the input. At date n:1, when realizing his pro®ts from selling the input to the next buyer, B1 has enough cash to pay. However, if he does not do so voluntarily, S1 's has to incur costs x in order to enforce payment of p1 . Suppose that B1 refuses to pay the full price p1 on which the two parties agreed at date 1 but offers to pay p~1 p1 ÿ x instead. If this happens, S1 can either accept this payment or enforce p1 at cost x. In equilibrium he will accept B1 's reduced payment. At date 1, the two parties have to agree on a price p1 . Since B1 's investment i is already sunk at this date, this investment is not taken into account in the bargaining. This is what constitutes the holdup problem of buyer B1 . However, the two parties anticipate at date 1 that B1 will exploit his position after delivery of the input good and pay a reduced price at date n.1. Recall that we have assumed Nash bargaining whenever possible. This implies that a price p1 is chosen such that v1 ÿ
p1 ÿ x p1 ÿ x $ p1
v1 x; 2
8:1
meaning that in anticipating B1 's future price reduction, S1 marks up p1 in the ®rst place, if this is possible. However, in¯ating the input price in anticipation of the price reduction at date n:1 will not always be possible. B1 's liquidity
The Output Decline: The Lack of Trust and Liquidity
145
constraintÐthe cash he gets when he himself sells the good to the next buyer, v1 Ðputs a bound on the maximum payment that can be enforced at cost x. Thus, in order to fully capture the subsequent price reduction, S1 may want to in¯ate the price more than can credibly be enforced as payment at date n:1, since, even at cost x, B1 cannot be forced to pay more than he has in his pockets at date n:1. Thus v1 x; v1 : p1 min
8:2 2 Only if enforcement costs are low (i.e., x < v1 =2) will S1 be able to pass on x in the price markup. In this case the fact that B1 is liquidity constrained does not prevent S1 's and B1 's equally sharing the surplus, v1 . If x is suf®ciently large (i.e., x > v1 =2), B1 's liquidity constraint will make it impossible for S1 to pass on these costs to him. B1 's cash from the sale to the next buyer will simply not be enough to fully cover these costs. In this case B1 can exploit the fact that he is liquidity constrained to capture more than half of the surplus. If the enforcement cost excede the total value of the transaction (i.e., x b v1 ), then B1 captures the entire surplus and S1 cannot guarantee himself a positive payoff. The following payoff functions summarize these three cases. If S1 delivers the input good, then for a given v1 , the payoff of B1 is 8 v1 v1 > > if x a , > ÿi > > 2 2 < 1 PB v1 v1
8:3 > a x a v1 , x ÿ i b ÿ i if > > 2 2 > > : v1 ÿ i if x > v1 . Similarly the payoff 8 v1 > > > > > <2 PS1 v1 > v1 ÿ x < > > 2 > > : 0
of S1 is given by if x <
v1 , 2
v1 a x a v1 , 2 if x > v1 .
if
8:4
Thus B1 's liquidity constraint gives him some bargaining advantage because credit enforcement is not costless to S1 and the maximum payment that can be enforced is ®nite. If credit enforcement is a suf®ciently severe problem, B1 can use his bargaining power to shift
146
Contract Enforcement in Transition Economies
the surplus in his favor. Otherwise, the bargaining is either not affected by the presence of enforcement costs (when x is low) or S1 refuses to participate in the deal (when x is very large). So far we have taken the value of the ®rst production step, v1 , as given. We still have to determine how v1 is affected by the value of the ®nal product, v, by the number of production steps, n, and by the fact that all buyers are liquidity constrained and that credit enforcement is costly. For this purpose we have to solve the game recursively. The following lemma characterizes v1 as a function of v, n, and x. lemma 8.1 v1
x; n; v v1
x; n; v < Proof
The value of production at step 1 is v 2 nÿ1 v 2 nÿ1
if x < if x b
v 2 nÿ1 v 2 nÿ1
;
8:5 :
See appendix II-A.
The important thing to note here is that if x is small enough, it does not affect the value of production at step 1. The reason is that in all subsequent production steps, x can be fully covered by a price markup and hence does not affect the equal sharing of the surplus. Bargaining Power and Output Growth We can now state the conditions under which production takes place at the ®rst and all subsequent production steps. S1 agrees to deliver the input good at date 1 on a credit basis if and only if PS1 > 0. This is the case if and only if v > x; 2 nÿ1
8:6
because in this case v1 v=2 nÿ1 by lemma 1 and x < v1 , so PS1 v1 ÿ x > 0 (see equation 8.4). If x b v=2 nÿ1 instead, then v1 < v=2 nÿ1 , and hence x > v1 , so PS1 0 (see equation 8.4). At date 0, B1 is willing to engage in the up-front investment i if and only if (8.6) is satis®ed and in addition v
8:7 i a max x; n : 2
The Output Decline: The Lack of Trust and Liquidity
147
Note that B1 's payoff is v=2 n ÿ i if x < v=2 n , following from lemma 1 and equation (8.3), and it is x ÿ i if v=2 n < x < v=2 nÿ1 . The following proposition states under which conditions production will take place in the presence of both the holdup problem and the credit problem. proposition 8.1 i. Suppose that there exists a holdup problem but no credit problem, that is, i > 0 and x 0. Then production takes place if and only if v > i: 2n
8:8
ii. Suppose that there exists a credit problem but no holdup problem, that is, i 0 and x > 0. Then production takes place if and only if v 2 nÿ1
> x:
8:9
iii. Suppose that there exist both a holdup problem and a credit problem, that is, i > 0 and x > 0.
. If x a v=2 n , then production takes place if and only if v b i: 2n
8:10
. If v=2 n < x < v=2 nÿ1 , then production takes place if and only if x b i:
8:11
. If v=2 nÿ1 < x, then no production takes place. Proof
See appendix II-A.
The ®rst part of this proposition restates the Blanchard and Kremer result, which says that production will take place when B1 's share of the value of production suf®ces to cover B1 's investment costs. Thus the larger are the number of production steps n, the smaller is B1 's share of the value of production and thus the more severe is the holdup problem. The second part of the proposition makes Calvo and Coricelli's point that production might not take place due to the lack of credit even in the absence of the holdup problem. Output collapses when S1 's share of the value of produc-
148
Contract Enforcement in Transition Economies
tion does not suf®ce to cover S1 's enforcement costs. Thus the larger are the number of production steps n, the smaller is S1 's share of the value of production and thus the less attractive it is for S1 to grant a credit to B1 . The last part of the proposition is particularly interesting. It shows that the presence of a liquidity and credit constraint can alleviate B1 's holdup problem. This is the case if and only if v v < i a x < nÿ1 : 2n 2
8:12
Without a liquidity constraint and enforcement costs, B1 's payoff would be v=2 n , namely half the value of production at the ®rst production step, and if i > v=2 n , then no production would take place at all. However, if enforcement costs are suf®ciently high, B1 can exploit this fact to capture more than one-half of the production value. B1 's ex post bargaining power has to be suf®ciently large to cover his ex ante investment; that is, i a x in order for production to take place. Since S1 needs a positive payoff, enforcement costs may not be too high, either, x < v=2 nÿ1 . Thus production takes place if i a x < v=2 nÿ1 . Creating a Hostage As we have seen, S1 may not be willing to deliver the input good if the credit problem is too severe, that is, if x b v=2 nÿ1 . Thus, if the buyer has no cash and the legal system to enforce payment is poorly developed, a potentially valuable transaction does not take place. In this section we investigate to what extent barter can help under these circumstances. We will show that barter can be used as a hostage, namely as a commitment device that prevents the buyer from fully exploiting his bargaining power due to the enforcement cost. In this sense barter creates a deal-speci®c collateral that helps to alleviate the holdup problem when credit enforcement is prohibitively costly. Suppose that B1 can produce one unit of a barter good, but only after date 1.8 Let w denote the value of the barter good, and let k denote B1 's production cost. If B1 sells this barter good to someone outside the production chain, he does so at a cash price pBC
w k= 2, assuming again Nash bargaining. This would give B1 a payoff of
w ÿ k=2. However, B1 can also use this barter good as a hostage to improve his creditworthiness. In this case B1 promises to deliver the barter good to S1 when credit repayment is due. The price for this
The Output Decline: The Lack of Trust and Liquidity
149
barter good, pB , is ®xed together with p1 before S1 decides about his input delivery. Of course, given that the two parties engage in Nash bargaining whenever possible, they negotiate prices p1 and pB such that they split the surplus of both transactions equally, taking into account the renegotiation on p1 at date n:1. This means that p1 and pB have to be ®xed such that
p1 ÿ x w ÿ pB v1 ÿ
p1 ÿ x pB ÿ k;
8:13
where the left-hand side represents S1 's payoff and the right-hand side B1 's payoff from carrying out both transactions. Solving this equation for p1 leads to v1 wk v1
8:14 ÿ pB x ÿ
pBC ÿ pB ; p1 x ÿ 2 2 2 where pBC
w k=2 is the price for the barter good in a cash transaction, as argued above. Recall that the price p1 that can be enforced is bounded above by v1 . Thus, for x > v1 =2, that is, when enforcement cost prevent an equal split of the surplus in the input trade, an increase in x must be compensated by a reduction in
w k=2 ÿ pB to induce the Nash bargaining solution. What this effectively means is that the inclusion of the barter trade allows B1 to shift some of the pro®t back to S1 by discounting the price of the barter good pB by an amount of pBC ÿ pB . Note, however, that pB cannot be chosen arbitrarily small because B1 cannot be forced to deliver the barter good as promised but has to be induced to do so voluntarily. If B1 cheats on S1 and refuses to deliver, all S1 can do, given that B1 has signed a contract that promises delivery of the barter good, is to try to prevent a sale of the barter good to someone else. We assume that S1 succeeds with such an attempt with probability
1 ÿ p which reduces B1 's potential payoff from selling the barter good to p
w ÿ k=2, where p a 1. This implies that B1 voluntarily delivers the barter good if and only if pB ÿ k b p
wÿk ; 2
8:15
meaning that his payoff from delivering the barter good to S1 must be at least as high as his payoff from trying to sell it to someone else.
150
Contract Enforcement in Transition Economies
Rearranging this expression leads to wk wÿk ÿ pB a
1 ÿ p 1 z; 2 2
8:16
namely B1 will discount the price for the barter good by an amount that equals at most what S1 can take away from him due to the fact that B1 has signed the barter contract. Using (8.16) in (8.14), we see that this constraint puts a lower bound on the markup for price p1 , if the surplus is to be split equally, that is, p1 b
v1 x ÿ z: 2
8:17
We can interprete z as the commitment value or hostage created by the barter contract. The larger is this value z, the less the credit enforcement cost negatively affects S1 's willingness to participate in this input deal. Considering now B1 's decision at date 0, under what conditions is he be willing to make investment i in the relationship with S1 ? Note that the alternative to investing i and carry out both the production of good 1 and the barter good is to produce only the barter good and sell it for price pBC
w k=2. This implies that B1 will undertake the investment if and only if the prices pB and p1 chosen at date 1 are such that v1 ÿ
p1 ÿ x pB ÿ k ÿ i b
wk ÿ k: 2
8:18
The following proposition characterizes how barter affects the production decision: proposition 8.2 Suppose there exists a holdup problem and a credit problem. Suppose further that S1 and B1 can use barter to create a hostage of a given size z, where z 1
w ÿ k=2
1 ÿ p.
. If x ÿ z a v=2 n , then production takes place if and only if v b i: 2n
8:19
. If v=2 n < x ÿ z < v=2 nÿ1 , then production takes place if and only if x ÿ z b i:
. If v=2 nÿ1 < x ÿ z, then no production takes place.
8:20
The Output Decline: The Lack of Trust and Liquidity
Proof
151
See appendix II-A.
Note that the size of the hostage z created by barter depends on two things. First, it depends on the value of the good offered as a means of payment in barter. When sold on the market outside of barter, this value is
w ÿ k=2 for the buyer (always assuming Nash bargaining). Second, the size of z depends on B1 's payoff when signing the barter contract and defaulting on payment which is expressed by p
w ÿ k=2. The difference between these two payoffs is determined by the parameter p, and this captures the commitment value that B1 achieves by agreeing to repay the trade credit with goods rather than cash. By doing so, B1 reduces his chances to sell the barter good to someone else than S1 . The parameter p can be thought of as a measure of how well the input seller can label the barter good as belonging to him. The smaller that p is, the less ``anonymous'' is the barter good and the smaller is B1 's cheating surplus from defaulting on payment. Thus the smaller that p is, the larger is the commitment value of barter and the larger is the hostage z.9 B1 uses the barter contract as a commitment to give S1 more than half of the value of the barter transaction, as a compensation for the fact that S1 's payoff in the input transaction is too low due to credit enforcement cost. As a consequence barter reduces the creditworthiness problem caused by the enforcement costs x. This is re¯ected in the proposition by a shift of the parameter range for which the input transaction takes place. The benchmark is no longer x but x ÿ z. To conclude, in this chapter we have established a link between the output decline, inter-®rm arrears, and barter. We have seen that ®rms can use arrears to avoid the problems associated with complexity and speci®city. The fact that input suppliers have to worry about being paid when they extend trade credits to their buyers means that they cannot exploit their buyers' dependence on these inputs. Thus the presence of arrears helps avoid the output's collapse even more than would otherwise be the case. However, very large arrears become counterproductive for maintaining production, because credit enforcement becomes very costly and thus input suppliers will refuse to extend credit to their buyers. Under these circumstances barter trade is the only way to maintain production. Barter creates a hostage that can be used to collateralize a trade credit when the ®rm's creditworthiness problem is severe. This function of barter explains why ®rms are able to give loans to each other when the banking sector is reluctant to provide capital.
This page intentionally left blank
9
Confronting the Data
In chapter 8 we have seen that introducing a credit constraint can alleviate the holdup problem. The fact that the input buyer is short of cash to pay for the inputs gives him bargaining power. This bargaining power in turn reduces the possibility that the input supplier can exploit the input buyer's need for the input. We have also seen that barter creates a deal-speci®c collateral which allows to deal with the problem of creditworthiness of ®rms. In this section we explore the predictions from our model with data of 165 barter deals in Ukraine in 1997.1 9.1
The Output Decline
Our model implies, similar to Blanchard and Kremer, that ®rms with more complex production will experience a more pronounced output loss. This can be seen by considering condition (8.8), v > i; 2n
9:1
which states that the larger is n, the more bargaining problems arise along the production chain and thus the more likely it is that the surplus of the input buyer at production step 1 will not cover his investment costs. The model implies further that the output decline will be less pronounced for ®rms short of cash. If ®rms are short of cash, they can use the credit constraint in the bargaining to prevent being held up by the input supplier. This can be seen from condition (8.11): x b i;
9:2
154
Contract Enforcement in Transition Economies
which states that the output collapse is less likely, the larger is the ex post bargaining power of the input buyer (as measured by x) relative to his ex ante investment i. However, if the ®nancial constraint becomes too large, it stops to work. When credit enforcement cost become too large, it may be too costly for the supplier to enforce payment, and thus the supplier may not be willing to deliver the input good. This can be seen from condition (8.12): v v < i a x < nÿ1 ; 2n 2
9:3
which states that credit enforcement costs have to be just right in order for the credit problem to alleviate the holdup problem. These costs have to be suf®ciently high to give the input purchaser enough bargaining power to allow him to cover his ex ante investment (x b i), but they cannot be too high; otherwise, the input supplier will refuse to participate in the deal (x < v=2 nÿ1 ). Thus we expect an inversely U-shaped relationship between ®nancial constraints and output growth. The model implies also that the ®nancial constraint should be less binding for bartering ®rms. Again, we expect an inversely U-shaped relationship between barter and output. If the ®nancial constraint is too severe for the input supplier to participate in the deal, barter contributes to maintaining production by relaxing this constraint. However, when the barter exposure becomes large, it might reduce the credit problem by so much that the input purchaser fails to be effective in capturing some of the rents from the input supplier and thus may not prevent the input purchaser from being held up.2 These predictions are consistent with the picture that emerges from table 7.6. The table shows that credit enforcement costs indeed exhibit a knife edge character when ®rm arrears are taken as a proxy for these costs. Firms with arrears below 10 percent and above 50 percent performed worse than the rest of the economy, whereas for intermediate values of ®rm arrears they performed better. A similar picture emerged for the ®rm's barter exposure. Firms with a barter share of output over 70 percent did less well, and those with a barter share of over 30 percent performed better compared to the economy as a whole. Thus our theoretical explanation seems to be consistent with the inversely U-shaped behavior between output growth, on the one hand, and ®rm arrears and barter, on the other.
Confronting the Data
155
To explore this relationship in more detail, we regress the relative output growth of the ®rm on Blanchard and Kremer's index of complexity, total arrears, and the barter share of the ®rm. We measure the relative growth performance of the ®rm by the mean percentage deviation of the output growth of the ®rm between 1994 and 1996 relative to GDP growth in Ukraine in the same period. The complexity variable is used by Blanchard and Kremer as a measure for the severity of the holdup problem. Complexity is an index that takes the value of zero if the sector uses only one input and approaches one when the sector uses several inputs from other sectors. The measure of complexity is constructed on the basis of the 1990 ``100sector'' input±output table for Russia. We matched the ISIC (International Standard of Industry Classi®cation) sector of our bartering ®rms with the sector of the complexity index given by Blanchard and Kremer. We use this variable for Ukraine, since both economies have very similar input±output structures. The ISIC classi®cation of our sample could not always be perfectly matched with Blanchard and Kremer's classi®cation of the index and might have introduced some noise into the complexity measure. The results are reported in table 9.1. Column 1 reports the result of a regression that includes only Blanchard and Kremer's index of complexity. The variable is negative and highly signi®cant, which con®rms Blanchard and Kremer's results.3 However, as equations (8.8) and (8.9) of proposition 8.1 show, the degree of complexity (the number of production steps n) worsens both the holdup problem and the credit problem. Thus the estimated effect of complexity on output growth might be due to the fact that ®rms are short of cash and face a credit constraint rather than due to the fact that they have no trust in their business partners. To distinguish between the two problems, we introduce total arrears of the ®rm into the equation as a proxy for the ®rm's credit constraint (columns 2 to 6).4 Arrears can be seen as a proxy for the credit enforcement costs x, which increase with the ®rm's indebtedness. As expected, the arrears variable has a positive sign and is highly signi®cant. The positive sign suggests that indeed the credit constraint enables the ®rm to deal with the holdup problem.5 Next, we include the ®rms' barter share into the equation (column 4). The variable has a negative and signi®cant effect on output growth. We also include a quadratic term of the barter share into the equation to capture the inversely U-shaped relationship between output growth and barter, which is signi®cant and positive.
156
Contract Enforcement in Transition Economies
Table 9.1 Output decline in the former Soviet Union: Dependent variable of relative ®rm growth All ®rms (1) Complexity
Barter < 30%
(2)
(3)
(4)
(5)
(6)
0.00 (0.22)
Arrears 2
0.07 (0.00)
(9)
(10)
0.01 (0.04)
0.01 ÿ0.01 (0.02) (0.00)
0.16 0.16 0.05 (0.00) (0.00) (0.14)
ÿ0.05 (0.00)
ÿ0.23 ÿ0.02 (0.00) (0.04)
0.10 0.13 (0.22) (0.09)
0.00 (0.00)
Barter Barter 2
2.18 (0.01)
Bank debt
N
(8)
ÿ10.60 ÿ10.48 ÿ6.58 ÿ11.20 ÿ11.42 ÿ0.57 ÿ30.92 ÿ5.70 ÿ4.76 ÿ0.91 (0.00) (0.00) (0.03) (0.00) (0.00) (0.80) (0.00) (0.25) (0.34) (0.85)
Arrears
R 2 adjusted
(7)
0.23 (0.00) 0.07 141
0.07 135
0.19 135
0.14 135
0.17 135
0.62 135
0.15 (0.00) 0.28 52
0.75 49
0.76 49
0.80 49
Source: Survey of 165 barter deals in the Ukraine in 1997. Notes: Numbers in parentheses are p-values. Percentage deviation of ®rm's output growth relative to the growth rate of GDP between 1994 and 1996. The number of cases exceeds the number of interviewed ®rms because each barter deal involves a selling and a buying ®rm. a. Firms with higher output growth than the growth rate of GDP. Complexity Complexity index of the industrial sector of the ®rm; the index is equal to zero if there is only one input, and tends to one if the sector uses many inputs Arrears Share of ®rm's arrears (tax, wage, and ®rm arrears) in percentage of output Barter Share of ®rm's barter in percentage of output Bank debt Firm's bank debt in percentage of output
To obtain a closer look at the inversely U-shaped relationship between arrears and barter, on the one hand, and output growth, on the other, we divided the data into the following subsamples: highbarter ®rms with a barter share of over 70 percent, low-barter ®rms with a barter share of less than 30 percent, and high-debt ®rms with total arrears of more than 40 percent of output. We also looked at the ®rms in the sample that performed better than the economy as a whole. The results for the two barter subsamples are given in columns 7 to 14 of table 9.1. Clearly, the regressions give a positive effect of barter on output growth for low-bartering ®rms and a negative one for high-bartering ®rms. It is interesting to see that for the sample of low-bartering ®rms, the inclusion of the arrears variable in the equation reduces the estimated effect of complexity on output substantially and the effect becomes insigni®cant at conventional levels.
Confronting the Data
157
Table 9.1 (continued) Barter > 70% (11)
(12)
Arrears > 40% (13)
(14)
(15)
(16)
High-growth ®rmsa (17)
(18)
(19)
(20)
(21)
(22)
ÿ0.90 ÿ1.84 ÿ3.69 ÿ2.27 ÿ18.36 ÿ19.46 ÿ12.26 ÿ2.82 ÿ59.96 ÿ21.20 ÿ30.00 ÿ26.91 (0.44) (0.08) (0.00) (0.02) (0.05) (0.05) (0.08) (0.00) (0.00) (0.00) (0.00) (0.00) 0.00 0.00 ÿ0.01 (0.01) (0.00) (0.07)
ÿ0.02 19
0.32 19
0.00 (0.64)
0.01 ÿ0.30 (0.40) (0.67)
ÿ0.04 0.00 (0.00) (0.97)
ÿ0.27 0.03 (0.00) (0.01)
0.11 (0.04)
0.29 (0.00)
0.80 19
0.84 19
0.12 24
0.09 24
0.57 24
0.99 24
0.12 (0.00)
0.10 (0.00)
0.08 (0.01)
0.08 (0.00)
0.08 (0.00)
0.04 (0.32) 0.77 27
0.93 27
0.95 27
0.95 27
Apparently for these ®rms the complexity variable is capturing more of a ®nancial shortage than that of an input shortage. Let us consider next the results for the sample of high-arrears ®rms given in columns 15 to 18. For highly indebted ®rms, arrears do not appear to play a role for output growth. These ®rms appear to have credit enforcement costs that are too large to make it worthwhile for the input supplier to participate in the deal. Moreover these ®rms could be so little creditworthy that even barter would not help them maintain production by getting trade credits from other ®rms. The results for the sample of high-growth ®rms are given in columns 19 to 22. These ®rms show a favorable growth performance likely because they used their credit constraint and barter activity effectively to avoid an input and ®nancial shortage. Finally, we include the share of bank debt in percent of the ®rm's output in the output growth regressions given in columns 6, 10, 14, 18, and 22 of table 9.1. This is an alternative way to capture whether or not ®rms faced a credit contraction problem. The positive and signi®cant coef®cient of the share of bank debt supports Calvo and Coricelli's view that credit contraction and the associated liquidity shortage have caused the output decline in Eastern Europe.6
158
9.2
Contract Enforcement in Transition Economies
The Shift in the Terms of Trade
In this section we test deal speci®c predictions of our model. We evaluate how the holdup problem and the credit constraint speci®ed in the previous sections are re¯ected in the terms of the barter contract. We have argued above that the holdup problem can be alleviated if the input buyer faces a credit constraint and barter is used when credit enforcement is costly for the seller. Thus we expect these problems to be re¯ected in the prices chosen in barter contracts as compared to the prices in cash deals where no such problems are present. Our model predicts that the holdup problem and the credit problem both shift the terms of trade of the barter contract in favor of the input supplier, either by an increase of p1 as compared to p1C or by a decrease of pB as compared to pBC or both. To see this, recall from equation (8.2) that the price chosen for the input good in barter is equal to v1
9:4 x; v1 : p1 min 2 Compare this price with the usual cash price for the input good if no credit enforcement and holdup problems are present. In this case the investment costs i can be contracted before investment takes place, and the buyer has no liquidity constraint and thus cannot use it to renegotiate the input price. Splitting of the surplus implies a cash price p1C : p1C ÿ i v1 ÿ p1C $ p1C
v1 ÿ i : 2
9:5
Thus p1C
v1 ÿ i=2 < minv1 =2 x; v1 p1 because the cash price re¯ects the investment cost i and does not include a markup for the credit enforcement cost x. Similarly, if the upper bound on p1 , v1 , is binding, then we expect a discount on pB as compared to the cash price pBC , as speci®ed in equation (8.14) given below: v1 wk v1 p1 x ÿ ÿ pB x ÿ
pBC ÿ pB :
9:6 2 2 2 Our model predicts further that the price discount on the barter good will be larger, the larger is the hostage z, meaning the smaller is
Confronting the Data
159
p and the larger is
w ÿ k=2. This can be seen in equations (8.15) and (8.16). Thus the more speci®c (the smaller is p) and the more liquid (the larger is
w ÿ k=2) is the barter good, the larger is the discount on pB and thus the more shifts occur in the terms of trade in favor of the input supplier. To measure the shift of the terms of trade in barter relative to the prices prevalent in cash transactions, we use the variable TOT. TOT is de®ned as the difference of SCASH and PCASH, where SCASH and PCASH are the percentage differences of barter prices compared to cash prices for the input good and the barter good, respectively. 7 Let p1C and p1 denote the price for the input good in cash and barter transactions, respectively. Similarly let the price for the barter good in cash and in barter transactions be pBC and pB . Thus the percentage price change for the input good is
p1 ÿ p1C =p1C , and the percentage price change for the barter good is (pB ÿ pBC =pBC . The net terms of trade effect is measured by TOT SCASH ÿ PCASH. To obtain a proxy for the severity of the holdup problem (a measure for n) on the input deal, we classi®ed the input good and the barter good of each transaction according to the complexity index given by Blanchard and Kremer. Using this method, we constructed a deal-speci®c complexity measure for both goods exchanged, SCOMPLEX and PCOMPLEX. Furthermore we used as a proxy for the creditworthiness (as a measure for x) of the input purchaser her total outstanding debt (®rm arrears, wage arrears, and tax arrears), PARREARS. The data allow us to distinguish whether the ®rm is on the selling or buying end of the transaction. We ®rst look at the price effect on each of the deals separately. In a next step we focus on the net effect on the terms of trade of both transactions together. Consider the regression on the percentage price change on the input deal SCASH given in columns 1 to 7 in table 9.2. The more complex is the input good, the more severe is the holdup problem in the input deal and thus the larger is the barter price p1 relative to the cash price p1C . Thus we expect a positive sign on the complexity index for the input good SCOMPLEX. This is supported by the regressions. The input-speci®c complexity measure is positive and signi®cant independent of the speci®cation.8 Furthermore we expect the input purchaser's indebtedness (PARREARS) to have a positive effect on SCASH, since the input seller in¯ates the barter input price p1 relative to the cash price p1C to cover the anticipated credit enforcement costs x. The coef®cient on PAR-
160
Contract Enforcement in Transition Economies
Table 9.2 Terms of trade effects scash (1)
pcash (2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
scomplex
0.21 0.20 (0.02) (0.03)
0.19 (0.03)
0.18 (0.04)
0.19 0.19 (0.02) (0.02)
0.20 ÿ0.11 (0.02) (0.16)
ÿ0.10 (0.18)
ÿ0.10 (0.24)
ÿ0.09 (0.27)
parrears
0.00 0.00 (0.60) (0.73)
0.00 (0.75)
0.00 (0.54)
0.01 0.01 (0.42) (0.36)
0.01 ÿ0.02 (0.35) (0.00)
ÿ0.02 (0.00)
ÿ0.02 (0.00)
ÿ0.02 (0.00)
pcomplex
11.28 11.49 (0.14) (0.15)
9.69 10.77 11.42 11.04 (0.22) (0.15) (0.13) (0.15)
pcoke
ÿ3.16 ÿ1.74 ÿ5.14 ÿ4.10 ÿ4.03 (0.40) (0.64) (0.18) (0.29) (0.30)
sstate
ÿ3.66 ÿ3.62 ÿ3.41 ÿ3.46 (0.06) (0.05) (0.07) (0.07)
pdistort
7.99 7.14 (0.01) (0.03)
repeat
N
2.12 (0.55)
1.57 (0.66) 1.43 (0.45)
7.22 (0.03)
ÿ1.80 ÿ1.77 (0.24) (0.25)
relation R 2 adjusted
ÿ11.10 ÿ13.42 ÿ12.72 (0.10) (0.08) (0.10)
ÿ1.98 (0.68) 0.06 65
0.08 64
0.08 58
0.12 58
0.21 58
0.22 58
0.21 58
0.19 65
0.21 64
0.21 58
0.20 58
Source: Survey of 165 barter deals in the Ukraine in 1997. Notes: Numbers in parentheses are p-values. scomplex Complexity index of input good; the index is equal to zero if the input good is produced with only one input and tends to one if input good is produced with many inputs parrears Total arrears of purchasing ®rm in percentage of output pcomplex Complexity index of barter good; the index is equal to zero if the barter good is produced with only one input and tends to one if the barter good is produced with many inputs pcoke Dummy variable equal to one if barter good is coke or petroleum sstate Dummy variable equal to one if selling ®rm is state owned pdistort Dummy variable equal to one if market for barter good is regulated repeat Dummy variable equal to one if seller and buyer have interacted frequently relation Dummy variable equal to one if seller is input supplier
REARS is zero and insigni®cant, suggesting that the input supplier has not been able to pass on these costs to the input purchaser. Our theory predicts for this case that the input purchaser needs to shift some of the pro®t back to the input supplier in order to make him participate in the deal by discounting the price for the barter good. Thus we expect a negative sign on the PARREARS variable in the regressions for PCASH. Looking at the regression results for the percentage price change on the barter good PCASH given in columns 8 to 14, we see this con®rmed by the data. PARREARS is negative and highly signi®cant.
Confronting the Data
161
Table 9.2 (continued) tot (12)
(13)
(14)
ÿ0.09 (0.25)
ÿ0.08 (0.30)
ÿ0.08 (0.31)
ÿ0.02 (0.00)
ÿ0.02 (0.00)
ÿ0.02 (0.01)
ÿ13.30 (0.08)
ÿ12.56 (0.10)
ÿ12.90 (0.09)
3.40 (0.38)
4.57 (0.24)
4.63 (0.24)
1.41 (0.45)
1.65 (0.38)
1.60 (0.40)
ÿ4.30 (0.17)
ÿ5.25 (0.10)
ÿ5.19 (0.11)
ÿ2.03 (0.19)
ÿ2.00 (0.20)
(15) 0.31 (0.16)
(16)
(17)
(18)
(19)
(20)
(21)
0.32 (0.00)
0.30 (0.01)
0.29 (0.01)
0.27 (0.01)
0.28 (0.00)
0.28 (0.00)
0.28 (0.00)
0.03 (0.00)
0.03 (0.00)
0.02 (0.00)
0.03 (0.00)
0.03 (0.00)
0.03 (0.00)
0.03 (0.00)
22.38 (0.02)
24.92 (0.01)
22.41 (0.02)
24.06 (0.01)
23.98 (0.01)
23.94 (0.01)
ÿ5.28 (0.24)
ÿ3.31 (0.45)
ÿ8.53 (0.05)
ÿ8.66 (0.05)
ÿ8.66 (0.06)
ÿ5.09 (0.30)
ÿ5.03 (0.02)
ÿ5.06 (0.02)
ÿ5.06 (0.02)
12.29 (0.00)
12.40 (0.00)
12.41 (0.00)
0.23 (0.90)
0.23 (0.90)
ÿ1.77 (0.72) 0.21 58
0.23 58
0.21 58
(22)
ÿ0.21 (0.97) 0.01 142
0.21 65
0.27 64
0.30 58
0.35 58
0.47 58
0.46 58
0.45 58
Consider next the net terms of trade effect of both transactions given in columns 15 to 22 of table 9.2. We expect a positive sign for SCOMPLEX and PARREARS in the TOT regressions, since a larger SCASH due to the holdup problem and a smaller PCASH due to the credit problem imply both a larger TOT. This is indeed the case. The data suggest that the holdup problem is re¯ected in an in¯ated price on the input deal; the input purchaser's credit problem appears to have been so severe that it had to be taken care of by price concessions on the barter side of the contract. Both problems have shifted the terms of trade in favor of the input seller. We predict two more variables to have affected the terms of the contract: the liquidity w and the anonymity p of the barter good. The more liquid and the less anonymous is the barter good, the larger is the hostage value of barter and thus the larger is the discount on the price of the barter good pB relative to the cash price pBC . Thus we expect a negative coef®cient of liquidity and anonymity in the PCASH regressions and a positive coef®cient for the same variables in the TOT regressions.
162
Contract Enforcement in Transition Economies
We measure the liquidity and anonymity of the barter good by PCOKE and PCOMPLEX.9 PCOKE is a dummy variable, taking the value of one if the barter good is coke or petroleum. Coke is a liquid good (everybody uses it for heating) that can be sold easily on the market at a known price. PCOMPLEX measures the complexity of the barter good. We use it as a proxy for the degree of speci®city of the barter good. If the complexity index for the barter good is large, and thus there are many production steps to get from the raw input to the ®nal good, we infer that the barter good can be potentially used only by a small number of ®rms. The more speci®c the good is for the creditor's use, the harder it will be for the debtor to cheat on repayment and to sell the good to someone else than the creditor. We therefore expect a negative coef®cient on PCOMPLEX in the PCASH regressions and a positive coef®cient in the TOT regressions. Turning to the results given in table 9.2, PCOKE has the wrong sign but is not signi®cant. PCOMPLEX has the expected sign and is highly signi®cant in all regressions. Additionally we include the variables SSTATE and PDISTORT to control for other distortions in the economy that might have in¯uenced the terms of the contract. SSTATE is a dummy variable taking the value of one if the selling ®rm is a state-owned enterprise. PDISTORT is a dummy of value one if the market for the barter good is regulated and thus pBC does not re¯ect market forces. It appears that when the seller is a state enterprise the input price is discounted and the barter price is in¯ated, suggesting that the state ®rms subsidized their buyers. In contrast, when the price for the barter good is regulated, then the contract is used to shift the terms of trade in favor of the seller rather than the buyer. Finally, we use the variables REPEAT and RELATION that capture the terms of the relationship between the input supplier and purchaser. RELATION is a dummy that takes the value of one if the seller is an energy or other input provider and zero otherwise. REPEAT is a dummy with the value of one if there is a history in the relationship between the input seller and the purchaser. RELATION measures the quality of the relationship and REPEAT the duration of the relationship between the parties. We expect both variables to have enhanced trust among the parties involved in the deal and thus to have an impact on the terms of the contract. Both variables are, however, not signi®cant in any of the regressions. In times of historic change, reputation does not appear to have governed the behavior of the parties.
Confronting the Data
163
In chapter 6 we established the puzzle that the terms of trade of the non-cash economy shifts in favor of the input supplier. We also explored in chapter 6 the virtual economy argument as one of the most in¯uential explanations for this shift. We have found that the argument is not consistent with the actual pricing behavior in barter deals. First, we did not ®nd a statistically signi®cant difference in the pricing behavior across sectors. Second, as we saw, the only sector that appears to suffer a loss from the non-cash economy is the manufacturing sector. In this chapter we offered a model based on the lack of trust and ®nancial discipline that is able to explain this shift in the terms of trade of the non-cash economy. Our model and the data suggest that barter prices on the ``sale'' side of the transaction are in¯ated because they re¯ect a trust problem and a credit problem between input suppliers and producing ®rms. Input suppliers are exploiting the fact that there are only a few suppliers around and thus they can charge higher prices for their inputs in barter deals compared to cash deals (this is how the trust problem materializes). If this price markup for inputs happened in cash deals, ®rms would refuse to buy those expensive inputs and prefer not to produce. Furthermore input suppliers have to incur cost to enforce payment (they have to involve the Ma®a or legal ®rms) that they want to be covered by the deal. If these costs become very large (which happens when legal institutions do not work properly or when ®rms are already very indebted), then input suppliers will refuse to deliver the inputs in cash deals due to high costs of credit enforcement, which makes the deal unpro®table for the supplier. Thus in a cash economy the lack of trust and ®nancial discipline prevent many pro®table trades from taking place. In a non-cash economy the deal can go through by choosing the right prices for the ``sale'' and the ``goods payment'' for the following reasons: First, by introducing a second pro®table transaction in the form of the ``goods payment,'' the producing ®rm can buy an in¯ated input and still make a pro®t.10 Second, the input supplying ®rm gets a discount on the barter good, which allows her to cover the credit enforcement cost. Thus the non-cash economy helps to maintain output that otherwise would collapse due to imperfect input and credit markets. The imperfection of input and credit markets are re¯ected in a shift in the terms of trade of barter. Through the in¯ated price for the ``sale'' and the price discount on the ``goods payment,'' the deal is actually saved by guaranteeing both parties a positive pro®t.
164
Contract Enforcement in Transition Economies
The shift in the terms of trade is the mechanism by which the noncash economy accomplishes to maintain output that otherwise would collapse in a cash economy. 9.3
Time and Cross-country Pattern
The purpose of our analysis in the second part of the book is to establish a link between the output decline, inter-®rm arrears, and barter in postsocialist economies. So far from our analysis the following picture has emerged. Arrears and barter are two sides of the same phenomenon. Arrears are a trade credit in cash and barter is a trade credit in goods. Both are a response to a banking failure. Firms lack the cash to pay for their inputs and banks refuse to provide capital. This has led to the phenomenon of inter-®rm arrears in which ®rms extend trade credits to each other. We argued further that barter comes into play when arrears become so critically large that ®rms refuse to extend further trade credits to their buyers for fear of not being paid. By introducing a hostage, barter allows the debtor ®rm to make a commitment to repay the loan and thus restores the creditworthiness for one speci®c deal. The data we presented in this chapter supports this view of barter and arrears. We then investigated the link between the output decline and arrears. We explored two explanations for the output fall in transition economies: the input shortage explanation due to disorganization and the ®nancial shortage explanation. We combine these two explanations and show that the ®nancial shortageÐthe fact that the buyer has no cash to pay for the inputs and requires a trade credit from his supplierÐcan help the ®rm to prevent from being exploited by his input supplier. This problem arises when the supplier's bargaining position is prohibitively strong due to high costs of switching suppliers. As we showed earlier in this chapter, we ®nd supporting evidence for this role of arrears as alleviating the trust problem in the Russian economy. Our theory is able to explain the cross-country as well as the time pattern of arrears and barter. We argue that the former Soviet Union differs from the early transition economies in Central Europe because disorganization and speci®city have posed a more severe problem. In open economies like Central Europe, entry of foreign ®rms alleviated the problems of small numbers. This is why we see the output
Confronting the Data
165
fall to be more pronounced in the former compared to the latter. In the former Soviet Union mechanisms other than international trade and foreign direct investment must have been at work to limit the adverse effect of speci®city. We argue that inter-®rm arrears and barter can be seen as such mechanisms to deal with problems arising in the transition. Inter®rm arrears are larger and growing faster in the former Soviet Union compared to Central Europe because they help deal with disorganization and holdup. Barter is observed in the former Soviet Union while being absent in Eastern Europe because the arrears crisis was more severe in the former than the latter region. Our model can also explain the pattern of arrears and barter over time in the former Soviet Union. The arrears crisis started in 1992 and 1993 in Russia, reaching almost 10 percent of GDP (®rm and tax arrears), while barter started to rise in 1994. The theory predicts barter to arise when arrears reach a critical level at which credit enforcement becomes so costly that only barter can maintain production. Our data suggest that arrears reached this critical level at around 30 to 40 percent of ®rms' sales. At this point, ®rms refused to extend further credit to each other out of the fear of not being paid. Barter then stepped in as the only way to maintain production. At this point, barter started to substitute for the nonactive banking sector as well as for trade credits in cash, which explains the explosive increase. Thus we argue that barter in Russia since 1994 has been triggered by a level of arrears at which production was unsustainable.
This page intentionally left blank
Appendixes to Part II
Appendix II-A Proof of Lemma 8.1 Suppose that all buyers are liquidity constrained and thus have to delay repayment. Consider now the ®nal step of production. This is exactly like the ®rst step, with the only exception that Bn does not have to make an up-front investment. Then the two parties share the joint surplus as follows: Sn receives pn ÿ x v=2 and Bn receives v ÿ pn x v=2, provided that x a v=2. If v=2 < x < v, then Bn receives x and Sn receives v ÿ x. If x b v, then no production takes place at the ®nal step. The value of production at step n ÿ 1 is equal to the payoff of the seller at the last production step. Solving the game recursively, this leads to the following value of production at step j; j 1; . . . ; n ÿ 1: 8 v v > > if x a nÿj , > > nÿj > 2 2 > > < v v v vj
1 ÿ x if nÿj < x < nÿjÿ1 , nÿjÿ1 > 2 2 2 > > > > > v > :0 if nÿjÿ1 a x. 2 Note that if x > v=2 nÿjÿ1 , production will not take place at step j 1 because the seller Sj1 will not be able to guarantee himself a positive payoff. Thus the value of production at step j is zero. r
168
Contract Enforcement in Transition Economies
Proof of Proposition 8.1 Production takes place if and only if equations (8.6) and (8.7) are satis®ed. i. If x 0, equation (8.6) is satis®ed by assumption and (8.7) is satis®ed if and only if i a v=2 n . ii. If i 0, equation (8.7) is satis®ed by assumption and (8.6) is satis®ed if and only if x < v=2 nÿ1 . iii. Suppose that i > 0 and x > 0. If x a v=2 n , then (8.6) is satis®ed by assumption and (8.7) is satis®ed if and only if i a v=2 n . If v=2 n < x < v=2 nÿ1 , then (8.6) is satis®ed by assumption and (8.7) is satis®ed if and only if i a x. If x > v=2 nÿ1 , then (8.6) is violated and no production takes place. r Proof of Proposition 8.2 Note ®rst that we can determine the value of v1 just as in lemma 8.1, with the only difference that now the enforcement cost x are reduced by the hostage z. Of course, barter needs to take place only at those production steps j where x > vj =2, that is, where the liquidity constraint prevents an equal sharing of the surplus at production stage j. With this in mind, we can solve the game recursively as done in lemma 8.1, with the only difference that now instead of x we have to consider x ÿ z whenever x is hitting this constraint. Thus we have v1
x; z v=2 nÿ1 if x ÿ zav=2 nÿ1 and v1
x; z < v=2 nÿ1 if x ÿ z > v=2 nÿ1 . Recall that p1 and pB are ®xed such that the parties share the surplus equally whenever possible. This implies that we can set pB
wk ÿz 2
2
and allows us in turn to set v1 p1 min x ÿ z; v1 : 2
3
Note that both S1 and B1 need to be willing to participate and make the necessary investment. For B1 it requires that v1 ÿ
p1 ÿ x pB ÿ k ÿ i a
wÿk 2
4
Appendixes to Part II
169
and for S1 it requires that
p1 ÿ x w ÿ pB >
wÿk : 2
5
Suppose that x ÿ z < v=2 nÿ1 . Then v1 v=2 nÿ1 . Suppose further that x ÿ z < v=2 n v1 =2. Then, using the equations for p1 and pB in (4) yields v1 v n b i; 2 2
6
and in (5) yields v1 v n > 0: 2 2
7
Note that this is condition (8.19) in proposition 8.2. Suppose next that x ÿ z < v=2 nÿ1 so that v1 v=2 nÿ1 , but that v=2 n v1 =2 < x ÿ z < v=2 nÿ1 v1 . Then, using the equations for p1 and pB in (4) yields x ÿ z b i;
8
and in (5) yields v1
v > x ÿ z: 2 nÿ1
9
Note that this is condition (8.20) in proposition 8.2. Finally, note that if x ÿ z > v=2 nÿ1 , then v1 < v=2 nÿ1 , and it is not possible to satisfy (5). Therefore no production takes place as speci®ed in proposition 8.2. r
170
Contract Enforcement in Transition Economies
Appendix II-B De®nition of Variables and Sample Statistics
Maximum
Standard deviation
1.00
687.90
101.068
65.30
0.00
687.90
157.947
Firm's bank debt in percentage of output
6.31
0.00
104.20
15.955
150
Firm's tax arrears in percentage of output
7.15
0.00
121.50
19.306
wage arrears
150
Firm's wage arrears in percentage of output
3.38
0.00
38.60
6.001
®rm arrears
138
Firm's inter-®rm arrears in percentage of output
30.15
0.70
626.00
90.887
relative ®rm growth
153
Percentage deviation of ®rm's output growth relative to the growth rate of GDP between 1994 and 1996
0.01
ÿ2.03
25.43
4.215
barter
165
Share of ®rm's barter in percentage of output
45.21
1.00
100.00
28.181
complexity
141
Complexity index of the industrial sector of the ®rm; the index is equal to zero if there is only one input and tends to one if the sector uses many inputs.
0.80
0.34
0.92
0.116
scomplex
142
Complexity of input good; the index is equal to zero if the input good is produced with only one input and tends to one if input good is produced with many inputs
0.77
0.30
0.92
0.139
Variable
Observations
arrears
Description
Mean
138
Share of ®rm's arrears (tax, wage, and ®rm arrears) in percentage of output
41.40
69
Total arrears of purchasing ®rm in percentage of output
bank debt
150
tax arrears
parrears
Minimum
Appendixes to Part II
171
Variable
Observations
pcomplex
145
Complexity of barter good; the index is equal to zero if the barter good is produced with only one input and tends to one if barter good is produced with many inputs.
0.75
0.19
0.92
0.134
tot
163
Net difference between cash and barter price in percent (scash ÿ pcash)
4.59
ÿ168.00
50.00
18.075
scash
163
Difference between the barter price and cash price in percent of the cash price for the input good
3.43
ÿ16.70
49.00
8.302
pcash
163
Difference between the barter price and cash price in percent of the cash price for the barter good
ÿ1.16
ÿ50.00
186.00
18.633
sstate
165
Dummy variable equal to one if selling ®rm is state owned
D 1, 49 observations
pdistort
165
Dummy variable equal to one if market for barter good is regulated
D 1, 36 observations
pcoke
148
Dummy variable equal to one if barter good is coke or petroleum
D 1, 16 observations
relation
164
Dummy variable equal to 1 if seller is input supplier
D 1, 88 observations
repeat
165
Dummy variable equal to 1 if seller and buyer have interacted frequently
D 1, 87 observations
Description
Mean
Minimum
Maximum
Standard deviation
This page intentionally left blank
10
Contracts in Trade and Transition: An Outlook
In this book we have explored the role of countertrade and barter in international trade and economic transition. We have seen that these forms of trade can be interpreted as institutions. They solve problems associated with weak contract enforcement when the legal system or reputation as an enforcement mechanism do not work. In this concluding chapter we speculate about the role that these alternative institutions will play in the future. How important will be the problems that countertrade and barter are supposed to solve and what alternative institutions exist to solve these problems? 10.1
Contracts in Trade: A Shift to Foreign Investment?
In international trade we have focused on two issues: the capital transfer problem and the technology transfer problem. The capital transfer problem relates to lending capital to countries whose high indebtedness makes it less likely that they will repay their debt. The technology transfer problem relates to the problem of selling up-todate technology to Eastern Europe or developing countries when the technology seller in an industrialized country has an incentive to deliver an inferior technology. Will these problems continue to be relevant for these regions in the future? A look at table 10.1 reveals that in many of the countries that have used countertrade in the 1980s, creditworthiness has deteriorated much since 1987. Take the example of the former Soviet Union whose debt to GDP ratio was 10 percent in 1987. This ratio increased to 26 percent in Russia, to 57 percent in Moldova, and to 43 percent in Kyrgyztan in 1999. Some of these countries did substantially reduce their external debt. Examples are Poland and Hungary. Overall, however, the transition economies' external debt has not improved over the last decade.
174
Contracts in Trade and Transition: An Outlook
Table 10.1 Debt to GDP ratio in selected countries 1987 Transition economies (average)
1990
1997
34
19
36
Albania
5
17
28
Bulgaria
29
57
101
Former Czechoslovakia
13
Ð
Ð
Ð
20
42
Slovak Republic Former East Germany
Ð 49
13 Ð
52 Ð
Hungary
78
67
55
Former Yugoslavia
35
20
Ð
Ð
22
35
Czech Republic
Croatia Macedonia, FYR
Ð
64
71
Slovenia
Ð
15
n.a.
Poland
70
89
29
Romania Former Soviet Union
17 10
3 6
33 27
Armenia
Ð
7
38
Azerbaijan
Ð
1
12
Belarus
Ð
4
5
Estonia
Ð
4
14
Georgia
Ð
16
28
Kazakhstan
Ð
6
19
Kyrgyzstan Latvia
Ð Ð
7 4
43 9
Lithuania
Ð
4
16
Moldova
Ð
5
57
Russian Federation
Ð
10
26
Tajikistan
Ð
13
45
Turkmenistan
Ð
5
63
Ukraine
Ð
3
22
Ð 79
5 58
11 56
Uzbekistan East Asia (average) Indonesia
69
64
65
Malaysia
77
40
51
Philippines Developing countries (average)
91
69
53
103
162
83
Algeria
37
48
69
Argentina
56
46
39
Brazil China
44 12
28 16
24 17
Contracts in Trade and Transition: An Outlook
175
Table 10.1 (continued) 1987
1990
1997
Cyprus
57
Ð
Ð
Ecuador
107
123
87
Egypt
163
97
39
India
22
28
25
Iran Nicaragua Syria
5
8
10
208 146
1081 148
306 126
Togo
103
80
93
Zambia
327
242
185
55
Ð
58
Zimbabwe
Sources: Global Development Finance 1988, 1997, 1999; The World Bank.
One can safely say that the capital transfer problem will continue to be relevant for transition economies and developing countries in the future. More recently new countries with a problem of creditworthiness have emerged. The ®nancial crisis in East Asia made countries in this region less creditworthy due to large outstanding domestic debt. The ®gures in table 10.1 suggest this to be the case despite the fact that the outstanding foreign debt of the East Asian countries has improved over the last decade. We now turn to the technology transfer problem. Will the transition economies and developing countries depend on the import of Western technology in order to close the income gap to the industrialized countries in the future? Since the collapse of communism the income gap between transition economies and western market economies has widened. As can be observed in table 10.2, after liberalization, output declined in all of the transition economies. The output decline has been more drastic in the countries of the former Soviet Union compared to the early transition economies like Hungary, Poland, and the Czech Republic. Thus the issue of technology transfer to close the income gap is becoming more important for transition economies than it used to be in the past. Have the prospects of these regions to solve the capital transfer and the technology transfer problem improved since the 1980s? Under central planning, in Eastern Europe alternative mechanisms to solve the capital and technology transfer problem that exist in market
176
Contracts in Trade and Transition: An Outlook
Table 10.2 Growth of real GDP in transition economies, 1990 to 1997 (in % per annum) Growth of real GDP 1990 CEE
1991
1992
1993
1994
1995
1996
1997
ÿ6.4
ÿ14.2
ÿ7.2
0.2
3.4
4.3
2.9
0.8
Albania
ÿ10.0
ÿ28.0
ÿ7.2
9.6
9.4
8.9
9.1
ÿ7.0
Bulgaria
ÿ9.1
ÿ11.7
ÿ7.3
ÿ1.5
1.8
2.1
ÿ10.9
ÿ7.4
Croatia
ÿ7.5
ÿ17.0
ÿ11.7
ÿ0.9
0.6
1.7
4.3
6.3
Czech Republic
ÿ1.2
ÿ11.5
ÿ3.3
0.6
2.7
5.9
4.1
1.2
Hungary
ÿ3.5
ÿ11.9
ÿ3.1
ÿ0.6
2.9
1.5
1.3
4.0
FYR Macedonia
ÿ7.5
ÿ17.0
ÿ21.1
ÿ9.4
ÿ2.7
ÿ1.6
0.9
1.5
ÿ11.6 ÿ5.6
ÿ7.0 ÿ12.9
2.6 ÿ8.8
3.8 1.5
5.2 3.9
7.0 6.9
6.1 3.9
6.9 ÿ6.6
Poland Romania Slovak Republic
ÿ0.4
ÿ15.9
ÿ6.7
ÿ3.7
4.6
6.8
7.0
5.7
Slovenia
ÿ7.5
ÿ8.9
ÿ5.5
2.8
5.3
4.1
3.2
3.7
Baltics
ÿ2.3
ÿ8.3
ÿ25.5
ÿ13.8
ÿ3.6
2.3
4.0
5.7
Estonia
ÿ2.3
ÿ7.9
ÿ21.6
ÿ8.2
ÿ1.8
4.3
4.0
5.0
Latvia
ÿ2.3
ÿ11.1
ÿ35.2
ÿ16.1
2.1
0.3
2.8
6.0
Lithuania
ÿ2.3
ÿ6.0
ÿ19.6
ÿ17.1
ÿ11.2
2.3
5.1
6.0
Russia
ÿ2.3
ÿ5.0
ÿ14.5
ÿ8.7
ÿ12.6
ÿ4.0
ÿ2.8
0.8
Other FSU
ÿ6.3
ÿ8.4
ÿ21.2
ÿ12.1
ÿ15.2
ÿ5.5
ÿ1.9
1.4
Armenia
ÿ2.2
ÿ12.4
ÿ52.6
ÿ14.1
5.4
6.9
5.8
3.3
Azerbaijan
ÿ11.7
ÿ0.7
ÿ22.1
ÿ23.1
ÿ18.1
ÿ11.0
1.3
5.0
Belarus
ÿ2.3
ÿ1.2
ÿ9.7
ÿ7.6
ÿ12.6
ÿ10.4
2.8
10.0
Georgia
10.0
ÿ12.4
ÿ20.6
ÿ44.8
ÿ25.4
ÿ11.4
2.4
10.5
Kazakhstan
ÿ2.3
ÿ11.0
ÿ5.3
ÿ10.6
ÿ12.6
ÿ8.2
0.5
2.1
Kyrgyzstan Moldova
ÿ2.3 ÿ2.4
ÿ7.9 ÿ17.5
ÿ13.9 ÿ29.1
ÿ15.5 1.0
ÿ20.1 ÿ31.0
ÿ5.4 ÿ1.4
5.6 ÿ7.8
6.2 1.3
ÿ32.6
ÿ7.1
ÿ28.9
ÿ11.1
ÿ21.4
ÿ12.5
ÿ28.3
2.2
Turkmenistan
Tajikistan
ÿ2.3
ÿ4.7
ÿ5.3
ÿ10.0
ÿ18.8
ÿ8.2
ÿ3.0
ÿ24.0
Ukraine
ÿ3.6
ÿ8.7
ÿ9.9
ÿ14.2
ÿ22.9
ÿ12.2
ÿ10.0
ÿ3.4
4.3
ÿ0.5
ÿ11.1
ÿ2.3
ÿ4.2
ÿ0.9
1.6
2.4
Mongolia
ÿ5.6
ÿ9.2
ÿ9.5
ÿ3.0
2.3
6.3
2.6
4.0
Germany
5.5
2.1
ÿ1.2
1.5
1.6
0.9
1.7
1.9
Uzbekistan
Sources: International Monetary Fund and Statistisches Bundesamt.
Contracts in Trade and Transition: An Outlook
177
economies were not available. Foreign ownership of assets was prohibited. Thus foreign direct investment and joint ventures were rare before 1989. Today, when the technology seller can own and control the economic activity in another country as is the case with foreign direct investment, the technology ®rm in general will have highpowered incentives to use the appropriate technology because it controls the assets and enjoys the pro®ts generated with those assets. Foreign direct investment is now another way for the technology ®rm to internalize the bene®ts of its action and thus to solve the technology transfer problem. Countertrade is a contract that achieves the same result in a market transaction when ownership of assets is not an option. What are the advantages and disadvantages of foreign direct investment for solving the capital transfer problem? After the international debt crisis of the 1980s, it has been argued that foreign direct investments are less vulnerable to country risks because of the ¯exible payment schedules and extended property rights associated with the investment. But the experience of developing countries shows that direct investments are also subject to sovereign risks. Just as a government can choose to default on its debt service, it can choose to expropriate the assets of a direct investment. In addition direct investments face the risk of creeping expropriation through, say, taxes, speci®c import or export duties. Thus a priori it is not clear whether or not foreign direct investment is better in solving the capital transfer problem than credit arrangements secured with countertrade contracts.1 Insecure property rights and drastic changes in the policy environment in transition economies may thus be the reason why the investment ¯ows to some of the transition economies are much smaller than has been expected.2 Table 10.3 illustrates that countertrade as a contractual solution to the capital and technology transfer problem has declined in most of these countries since the fall of communism. Foreign direct investment and joint ventures as solutions to contractual hazards have increased in Hungary, the Czech Republic, and Poland, while the former states of the Soviet Union have been less successful in attracting foreign investors. Our work predicts that as long as the political and legal environment is risky in these economies, countertrade will continue as an institutional arrangement to secure the technology and capital transfer to these countries.
178
Contracts in Trade and Transition: An Outlook
Table 10.3 Institutional structure of East±West trade Exports imports Before 1989 After 1989
Barter countertrade a
@60%
Joint ventures, foreign direct investment b
@40% 1990
0%
1995
1990
1992
1998
Eastern Europe
1.0%
4.6%
Bulgaria
0.1%
0.9%
8.0%
0.4%
10.4%
Croatia
8.6%
Czech Republic
1.5%
9.7%
8.8%
Hungary Poland
3.5% 0.7%
13.2% 1.8%
7.5% 20.0%
1.2%
17.2%
Slovakia
0.2%
2.6%
3.9%
Slovenia
0.1%
1.8%
1.6%
Baltics
5.6%
13.5%
Estonia
14.3%
12.1%
Latvia
4.0%
11.2%
Lithuania FSU
1.7% 4.3%
16.0% 13.4%
Romania
@10%
@6%
Belarus Kazakhstan Russia
@50%
Ukraine
@50%
Total
@20%c 0.5%
0.9%
4.8%
21.3%
51.3%
3.9%
8.5%
9.8%
11.0%
4.5%
10.0%
Sources: Economic Commission of Europe, ECE, No. 3, 1999; Vienna Institute for Comparative Economic Studies; OECD: East±West Trade: Recent Developments in Countertrade 1981. a. Percentage of total trade. b. Incoming foreign direct investment ¯ows as a ratio of total imports of the respective country. c. 1996
Contracts in Trade and Transition: An Outlook
10.2
179
Contracts in Transition: An Institutional Trap?
In the context of economic transition we dealt with the problem of liquidity and the lack of trust in the economies of the former Soviet Union. We have argued that inter-®rm arrears and barter are a response to an underdeveloped ®nancial and legal system in which contracts and credit repayment are only weakly enforced. As long as proper ®nancial and legal institutions do not exist in these countries, barter and arrears will continue to play a role to maintain production. Barter seems to have established itself as an economic institution to deal with the banking failure and capital market imperfections in transition economies.3 We have argued in this book that barter has produced short-term bene®ts by allowing these economies to maintain production without a functioning banking sector. However, the short-term ef®ciency gain might come at the costs of long-term ef®ciency losses. As Huang, Marin, and Xu (2001) suggest, in these economies barter might have established itself as an institution that hinders the banking sector from developing. The banking sector in the former Soviet Union is hindered from developing by the soft budget constraint of the government and by the high costs of bank lending. Bank lending is costly in these countries because of severe asymmetric information problems between creditors and lenders. This gives rise to an adverse selection problem where good risk ®rms turn away from bank lending because they are subsidizing bad risk ®rms. Good risk ®rms ®nd barter to be a cheaper way to raise liquidity. This in turn generates an externality on the lending rates of banks. The more liquidity is ®nanced by barter trade, the more costly becomes bank lending and even fewer ®rms will turn to banks for credit. This gives rise to an institutional trap for the banking sector to develop. Kranton (1996) suggests a different reason why barter might generate an institutional trap. She points to a coordination problem that might arise in a system of personalized exchanges. She argues that when more people engage in barter and ``personalized'' exchange, market search costs increase, and thus it becomes harder to exchange goods for money in an anonymous market. This way the incentive to maintain ``personalized'' exchange increases. Barter and personalized exchange get ``locked in'' and thus will persist over time even though it is a less ef®cient form of exchange.
180
Contracts in Trade and Transition: An Outlook
A further reason why a barter economy might be a self-perpetuating system is that second-best solutions may deter political demand for better institutions. Thus an economy might be stuck with a secondbest institution when a more ef®cient would be available.4 Can a ®rst sign of such an institutional trap be identi®ed in the data across transition economies? If such a mechanism were at work, we would expect that (1) in transition economies with underdeveloped ®nancial and legal systems barter will be more predominant and (2) transition economies with a large and increasing exposure to barter trade will see their ®nancial sector not develop. In table 10.4 we look at these questions for 20 transition economies. Columns 1 to 3 examine whether in 1996 differences in the importance of barter in a country (measured by the share of barter in percent of sales) can be accounted for by differences in the development of their ®nancial and legal system. As a proxy for ®nancial sector development, we use the EBRD index of effectiveness of ®nancial and commercial regulations. These indexes should be interpreted with caution, since they re¯ect the subjective perception of ®rms how well the regulations work in these countries rather than being an objective indicator of ®nancial development.5 The indexes go from 1 to 4, with 1 giving a least effective regime and 4 indicating a very effective regime. Additionally we use differences in bank intermediation across transition economies as a proxy for the presence of a banking failure (measured by credit to the private sector in percent of GDP, bank credit). The more bank credit is available the less we expect countries to turn to barter trade. Looking at column 1, bank credit is, as expected, negative but not signi®cant at conventional levels. Turning to the two EBRD indexes measuring ®nancial and legal regulation, we predict on both variables a negative sign. Countries with better legal and ®nancial systems are expected to have less barter. It turns out that both variables are signi®cant, but only the legal ef®ciency variable has the expected sign. Additionally we included the exchange rate and the share of exports in GDP as variables accounting for differences in barter. A country with not too strong a currency will ®nd it easier to sell its goods abroad. This allows its ®rms to ®nance production from internal cash ¯ow rather than by raising liquidity through barter. The coef®cient on the exchange rate is negative and signi®cant, suggesting that weaker exchange rates indeed appear to have played such a liquidity generating role.6 Further we expect the share of exports to
Contracts in Trade and Transition: An Outlook
181
Table 10.4 Macro- and microeconomic determinants of barter: Dependent variable barter 1996 (1) Bank credit Financial regulation Legal ef®ciency
Percent change, 1996±1999 (2)
(3)
(4)
ÿ0.17 (0.428) 7.52 (0.050)
0.02 (0.556) 5.65 (0.083)
ÿ4.98 ÿ7.52 (0.100) (0.006)
Exchange rate
ÿ0.00 (0.080)
Export share
(6)
(7)
0.04 (0.262)
6.71 (0.063)
ÿ0.13 (0.313)
ÿ8.17 ÿ0.12 ÿ0.18 ÿ0.17 ÿ0.12 (0.005) (0.067) (0.020) (0.055) (0.163) ÿ0.00 (0.063)
0.01 (0.183)
ÿ14.07 (0.420)
Export R 2 adjusted N
(5)
0.01 (0.458)
ÿ0.29 ÿ0.29 ÿ0.37 (0.037) (0.076) (0.045) ÿ0.46 (0.010)
0.16 18
0.32 19
0.31 19
0.36 18
0.29 18
0.16 19
0.18 19
Sources: World Business Environment Survey, World Bank±EBRD 1999, EBRD Transition Report 1999; IMF International Financial Statistics 1999. Notes: The percent change regressions in columns 4 to 7 were run for ®rst differences between 1996 and 1999 except for the indexes of legal ef®ciency and ®nancial regulation where absolute differences are used; p-values are in parentheses. Barter Share of barter in percentage of sales in 20 transition countries Bank credit Credit to the private sector in percentage of GDP in 20 transition countries Financial regulation EBRD index of effectiveness of ®nancial regulation; index goes from 4 (very effective) to 1 (least effective) Legal ef®ciency EBRD index of effectiveness of commercial regulation; index goes from 4 (very effective) to 1 (least effective) Exchange rate Local currency in U.S.$ in 20 transition countries Export share Exports in percentage of GDP in 20 transition countries Export Export growth between 1996 and 1999 in 20 transition countries
182
Contracts in Trade and Transition: An Outlook
help a country to reduce its exposure to barter because of the following reasons. First, countries with a larger trade exposure will have an outside option to buy inputs from foreign ®rms and thus will be less exposed to the problem of trust inside the country. Barter as an institution to deal with this problem will be less needed. Second, countries with a larger share of goods sold to foreign countries will have more cash available to ®nance production lowering their need to turn to barter trade. Looking at column 3 of table 10.4, we ®nd, however, the export share not to be signi®cantly different from zero. Columns 4 to 7 examine whether the evolution of barter over the period 1996 to 1999 can be explained by changes in the problem of banking failure and changes in the legal as well as the macroeconomic environment. It appears from the table that the remonetization of the economies in transition in this period is exclusively driven by improvements in the legal ef®ciency (a larger number in the absolute difference in this index between 1999 and 1996 indicates an improvement) and by a strong export growth in this period. Improvements in bank intermediation or in ®nancial regulation appear not to have contributed to abolishing the non-cash economy. This indicates that a banking-barter trap might indeed be at work here. 10.3
What Can Be Done?
Given our explanation for the role of barter and countertrade in international trade and transition economies, what follows for policy? In international trade, countertrade is an institutional response to imperfect international capital and technology markets. Foreign direct investment solves the same contractual hazards and thus should be seen as an alternative institution dealing with the technology and capital transfer problem. Some governments in transition economies have tried to increase the foreign investors involvement in their markets by giving tax holidays or other forms of support, with varied success. But we expect foreign direct investment to increase when these economies guarantee property rights and avoid drastic changes in the policy environment. In transition economies, the picture looks somewhat different. Here domestic barter is an institutional response to imperfect domestic ®nancial and input markets. What can be done in order to remone-
Contracts in Trade and Transition: An Outlook
183
tize the economies of the former Soviet Union? Given that the ®nancial squeeze is the source of barter and arrears in transition economies, one obvious candidate is expansionary monetary policy to overcome the liquidity shortage. But expansionary monetary policy might make matters worse. In a barter economy like Russia's, a monetary expansion can have perverse effects. In the framework of the model we developed in chapter 8, reducing arrears by fusing liquidity into the economy could eliminate input buyer's bargaining power, thereby robbing the instrument to deal with the ``disorganization'' problem of the economy. This might lead to less rather than more output. Thus, in a distorted economy like Russia's, eliminating one distortionÐarrears through as monetary expansionÐwithout removing the other distortionÐthe trust problem in the economyÐ will not lead to desired outcomes. Monetary policy might lead either to lower output, because production breaks down due to the trust problem, or to higher in¯ation, because input suppliers will in¯ate their inputs to exploit their bargaining power when the trust problem is not too severe. In other words, a monetary expansion in a barter economy works like introducing partial reform in an overall distorted economy.7 Since the ®nancial crisis of 1998, barter has fallen by more than 20 percent in Russia and Ukraine. This fall provides some clue of possible ways to remonetize the economies of the former Soviet Union. Our theory suggests that the way to make the non-cash economy disappear is to address the problem of banking failure. The banking failure is the source of the barter economy. What is the source of the banking failure and how has the ®nancial crisis of August 1998 affected the banking sector? The ®nancial crisis in Russia led to a dramatic decline in interest rates and exchange rates. In Russia the ruble depreciated by more than 50 percent, and commercial banks three months lending rates declined by about 40 percent while the treasury bills market collapsed. As Huang, Marin, and Xu (2001) suggest, the ®nancial crisis acted as a trigger for banking development in Russia by allowing the economy to get out of a banking±barter trap. With the collapse of the treasury bills market the banking sector lost an important market to invest its money. Banks turned back to the real sector for business opportunities. They lowered interest rates and started to lend to ®rms. Lower interest rates in turn made it attractive for some ®rms
184
Contracts in Trade and Transition: An Outlook
to borrow from banks rather than to raise liquidity through barter trade. This is why barter has declined with the ®nancial crisis in Russia. Huang, Marin, and Xu's analysis indicates that the most important instrument to remonetize the Russian economy is to remove the soft budget constraint of the government. This soft budget constraint was the prime reason why the treasury bills market offered incredibly high rates of return for banks. Banks realized higher returns on ®nancing the government budget than on ®nancing the real sector of the economy. Thus the government budget de®cit led to a crowding out of bank lending to the real sector. The ®nancial crisis removed this opportunity and this way stimulated bank lending to the real sector.8 The exchange rate depreciation of the ruble also played an important role in remonetizing the Russian economy. In table 10.4 we see that a strong export growth is an important vehicle to remove the non-cash economy in transition countries. Evidently, as the table shows, export growth helps reduce the barter economy by offering an outside opportunity to the trust problem of the economy and by generating internal liquidity to ®nance production. Huang, Marin, and Xu (2001) suggest an additional reason why export growth helps remonetize the economy. Strong export growth lowers the credit risk of the real sector creating an incentive for banks to lend to this sector at lower rates. The exchange rate depreciation stimulated this export growth by improving ®rms' competitiveness in international markets and helping them gain market access. Thus the depreciation of the ruble helped the ®nancial sector connect with the real sector by giving the banking sector an incentive to lend to the real sector. If the transition economies can establish themselves on the international markets after the exchange rate induced entry, the ®nancial crisis might have a lasting positive effect on the credit risk of the real sector and on bank lending. Hence the ®nancial crisis might have helped lower the need for barter in the future.
Notes
Chapter 1 1. OECD: East±West Trade: Recent Developments in Countertrade, 1981. 2. For a recent estimate for Russia, see Commander and Mumssen (1998) and for an estimate for Ukraine, see Marin, Kaufmann, and Gorochowskij (2000). 3. The reason why the estimates of barter of Commander and Mumssen (1998) on Russia and of Marin, Kaufmann, and Gorochowskij (2000) on Ukraine differ from the World Business Environment Survey estimates is that the former two studies include bartering ®rms only while the latter considers bartering as well as nonbartering ®rms. 4. See, for example, Sachs (1989). Gooptu and Soledad (1992) emphasize that commercial banks were reluctant to provide new loans unless they were insured by the creditor's governments. 5. Marin (1990). 6. See Fischer (1996), Fantapie (1994), and Jalloh (1990). For the importance of countertrade in the Uruguay Round, see Guisinger (1987). 7. For enforcement of law in transition economies, see Hay and Shleifer (1998), Greif and Kandel (1994), Johnson, McMillan, and Woodruff (1999), and Murrell (2000). For institutions that govern international trade, see Yarbrough and Yarbrough (1992). 8. For an overview of other institutions governing international trade in the Commercial Revolution, see Greif (1992). Recently Greif and Kandel (1994) applied this framework to economic transition. 9. In their ex post assessment after ten years of transition Fischer and Sahay (2000), however, argue that the reform consensus included elements of legal reforms that are now thought to have been overlooked.
Chapter 2 1. Almost all previous empirical studies on barter use macro data and test (on the basis of relatively few observations), for example, how debt ratios of various countries affect the estimated volume of barter in these respective countries. 2. World Debt Tables, World Bank (1994). 3. For other forms of international business cooperations, see Oman (1984).
186
Notes to Pages 22±62
4. The following discussion is based on material presented in Marin (1990). 5. Whether the countries have, in fact, been successful in enforcing destiny clauses in order to secure the additionality of exports cannot, however, be decided on the basis of the data. Some observers have expressed scepticism about the enforceability of supplemented trade and have therefore asserted that countertrade will de¯ate prices of traditional exports leading to a deterioration of the countertrading countries' terms of trade in the long run (Goldstein 1984). 6. The Western trading partners were asked to judge the quality of the goods they received relative to comparable Western goods. In 85 percent of the cases the quality was determined to be excellent or above average. Only 10 percent of the transactions was found to be bad or very bad. We have no information on quality in the remaining 5 percent. 7. In 1984, 10 to 20 percent of oil exports from OPEC countries are estimated to have been bartered (Banks 1985). 8. Guriev and Kvassov (2000) apply this price discrimination explanation to domestic barter in Russia. 9. For the reasons why better futures markets do not exist, see Newbery and Stiglitz (1985).
Chapter 3 1. This chapter draws on Marin and Schnitzer (2002a, b). 2. For a survey on the large literature on the sovereign debt problem, see Eaton and Fernandez (1995). 3. An extensive discussion of the legal aspects of these actions can be found in Bulow and Rogoff (1989b). 4. Period 2 is the ®rst period in which a payment is due. Since the problem is completely stationary, it suf®ces to show that it is not pro®table to deviate in this period. 5. If there is a supplier different from A with production cost smaller than c1 , B should have dealt with him in the ®rst place already. 6. This is exactly the structure of ``repayment schemes'' in case of foreign direct investment. When countries defaulted on their sovereign debt in the international debt crisis of the 1980s, foreign direct investments were sometimes recommended as an alternative to international credits because of the ¯exible payment schedules associated with the investment. 7. Appendix I-B describes the variables included in the following empirical analysis and reports on some sample statistics.
Chapter 4 1. This chapter uses material presented in Marin and Schnitzer (1998, 2002a). 2. In standard oligopoly models one should expect higher markups in differentiated goods as compared to homogeneous goods because of their higher market power. Our
Notes to Pages 62±80
187
point is that for a given level of market power, differentiated goods are subject to more asymmetric information problems and hence more quality problems. Note that the asymmetry and quality problems arise not only immediately for the buyer herself but even more so when she tries to resell the goods on the world market. Since producers of differentiated goods in Eastern Europe and developing countries do not have a reputation for producing high quality yet, these goods are dif®cult to market and can be sold only at a discount price compared to the same type of goods produced by Western producers. 3. Casella and Rauch (1998) introduce information-sharing networks as additional determinants of trade in differentiated goods and as a mechanism to overcome incomplete information in international markets. Our observation that the use of homogeneous goods as collateral goods has advantages because of their greater liquidity, as well as disadvantages because of their greater anonymity, is similar to a recent argument given by Myers and Rajan (1998) on the paradox of liquidity in the context of external ®nancing of ®rms. On the one hand, it is easier to raise capital if the ®rm can offer more liquid assets as collateral. On the other hand, greater asset liquidity reduces the ®rm's ability to commit to a speci®c course of action. 4. Appendix I-B describes the variables included in the following empirical analysis and reports on some sample statistics. 5. That basic goods seem to be more anonymous than investment goods is reasonable because the former are often ¯uid goods (e.g., oil) for which property rights are harder to de®ne than for bulky products (e.g., a machine). 6. We use the country variable DEBT as a proxy for the creditworthiness because we do not formally distinguish between the creditworthiness of the country and of the importing ®rm. This is sensible because in formerly centrally planned economies and in many developing countries, the importing side is a state-®rm or a state trade organization. Thus the creditworthiness of the importing ®rm and the state cannot really be disentangled. Furthermore whether or not a ®rm is creditworthy depends crucially on the willingness of state authorities to enforce credit repayment. The more indebted is the state, however, the less it is in the interest of the state to provide this service to foreign creditors. Thus the more indebted is the state as a whole, the less creditworthy are ®rms within the state. 7. Caves and Marin (1992) ®nd evidence that barter allows Western exporters to price discriminate. However, this does not lead to a clear prediction of which type of barter goods are used for the transaction.
Chapter 5 1. This chapter borrows from Marin and Schnitzer (1995). 2. Similarly the seller of a production plant may have superior information regarding the quality and/or future value of the commodity to be produced with the plant, which leads to a problem of adverse selection. As we mentioned in chapter 2, the negative implications of both incentive problems on international technology transfer have been emphasized by several authors; see Mirus and Yeung (1986), Kogut (1986), Hennart (1989), and Abdallah-Khodja (1984). 3. All of our results can be generalized to the case of a continuous quality variable; see note 9 below.
188
Notes to Pages 80±89
4. We rule out the possibility that the level of quality can be deduced from A's production cost or B's pro®ts. Both parties can shift accounting costs and pro®ts between different divisions and activities, so it becomes impossible for the courts to infer the quality from accounting data. 5. These are important considerations for ®rms in Eastern European or developing countries. See, for example, Hennart (1989). 6. A similar theoretical effect can be found in Bolton and Scharfstein (1990). They show that a ®rm will not default on its debt if the availability of a second loan is made contingent on the repayment of the ®rst one. 7. In the corporate debt literature there is an interesting parallel to our use of a ®nancial constraint as an incentive device. Hart (1991) considers a manager whose action has an externality on outside investors. The manager might not want to liquidate the ®rm when it would be pro®table to do so. The optimal debt repayment schedule is designed so that the manager faces a ®nancial constraint and is forced to come to the capital market for additional funds. This way he can be induced to liquidate the ®rm when it is in the interest of outside investors to do so. 8. Rewriting condition (5.3) as p2 > p1 ÿ a highlights the similarity with the condition for debt repayment in the sovereign debt literature. The debtor will meet his debt service obligations only if he expects to receive a subsequent net cash transfer from creditors if he repays. See Eaton (1993). 9. Note that b cannot be used as collateral because it is an unveri®able spillover bene®t that cannot be seized by potential creditors. 10. If quality is a continuous variable, then (5.5) and (5.6) have to be replaced by a condition that B can ®nance c2 if and only if ef®cient quality is delivered. 11. A strictly positive b is consistent with the export pattern that we observe for countertrade. As we showed in chapter 2, 67 percent of all countertrade exports in our sample from Eastern European or developing countries are manufactured investment and consumption goods as opposed to only 35 to 40 percent in total trade. As argued above, marketing these goods through a Western ®rm opens new markets for B and generates a positive spillover on his future exports to Western markets. In fact, as we showed in chapter 4, investment and consumption goods are more likely to be chosen as export goods if the ratio of technology imports is large. 12. Appendix I-B describes the variables included in our following empirical analysis and reports on some sample statistics. 13. Recall that our theory requires that B face a ®nancial constraint when A delivers low quality. This is the case only when the deal is suf®ciently large. A back-of-theenvelope calculation suggests that the deals have been large enough to make the foreign trade organisations (FTO) in Eastern Europe face this constraint. In 1985 the average countertrade technology imports to former CSFR in our sample made up 26 percent of the trade value of a typical FTO. Similar calculations for Hungary and Poland make the typical countertrade technology import 18 percent and 9 percent, respectively, of the average trade value of the typical FTO. 14. The mean compensation ratio of counterpurchase contracts in our sample is 61.2 (with a variance of 45.31) as opposed to 107.4 for barter (variance 28.77). An analysis of variance (ANOVA) and a Kolmogorov-Smirnov test both con®rm that the distributions of compensation ratios for barter and counterpurchase differ signi®cantly.
Notes to Pages 91±123
189
15. The reader might ask whether this negative impact of technology imports on the compensation ratio is simply due to the fact that technology imports like factories are typically very large and that there may not be enough valuable exports to offset these imports. To check this possibility, we calculated the correlation coef®cient between the import values and the compensation ratios. The result ÿ0:03 indicated that the two variables are not correlated. This con®rms that the negative coef®cient in our regression is not just driven by size effects.
Chapter 6 1. This chapter draws on Marin, Kaufmann, and Gorochowskij (2000) and on Marin and Schnitzer (1999). 2. See ®gure 1.1. The percentage estimates of barter of various data sources differ because of differences in de®nitions and methodologies. For an illustration of this compare table 6.1 with table 1.1. 3. See also Recanatini and Ryterman (2000) who document the positive role of business associations for enterprise performance in Russia. 4. In chapter 9 we empirically test the role of reputation for contract enforcement, and we ®nd that reputational concerns have no signi®cant effect on the terms of barter transactions negotiated among the parties we surveyed. 5. For a similar view, see Ickes and Ryterman (1992, 1993). For arrears as a way for ®rms to behave opportunistically, see Perotti (1995). 6. See European Bank for Reconstruction and Development, Transition Report 1997, pp. 26±27. 7. The possibility of hiding in barter might make it an instrument for the unof®cial economy and corruption. We explore the connection between barter, the unof®cial economy, and corruption in more detail across transition economies in Kaufmann and Marin (2000). For the role of the unof®cial economy to cope with the transition, see Johnson, Kaufmann, and Shleifer (1997). 8. This section draws on Marin (2002). 9. We obtained this information from the following question: ``What is the percentage price difference between the price you charge/you are charged for this particular good in this barter deal as compared to the typical price you charge/you are charged for the same product in cash deals?'' 10. The average percent price differential between barter and cash appears to be low from table 6.8. These averages hide the actual variation in the price differentials, because in 45 percent of the deals the non-cash and cash prices were equal. For the distribution of the price differentials, see table 6.7. 11. Moreover in table 6.5 we show that the ®rm's barter exposure does not increase for less ef®cient ®rms. 12. The fact that the terms of trade is shifting in 45 percent of the deals toward the ``sale'' side of the barter deal casts further doubts on the explanation that barter is driven by tax motives. As table 6.7 shows, it is only in 10.5 percent of the deals in which ®rms could potentially hide some of their pro®ts lowering their tax base. This
190
Notes to Pages 123±144
number corresponds roughly to the number of cases in which the interviewed ®rms gave taxes some importance for undertaking a barter deal; see table 6.10. 13. For similar results, see Commander and Mumssen (1998) and Russian Economic Barometer (1997) for Russia and Transition Report (1997).
Chapter 7 1. This chapter draws on Marin, Kaufmann, and Gorochowskij (2000). 2. In chapter 4 we provide empirical evidence for the better credit enforcement properties of goods as payment in international barter. 3. See also Begg and Portes (1993). 4. For a theoretical discussion of ®nancing transition economies in Eastern Europe, see HolmstroÈm (1996). 5. Similarly the study by Hendley, Murrell, and Ryterman (2000) on managers' perception of law enforcement in Russia suggests that legal institutions do work to some extend in these countries.
Chapter 8 1. This chapter draws on Marin, Kaufmann, and Gorochowskij (2000) and Marin and Schnitzer (1999). 2. Williamson (1975) refers to situations or relationships as speci®c when there is little outside opportunity and thus switching business partners is very costly. 3. In their study on contract enforcement in transition countries, Johnson, McMillan, and Woodruff (1999) measure how easily customers can ®nd alternative suppliers by the number of competitors located near (within 1 km of) the interviewed manufacturer. Most manufacturers (79 percent) reported no manufacturers of similar products located nearby, and on average there is less than one ®rm located within 1 km of manufactures. This is seen as evidence that the search costs for customers are nontrivial. In chapter 10 we provide empirical evidence from our data set that supports the assumption that the holdup problem is on the buyer's side. 4. In chapter 4 we show that in international barter a payment in goods has better credit enforcement properties than cash because it is a less anonymous means of payment. 5. Hendley, Murrell, and Ryterman (2000) report from their survey of contract enforcement in Russia that 42 percent of purchases involved barter because of the selfenforcing nature of this type of contract. 6. This allows us to solve the bargaining problem in each step of the production by simple analogy. It is straightforward to extend our analysis to enforcement costs that are increasing in the payment to be enforced. In this case the problem of creditworthiness becomes more severe at later production steps, which makes it more dif®cult to guarantee S1 's participation in the deal earlier on. 7. Bj might be involved in other production chains with revenues vj0 . But we assume that Sj has no knowledge about Bj 's revenues outside this particular production chain
Notes to Pages 144±159
191
and thus cannot use those revenues to enforce payment. Note, however, that this is without loss of generality. Allowing for higher enforceable payments leads only to level effects but does not affect the qualitative results of our model, as long as the maximum enforceable payment is ®nite. 8. If B1 could deliver the barter good right away, he would not be liquidity constrained because he could use the barter good as payment in kind. 9. See chapter 4 where we discuss the property of anonymity of barter goods in the context of a theory of money. Note that the mechanism by which a hostage is created here differs from the one described in chapter 5. In the present chapter a hostage is created to control the credit enforcement problem when the buyer agrees to repay the loan in goods rather than money. In chapter 5 a hostage is created to control the technology transfer problem when the technology buyer in the developing country has not enough cash in her pocket, and thus is unable to produce the good when the seller in the industrial country offers inferior technology.
Chapter 9 1. This chapter draws on Marin and Schnitzer (1999). 2. Note that this empirical prediction also holds if barter does not involve a credit relationship but if the goods used as payment are available right away. 3. Note that Konings and Walsh (1999) instead ®nd that disorganization did not constrain employment and productivity growth in newly established private ®rms in Ukraine. 4. Ickes and Ryterman (1993) also see arrears in Russia as a response to a liquidity shortage in the economy. 5. In their survey on contract enforcement in Eastern Europe and the former Soviet Union, Johnson, McMillan, and Woodruff (1999) ®nd that the higher are the search costs of customers, the more likely they are to be offered trade credits by their manufacturers. Customers with high search costs, so they argue, are reliable creditors because they are locked in the relationship with the input supplier. We ®nd the causality to be quite reverse. The fact that customers are given trade credits gives them the necessary counterbalancing power to deal with the holdup problem they are exposed to due to high costs of ®nding an alternative supplier. 6. Calvo and Coricelli run a similar regression between output and credit for Poland. They ®nd a point estimate between 0.2 and 0.6 depending on speci®cation, suggesting that a 10 percent contraction of credit results in an output decline between 2 and 6 percent. Note further that Blanchard and Kremer report evidence based on a survey among 500 ®rms in Russia that suggests that the ®nancial constraint was the most important shortage experienced by enterprises (see their table IV). Between 1993 and 1995 over 60 percent of the ®rms experienced a shortage of ®nancial resources compared with only over 20 percent of the ®rms experiencing shortages of material. 7. For a description of how we obtained this price information, see note 6 to chapter 6. 8. These empirical ®ndings justify the assumption we made in chapter 8 that the holdup problem arises on the buyer's rather than on the seller's side. If Blanchard and Kremer's formulation of the holdup problem on the seller's side were valid, the input
192
Notes to Pages 159±184
price would be lower rather than higher in barter as compared to cash deals. Thus in this case we would have expected a negative rather than a positive coef®cient on the complexity index. 9. For the concept of liquidity in an incentive theory of money, see Banerjee and Maskin (1996). 10. The model actually predicts that in equilibrium the holdup problem is solved and thus input prices will not be in¯ated. A look at table 6.7 reveals that in 73.6 percent of the deals this was indeed the case.
Chapter 10 1. See Schnitzer (2000, 1999c). 2. For a recent analysis of foreign direct investment in transition economies, see Lankes and Venables (1996). 3. For the development of banking and ®nancial markets in transition economies, see Anderson, BergloÈf, and Miszei (1996), Schnitzer (1999a, b), Transition Report (1998), Warner (1998), Anderson and Kegels (1998), and Calvo and Frenkel (1991). 4. See Transition Report (1999). 5. For the exact de®nition of these indexes, see EBRD, Transition Report (1999). 6. Ideally we would want a measure for whether the currency is overvalued or undervalued. This would have required an estimate of PPP (purchasing power parity) exchange rates for these countries. The exchange rates used in the regressions should be considered as a crude measure for this effect. 7. For the argument why partial reform might make things worse in an overall distorted economy, see Murphy, Shleifer, and Vishny (1992). 8. Shleifer and Treisman (2000) argue that the grand bargain between the banking sector and the government prevented the government de®cit from being ®nanced by printing money and thus avoided hyperin¯ation in Russia.
References
Abdallah-Khodja, Kemal. 1984. Algeria's experience with new international investment relations. In Charles Oman, ed., New Forms of International Investment in Developing Countries: The National Perspective, Paris: OECD. Abreu, Dilip. 1988. Towards a theory of discounted repeated games. Econometrica 56: 383±96. Amann, Erwin, and Dalia Marin. 1994. Risk-sharing in international trade: An analysis of countertrade. Journal of Industrial Economics 42: 63±77. Anderson, Ronald W., and Chantal Kegels. 1998. Transition Banking: Financial Development of Central and Eastern Europe. Oxford: Oxford University Press. Anderson, Ronald W., Erik BergloÈf, and Kalman Mizsei. 1996. Banking Sector Development in Central and Eastern Europe. Forum Report of the Economic Policy Initiative, No. 1. London: CEPR. Aslund, Anders. 1995. How Russia Became a Market Economy. Washington, DC: Brookings Institution. Balcerowicz, Leszek. 1994. The transition to the market economy: The Polish case, 1989±93. Economic Policy 9: 71±97. Banerjee, Abijit, and Eric Maskin. 1996. A Walrasian theory of money and barter. Quarterly Journal of Economics 111: 955±1005. Banks, Gary. 1983. The economics and politics of countertrade. World Economy 6: 159± 82. Banks, Gary. 1985. Constrained markets, ``surplus'' commodities and international barter. Kyklos 38: 249±67. Barkas, James M. 1987. A bank's role in facilitating and ®nancing countertrade. In Christopher M. Korth, ed., International Countertrade. New York: Quorum Books. Begg, David, and Richard Portes. 1993. Enterprise debt and economic transformation: Financial restructuring in Central and Eastern Europe. In Colin Mayer and Xavier Vives, eds., Capital Markets and Financial Intermediation. Cambridge: Cambridge University Press. Blanchard, Olivier, and Michael Kremer. 1997. Disorganization. Quarterly Journal of Economics 112: 1091±1126.
194
References
Bolton, Patrick, and David Scharfstein. 1990. A theory of predation based on agency problems in ®nancial contracting. American Economic Review 80: 94±106. Borenzstein, Eduardo, Andrew Berg, Ratna Sahay, and Jeromin Zettelmeyer. 1999. The evolution of output in transition economies: Explaining the differences. Mimeo. International Monetary Fund, Washington. Bulow, Jeremy, and Kenneth S. Rogoff. 1988. The buyback boondoggle. Brookings Papers on Economic Activity 2: 675±98. Bulow, Jeremy, and Kenneth S. Rogoff. 1989a. Sovereign debt: Is to forgive to forget? American Economic Review 79: 43±50. Bulow, Jeremy, and Kenneth S. Rogoff. 1989b. A constant recontracting model of sovereign debt. Journal of Political Economy 97: 155±78. Bussard, Willis A. 1987. An overview of countertrade practices of corporations and individual nations. In Christopher M. Korth, ed., International Countertrade. New York: Quorum Books. Business International Corporation, eds. 1984. Threats and Opportunities of Global Countertrade. New York: Business International Corporation. Calvo, Guillermo A., and Fabrizio Coricelli. 1995a. Inter-enterprise arrears in economies in transition. In Robert Holzmann, JaÂnos GaÂcs, Georg Winckler, eds., Output Decline in Eastern Europe. Dordrecht: Kluwer Academic. Calvo, Guillermo A., and Fabrizio Coricelli. 1995b. Output collapse in Eastern Europe: The role of credit. In Mario I. Blejer, Guillermo Calvo, Fabrizio Coricelli, and Alan Gelb, eds., Eastern Europe in Transition: From Recession to Growth? Washington, DC: World Bank. Calvo, Guillermo A., and Jacob A. Frenkel. 1991. Credit markets, credibility, and economic transformation. Journal of Economic Perspectives 5: 139±48. Casella, Alessandra, and James E. Rauch. 1998. Overcoming informational barriers to international resource allocation: Prices and group ties. NBER Working Paper 6628. National Bureau of Economic Research, Cambridge, MA. Casson, Mark, and Francis Chukujama. 1990. Countertrade: Theory and evidence. In Peter J. Buckley and Jeremy Clegg, eds., Multinational Enterprises in Less Developed Countries. London: Macmillan. Caves, Richard E. 1974. The economics of reciprocity: Theory and evidence on bilateral trading arrangements. In Willy Sellekaerts, ed., International Trade and Finance: Essays in Honour of Jan Tinbergen. London: Macmillan, pp. 17±54. Caves, Richard E., and Dalia Marin. 1992. Countertrade transactions: Theory and evidence. Economic Journal 102: 1171±83. Chan, Raissa, and Michael Hoy. 1991. East±West joint ventures and buyback contracts. Journal of International Economics 30: 331±43. Choi, Chong J., and Daniel Maldoom. 1992. A simple model of buybacks. Economic Letters 40: 77±82. Clement, Cynthia, and Peter Murrell. 2000. Assessing the value of law in transition economies: An introduction. In Peter Murrell, ed., Assessing the Value of Law in Transition Economies. Ann Arbor: University of Michigan Press, forthcoming.
References
195
Commander, Simon, and Christian Mumssen. 1998. Understanding barter in Russia. Mimeo. European Bank for Reconstruction and Development, London. Eaton, Jonathan. 1991. Debt relief and the international enforcement of loan contracts. Journal of Economic Perspectives 4: 43±56. Eaton, Jonathan. 1993. Sovereign debt: A primer. World Bank Economic Review 7: 137±72. Eaton, Jonathan, and Raquel Fernandez. 1995. Sovereign debt. In Gene M. Grossman and Kenneth S. Rogoff, eds., Handbook of International Economics. Amsterdam: North Holland, pp. 2031±77. Eaton, Jonathan, and Mark Gersovitz. 1981. Debt with potential repudiation: Theoretical and empirical analysis. Review of Economic Studies 48: 289±309. Ellingsen, Tore. 1998. Payments in kind. Mimeo. Stockholm. Ellingsen, Tore, and Lars A. Stole. 1996. Mandated countertrade as a strategic commitment. Journal of International Economics 40: 67±84. Fantapie Altobelli, Claudia. 1994. KompensationsgeschaÈfte im internationalen Marketing, Heidelberg: Physica-Verlag. Fischer, Heiko. 1996. Countertrade im Osthandel. Wiesbaden: Gabler, Deutscher UniversitaÈts Verlag. Fischer, Stanley, and Ratna Sahay. 2000. Taking stock. Finance and Development 37: 2±6. Gaddy, Clifford G., and Barry W. Ickes. 1998. Russia's virtual economy. Foreign Affairs 77: 53±67. Gaidar, Yegor. 1995. Russian reform. In Yegor Gaidar and Karl Otto PoÈ hl, eds., Russian Reform/International Money, Cambridge: MIT Press. Goldstein, Elizabeth Ann. 1984. A theoretical model of countertrade. Research Paper 8404. Federal Reserve Bank of New York. Gooptu, Sudarshan, and Maria Soledad Martinez Peria. 1992. Factors that affect shortterm commercial bank lending to developing countries. Working Paper. World Bank, Washington, DC. Greif, Avner. 1992. Institutions and international trade: Lessons from the commercial revolution. American Economic Review 82: 128±33. Greif, Avner, and Eugene Kandel. 1994. Contract enforcement institutions: Historical perspective and current status in Russia. In Edward Lazear, ed., Economic Transition in Eastern Europe and Russia: Realities of Reform. Stanford: Hoover Institution Press. Greif, Avner, Paul Milgrom, and Barry R. Weingast. 1994. Coordination, commitment, and enforcement: The case of the merchant guild. Journal of Political Economy 102: 745± 76. Group of Thirty. 1985. Countertrade in the World Economy. New York. Guisinger, Stephen E. 1987. Investment related to trade. In J. Michael Finger and Andrzej Olechowski, eds., The Uruguay Round: A Handbook for the Multilateral Trade Negotiations. Washington DC: World Bank. Guriev, Sergei, and Dmitry Kvassov. 2000. Barter for price discrimination. CEPRDiscussion Paper 2449. London.
196
References
Hammond, Grant T. 1990. Countertrade, Offsets, and Barter in International Political Economy. London: Pinter. Hart, Oliver. 1991. Theories of optimal capital structure: A principal-agent perspective. Mimeo. Massachusetts Institute of Technology. Hart, Oliver. 1995. Firms, contracts, and ®nancial structure. Oxford: Oxford University Press. Hart, Oliver, and John Moore. 1995. Debt and seniority: An analysis of the role of hard claims in constraining management. American Economic Review 85: 567±85. Hay, Jonathan R., and Andrei Shleifer. 1998. Government in transition: Private enforcement of public laws; a theory of legal reform. American Economic Review 88: 398± 403. Hay, Jonathan R., Andrei Shleifer, and Robert W. Vishny. 1996. Toward a theory of legal reform. European Economic Review 44: 559±67. Hendley, Kathryn, Peter Murrell, and Randi Ryterman. 2000. Law works in Russia: The role of legal institutions in the transactions of Russian enterprises. In Peter Murrell, ed., Assessing the Value of Law in Transition Economies. Ann Arbor: University of Michigan Press, forthcoming. Hennart, Jean-Francois. 1989. The transaction-cost rationale for countertrade. Journal of Law, Economics, and Organization 5: 127±53. Hennart, Jean-Francois. 1990. Some empirical dimensions of countertrade. Journal of International Business Studies 21: 243±70. Hennart, Jean-Francois, and Erin Anderson. 1993. Countertrade and the minimization of transaction costs: An empirical examination. Journal of Law, Economics, and Organization 9: 290±313. HolmstroÈm, Bengt. 1996. Financing of investment in Eastern Europe: A theoretical perspective. Industrial and Corporate Change 2: 205±37. HolmstroÈm, Bengt, and Jean Tirole. 1989. The theory of the ®rm. In Richard Schmalensee and Robert D. Willig, eds., Handbook of Industrial Organization, vol. 1. Amsterdam: North Holland, pp. 61±133. Huang, Haizhou, Dalia Marin, and Chenggang Xu. 2001. Financial crisis, economic recovery, and banking development in Russia. Mimeo. International Monetary Fund. University of Munich, and London School of Economics. Ickes, Barry W., and Randi Ryterman. 1992. The inter-enterprise arrears crisis in Russia. Post-Soviet Affairs 8: 331±61. Ickes, Barry W., and Randi Ryterman. 1993. Roadblock to economic reform: Interenterprise debt and the transition to markets. Post-Soviet Affairs 9: 231±52. International Monetary Fund. Annual Report on Exchange Arrangements and Exchange Restrictions. Washington DC: IMF, various years. International Monetary Fund. 1999. Russian Federation: Recent economic developments. IMF Staff Country Report 99/100. Washington, DC. Jalloh, Balimo. 1990. Countertrade: Kompensations- und OffsetgeschaÈ fte erfolgreich abwickeln. Landsberg: Verlag Moderne Industrie.
References
197
Johnson, Simon, Daniel Kaufmann, and Andrei Shleifer. 1997. The unof®cial economy in transition. Brookings Paper on Economic Activity, Fall (2): 159±239. Johnson, Simon, John McMillan, and Christopher Woodruff. 1999. Contract enforcement in transition. Paper presented at the 5th Nobel Symposium in Economics: The Economics of Transition. Kaufmann, Daniel, and Dalia Marin. 2000. The non-cash economy: Evidence from 20 transition economies. Mimeo. World Bank and University of Munich. Kiyotaki, Nobuhiro, and Randall Wright. 1993. A search-theoretic approach to monetary economics. American Economic Review 83: 63±77. Klaus, Vaclav. 1997. Renaissance: The Rebirth of Liberty in the Heart of Europe. Washington DC: Cato Institute. Kletzer, Kenneth M., and Brian Wright. 1990. Sovereign debt renegotiations in a consumption-smoothing model. Mimeo. University of California at Santa Cruz. Kogut, Bruce. 1986. On designing contracts to guarantee enforceability: Theory and evidence from East±West trade. Journal of International Business Studies 17: 47±61. Konings, Jozef, and Patrick Paul Walsh. 1999. Disorganization in the transition process: Firm-level evidence from Ukraine. Economics of Transition 7: 29±46. Kranton, Rachel E. 1996. Reciprocal exchange: A self-sustaining system. American Economic Review 86: 830±51. Krugman, Paul R. 1992. Currencies and Crises. Cambridge: MIT Press. Lankes, Hans-Peter, and Anthony J. Venables. 1996. Foreign direct investment in economic transition: The changing pattern of investments. Economics of Transition 4: 331± 47. Magenheim, Ellen, and Peter Murrell. 1988. How to haggle and to stay ®rm: Barter as hidden price discrimination. Economic Inquiry 26: 449±59. Marin, Dalia. 1990. Tying in international trade. World Economy 3: 445±62. Marin, Dalia. 2000. Monetary policy does not work in Russia's barter economyÐ German banking as a solution? Economic Systems 24: 87±90. Marin, Dalia. 2002. Trust versus illusion: What is driving demonetization in the former Soviet Union? Economics of Transition, forthcoming. Marin, Dalia, Kaufmann Daniel, and Bogdan Gorochowskij. 2000. Barter in transition economies: Competing explanations confront Ukrainian data. In Paul Seabright, ed., The Vanishing Rouble: Barter Networks and Non-monetary Transactions in Post-Soviet Societies. Cambridge: Cambridge University Press, pp. 207±35. Marin, Dalia, and Monika Schnitzer. 1995. Tying trade ¯ows: A theory of countertrade with evidence. American Economic Review 85: 1047±64. Marin, Dalia, and Monika Schnitzer. 1998. Economic incentives and international trade. European Economic Review 42: 705±16. Marin, Dalia, and Monika Schnitzer. 1999. Disorganization and ®nancial collapse. Paper presented at the 5th Nobel Symposium on Transition Economics, Stockholm, 1999, CEPR Discussion Paper 2245. Center for Economic Policy Research, London.
198
References
Marin, Dalia, and Monika Schnitzer. 2002a. The economic institution of international barter. The Economic Journal 112. Marin, Dalia, and Monika Schnitzer. 2002b. Creating creditworthiness through reciprocal trade. Review of International Economics 10. McMillan, John, and Christopher Woodruff. 2000. Private order under dysfunctional public order. Mimeo. Stanford University and University of California, San Diego. McVey, Thomas B. 1985. Overview of the commercial practice of countertrade. In Bart S. Fisher and Kathleen M. Harte, eds., Barter in the World Economy. New York: Praeger, pp. 9±36. Mirus, Rolf, and Bernard Yeung. 1986. Economic incentives for countertrade. Journal of International Business Studies 17: 27±39. Mirus, Rolf, and Bernard Yeung. 1987. Countertrade and foreign exchange shortages: A preliminary assessment. Weltwirtschaftliches Archiv 123: 535±44. Mirus, Rolf, and Bernard Yeung. 1993. Why countertrade? An economic perspective. International Trade Journal 7: 409±33. Mirus, Rolf, and Bernard Yeung, eds. 2001. The economics of barter and countertrade. Cheltenham, Northampton, UK: Edward Elgar. Mitchell, Janet. 1993. Creditor passivity and bankruptcy: Implications for economic reform. In Colin Mayer and Xavier Vives, eds., Capital Markets and Financial Intermediation. Cambridge: Cambridge University Press. Murrell, Peter. 1982. Product quality, market signaling and the development of East± West trade. Economic Inquiry 20: 589±603. Murrell, Peter, ed. 2000. Assessing the Value of Law in Transition Economies. Ann Arbor: University of Michigan Press, forthcoming. Murphy, Kevin M., Andrei Shleifer, and Robert W. Vishny. 1992. The transition to a market economy: Pitfalls of partial reform. Quarterly Journal of Economics 107: 889±906. Myers, Stewart C., and Raghuram G. Rajan. 1998. The paradox of liquidity. Quarterly Journal of Economics 113: 731±71. Newbery, David M. G., and Joseph E. Stiglitz. 1985. The Theory of Commodity Price Stabilization. Oxford: Clarendon Press. North, Douglas. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. Oman, Charles, ed. 1984. New Forms of International Investment in Developing Countries. Paris: OECD. Organisation for Economic Cooperation and Development. 1981. East±West Trade: Recent Developments in Countertrade. Paris: OECD. Organisation for Economic Cooperation and Development. 1985. Countertrade: Developing Country Practices. Paris: OECD. Organisation for Economic Cooperation and Development. 1997±98. OECD Economic Surveys: Russian Federation 1997. Paris: OECD.
References
199
Organisation for Economic Cooperation and Development. 1999±2000. OECD Economic Surveys: Russian Federation 2000. Paris: OECD. Parsons, John E. 1985. A theory of countertrade ®nancing of international business. Mimeo. Massachusetts Institute of Technology. Parsons, John E. 1987. Forms of GDR economic cooperation with the nonsocialist world. Comparative Economic Studies 29: 7±18. Perotti, Enrico C. 1995. Inertial credit and opportunistic arrears in transition. European Economic Review 42: 1703±25. Prendergast, Canice, and Lars Stole. 1996. Nonmonetary exchange within ®rms and industry. NBER Working Paper 5765. National Bureau Economic Research, Cambridge, MA. Prendergast, Canice, and Lars Stole. 1997. Barter, liquidity and market segmentation. Mimeo. University of Chicago. Prendergast, Canice, and Lars Stole. 1999. Restricting the means of exchange within organizations. European Economic Review 43: 1007±19. Recanatini, Francesca, and Randi Ryterman. 2000. Disorganization or self-organization. Mimeo. World Bank. Roland, GeÂrard, and Thierry Verdier. 1999. Transition and the output fall. Economics of Transition 7: 1±28. Rostowski, Jacek. 1993. The inter-enterprise debt explosion in the former Soviet Union: Causes and consequences, cures. ESR Discussion Paper 142. Center for Economic Performance, London. Russian Economic Barometer. 1999. Survey of Industrial Firms. Institute for World Economy and International Relations. Moscow. Sachs, Jeffrey D. 1989. Developing Country Debt and the World Economy. Chicago: University of Chicago Press. Schnitzer, Monika. 1999a. Enterprise restructuring and bank competition in transition economies. Economics of Transition 7: 133±55. Schnitzer, Monika. 1999b. On the role of bank competition for corporate ®nance and corporate control in transition economies. Journal of Institutional and Theoretical Economics 155: 22±46. Schnitzer, Monika. 1999c. Expropriation and control rights: A dynamic model of foreign direct investment. International Journal of Industrial Organization 17: 1113±37. Schnitzer, Monika. 2001. Debt versus foreign direct investment: The impact of sovereign risk on the structure of capital ¯ows to developing countries. Economica 69. Seabright, Paul, ed. 2000. The Vanishing Rouble: Barter Networks and Non-monetary Transactions in Post-Soviet Societies. Cambridge: Cambridge University Press. Shleifer, Andrei, and Daniel Treisman. 2000. Without a Map: Political Tactics and Economic Reform in Russia. Cambridge: MIT Press. Tirole, Jean. 1999. Incomplete contracts: Where do we stand? Econometrica 67: 741±81.
200
References
Transition Report. 1997. European Bank for Reconstruction and Development, London. Transition Report. 1998. European Bank for Reconstruction and Development, London. Transition Report. 1999. European Bank for Reconstruction and Development, London. UNCITRAL. 1993. Legal Guide on International Countertrade Transactions. Vienna: United Nations Commission on International Trade Law. Verzariu, Pompiliu. 1985. Countertrade, Barter and Offets: New Strategies for Pro®t in International Trade. New York: McGraw-Hill. Warner, Andrew M. 1998. The emerging Russian banking system. Economics of Transition 6: 333±47. Welt, Leo G. B. 1984. Trade without Money: Barter and Countertrade. New York: Law and Business/Harcourt Bracl Jovanovich. Welter, Reinhard. 1989. Collateral in international trade. In Norbert Horn, ed., The Law of International Trade Finance. Boston: Kluwer Law and Taxation Publishers. Williamson, Oliver. 1975. Markets and Hierarchies: Analysis and Antitrust Implications. New York: Free Press. Williamson, Oliver. 1983. Credible commitments: Using hostages to support exchange. American Economic Review 73: 519±40. Williamson, Oliver. 1985. The Economic Institution of Capitalism. New York: Free Press. Wood, Philip R. 1995a. Comparative Financial Law. Law and Practice of International Finance. London: Sweet and Maxwell. Wood, Philip R. 1995b. Comparative Law of Security and Guarantees. Law and Practice of International Finance. London: Sweet and Maxwell. World Bank. 1994. World Bank Debt Tables. Washington, DC. Yarbrough, Beth V., and Robert M. Yarbrough. 1992. Cooperation and Governance in International Trade. Princeton: Princeton University Press.
Index
Abdallah-Khodja, K., 29, 187 Abreu, D., 46 Africa, 18 After-sale service. See Technology, technological updates Albania, 18 Algeria, 29 Amann, E., 32, 72 Anderson, E., 19 Anderson, R. W., 192 Arrears, 11, 103±104, 112±14, 124, 127, 132±35, 155±56, 164 ®rm arrears, 7, 12, 103, 106, 108±10, 112, 124, 131, 135, 137, 151, 164 growth of, 103, 105 inter-®rm arrears (see Arrears, ®rm-arrears) tax arrears, 7, 103, 106, 122, 129, 133 wage arrears, 7, 103, 112, 133 Asia, 19 Aslund, A., 109 Asymmetry of information, 61±62, 64±65 Balcerowicz, L., 9 Banerjee, A., 61±62, 192 Bank, 127 bank credit, 131 bank debt, 109±10, 157 banking failure, 164, 179 banking sector, 13, 179 Bankrupty bankrupty law, 110±11 bankrupty procedure, 110±11, 127 Banks, G., 22, 25 Bargaining bargaining power, 12, 25, 114, 140±41, 145±46, 153 Nash bargaining, 144, 149, 151
Barkas, J. M., 38 Barter, 2, 5, 17, 20±21, 35, 47, 88, 105, 111, 113±14, 124, 127, 130, 133±35, 137, 141, 151, 154±55, 163±64, 181 advantages of barter, 49±51, 127 barter deal ``goods payment,'' 106, 114±15, 119, 122, 163 ``sale,'' 106, 114, 119, 122, 163 barter goods (see also Money) anonymity of barter goods, 10, 41, 50, 61±63, 161 choice of barter goods, 68 (see also Trade pattern) liquidity of barter goods, 10, 62±63, 68, 161 quality of export goods, 10±11, 18, 25± 26 barter transactions repurchase value, 20 (see also Compensation ratio) commodity composition of barter, 106± 107 growth of barter, 105 motives for barter, 123 disorganization (see Disorganization) export promotion, 23±24, 26 foreign exchange shortage (see Foreign exchange constraint) lack of credit (see Credit, lack of credit) market distortion, 26, 28, 123 price discrimination (see Price discrimination) soft budget constraint (see Soft budget constraint) tax avoidance, 111, 122±23 virtual economy (see Virtual economy)
202
Barter (cont.) pattern of barter cross-country pattern, 4, 103, 137, 164 time pattern, 3, 103, 137, 164 Begg, D., 190 BergloÈf, E., 192 Blanchard, O., 108, 124±25, 131, 137, 139± 43, 147, 153, 155, 159, 191 Bolton, P., 188 Borenzstein, E., 103 Bulgaria, 18 Bulow, J., 6, 43, 50±51, 186 Business International, 5 Bussard, W. A., 5, 21±22 Buyback, 2, 17, 20±21, 29±31 Calvo, G. A., 108, 125, 131, 147, 157, 191± 92 Capital markets, 110 imperfect capital markets, 1, 8, 179, 182 capital transfer problem, 173 Cartel agreement, 26, 28 Casella, A., 187 Cash-in-advance contracts, 43, 44, 48 Casson, M., 19 Caves, R. E., 27, 187 Central Europe, 2, 11, 103, 105, 108, 164 Central planning, 2, 13, 114, 137±39, 175 Chain of production, 108, 139±40, 142 Chan, R., 30, 77 China, 19 Choi, C. J., 30, 77 Chukujama, F., 19 Clement, C., 9 Collateral, 38, 40±41, 50, 55, 135 (see also Hostage) deal-speci®c collateral, 10, 32±33, 36, 49, 54, 78, 92, 127, 131, 148, 153 value of collateral, 54, 58, 127 choice of collateral goods, 68, 70 (see also Trade pattern) Collusive agreement. See Cartel agreement Commander, S., 185, 190 Commitment, 12, 31, 141, 150±51 commitment value of barter, 150±51 Communism, 9 collapse of communism, 13, 18, 175 Comparative advantage. See International trade theory Compensation ratio, 20, 23, 27±28, 55, 88
Index
Complexity of production, 151, 156±57, 162 complexity index, 155, 159, 162 Contracts contract enforcement, xiii, 1, 6, 8, 10, 12, 173 contract theory, 8 contractual problems, 1, 5±8, 182, enforcement mechanism, 33, 54 explicit contracts, 7 implicit contracts, 45±46 incomplete contracts, 9, 108, 139 Coricelli, F., 108, 125, 131, 147, 157, 191 Counterpurchase, 2, 17, 20±21, 31, 77, 88 Countertrade forms of countertrade (see Barter; Counterpurchase; Buyback) compensation ratio (see Compensation ratio) destiny clauses, 23±24 duration of contract, 18, 57, 73 shopping list, 25 Credit credit constraint, 36, 40, 61, 139, 153, 158 credit enforcement, 133 credit enforcement cost, 140±41, 143±45, 148±49, 151, 154±55, 159 credit enforcement mechanism, 128 credit evaluation, 128 credit market, 110 credit problem, 12, 33, 127, 147±48, 150, 154±55, 158, 161, 163 lack of credit, 108, 135 Creditworthiness, 11, 32, 35, 48, 55, 74, 78, 89, 110±11, 127, 129±30, 133, 153, 159, 164 Croatia, 2, 4 Czech Republic, 4, 75, 175, 177 Czechoslovakia (former), 18, 28, 103 Deal-speci®c collateral. See Collateral Debt bank debt (see Bank, bank debt) debt capacity, 54 debt-GDP ratio, 19, 173±75 debt overhang problem, 43, 44 ®rm debt, 11, 109±11, 129 outstanding debt, 129 Debt crisis domestic debt crisis in transition economies, 1 (see also Financial crisis)
Index
203
international debt crisis, xiii, 1, 4, 7, 8, 17, 43 Demonetization, 115 Disorganization, 12, 108, 137, 139, 164, 183 Double coincidence of wants, 10, 33, 61± 62
Goldstein, E. A., 186 Gooptu, S., 185 Gorochowskij, B., 185, 189±90 Greif, A., 8, 185 Group of Thirty, 5 Guisinger, S. E., 185 Guriev, S., 186
East Asia, 175 East Germany (former), 18 East±West trade, 5, 18, 26, 178 Eaton, J., 36, 43, 186 Economic Intelligence Unit, 5 Ef®ciency of ®rms, 113±14 Ellingsen, T., 31±32 European Judgment Convention, 39 European Bank for Reconstruction and Development (EBRD), 180, 192 (see also Transition Report) ®nancial and commercial regulations, effectiveness, 180 Export pattern. See Trade pattern Export promotion. See Barter, motives for barter Export revenues export revenue risk (see Uncertainty) stochastic export revenues (see Uncertainty)
Hammond, G. T., 5 Hart, O., 9, 188 Harvard Institute of International Development, 106 Hay, J. R., 109, 185 Hendley, K., 109, 190 Hennart, J.-F., 19, 21, 29±30, 187±88 Holdup problem, 12, 108, 137±44, 147± 48, 154±55, 158±59, 161 HolmstroÈm, B., 9, 190 Hostage, 11, 34, 77, 78, 84±85, 92, 141, 148, 150±51, 164. See also Collateral Hoy, M., 30, 77, Huang, H., 179, 183±84 Hungary, 4, 18, 22, 103, 110, 173, 175, 177
Fantapie Altobelli, C., 185 Fernandez, R., 186 Financial constraint, 7±8, 12, 34, 36, 79, 92, 108, 129, 131±32, 135, 139, 141±42, 144±45, 154, 158, 164, 179, 183 Financial crisis in East Asia, 175 in Russia, xiii, 2, 13, 183±84 Financial system, 179±80 Fischer, H., 185 Fischer, S., 185 Foreign direct investment, 9, 13, 29, 30, 34, 177 Foreign exchange constraint, 22±23 Frenkel, J. A., 192 Futures market, 18, 32 Gaddy, C. G., 112 Gaidar, Y., 9 GATT, 5, 17 Gazprom, 112. See also Virtual economy, natural resource sector Gersovitz, M., 43
Ickes, B. W., 112, 189, 191 Ifo Institut, 5 Import pattern. See Trade pattern Incentive problems, 8 adverse selection, 62 credit problem (see Credit, credit problem) double moral hazard problem, 30, 80, 84 holdup problem (see Holdup problem) moral hazard, 48, 61±62, 79 sovereign debt, 32, 36 technology transfer problem (see Technology, technology transfer) trust problem (see Trust) Incomplete contracts, 9, 108, 139 Input markets, 182 Input shortage, 12, 109, 124, 139, 164 Institutions, 5, 10, 173, 182 economic institution, xiii, 9 institutional approach, 5±6 institutional reforms, 108 institutional response, 1, 8 institutional trap, 13, 180 legal institutions, 109 Insurance mechanism, 32, 73. See also Uncertainty International Monetary Fund (IMF), 5
204
International Standard of Industry Classi®cation (ISIC classi®cation), 155 International trade theory, 8±9, 69, 72 comparative advantage, 69, 72 economies of scale, 9 factor endowment, 9 trade pattern (see Trade pattern) Iran, 19 Jalloh, B., 185 Johnson, S., 185, 189±91 Joint ventures, 9, 30, 177 Kandel, E., 185 Kaufmann, D., 185, 189±90 Kazakhstan, 4 Kegels, C., 192 Kiyotaki, N., 61 Klaus, V., 9 Kletzer, K. M., 44 Kogut, B., 29, 187 Konings, J., 108, 191 Kranton, R. E., 179 Kremer, M., 108, 124±25, 131, 137, 139± 43, 147, 153, 155, 159, 191 Krugman, P. R., 44 Kvassov, D., 186 Kyrgyztan, 173 Lankes, H.-P., 192 Legal Guide on International Countertrade Transactions, 37 Legal system, 6, 41, 173, 179 legal recourse, 38, 50 legal institutions (see Institutions) Liquidity constraint. See Financial constraint Long-term relationships, 7, 139 Macroeconomic stabilization, 2, 9, 103 Magenheim, E., 27 Maldoom, D., 30, 77 Manufacturing sector. See Virtual economy Marin, D., 27, 32, 72, 86, 179, 183±87, 189±91 Market distortion. See Barter, motives for barter Market power, 27 monopoly power, 27 monopsony power, 25 purchasing power, 25
Index
Marketing expertise, 25 Maskin, E., 61±62, 192 McMillan, J., 109, 185, 190±91 McVey, T. B., 25 Milgrom, P., 8 Mirus, R., 23, 30, 187 Mitchell, J., 111 Mizsei, K., 192 Moldova, 4, 173 Money commodity money, 62 monetary policy, 183 ranking of commodity money, 63, 66 special purpose money, 41, 61 Mumssen, C., 185, 190 Murphy, K. M., 192 Murrell, P., 9, 26±27, 109, 185, 190 Myers, S. C., 187 Natural resource sector. See Virtual economy Newbery, D. M. G., 186 Non-cash economy, 115, 119, 163 Non-cash transactions, 118±19, 122 OECD, 5, 18, 21±23 Oman, C., 185 OPEC, 26 Opportunistic behavior, 108, 139 Output decline, 11, 103±105, 108±109, 124±25, 131, 134±35, 137, 139, 146, 151, 153±54, 156±57, 163±64, 176 Ownership of ®rms, 110 Parsons, J. E., 29±30, 33, 77 Payment payment in cash, 10, 33, 35, 37, 61 payment in goods, 10, 33, 35, 10, 35, 61 Perotti, E. C., 189 Poland, 4, 18, 19, 22, 103, 109, 173, 175, 177 Portes, R., 190 Prendergast, C., 31 Price discrimination, 26, 27, 29 Price distortion, 27 Productivity of the ®rm, 112 Property rights, 36, 1, 44, 47, 50, 128 Rajan, R. G., 187 Rauch, J. E., 187 Recanatini, F., 189 Reciprocity, 5
Index
reciprocal enforcement convention, 39 reciprocal form of exchange, 17 Renegotiation, 81, 82, 85, 86, 141 Repudiation, 4, 43 (see also Debt, default on debt) Reputation, xiii, 6, 52, 54, 56, 139, 173 Resale restriction, 24, 25 (see also Countertrade, destiny clause) Restructuring, 111±12, 123 Retention of title, 39 Revolution, xiii Rogoff, K. S., 6, 43, 50±51, 186 Roland, G., 108 Romania, 18, 109 Rostowski, J., 131 Russia, 2±3, 10, 13, 103±105, 109, 114, 119, 122, 131, 155, 173, 183 Russian Economic Barometer, 2, 90 Ryterman R., 109, 189±91 Sachs, J. D., 185 Sahay, R., 185 Scharfstein, D., 188 Schnitzer, M., 86, 186±87, 189±92 Security. See Collateral Seniority rights, 43±44, 49 Shleifer, A., 109, 185, 189, 192 Slovakia, 109 Soft budget constraint, 109, 110±11, 131, 179, 184 Soledad Martinez Peria, M., 185 South America, 19 Sovereign debt problem, 32, 36 Soviet Union (former), 7±8, 10, 18, 75, 103, 105, 109, 137, 140, 143, 156±57, 164, 173, 175, 177, 179 Speci®c relationship, 108 business relationship, 138, 140, 143 business ties, 137±38 lock-in, 138 relationship speci®c investment, 64, 108, 139±40, 142, speci®city, 108, 138±39, 141, 151 State-owned ®rms, 18, 109, 111, 131, 162 Stiglitz, J. E., 186 Stole, L., 31 Taxes tax arrears (see Arrears, tax arrears) tax avoidance (see Barter, motives for barter)
205
Technology lemons technology, 78 technological relationship between import and export goods, 17, 31, 77, 83 technological updates, 26, 29, 77 technology transfer 7, 8, 29, 30, 31, 33, 34, 73, 77±78, 79, 89, 173, 175, 182 Terms of trade, 28, 115, 118±19, 122, 158± 63 Tirole, J., 9 Trade credit, 128±29, 131, 135, 140±41 Trade pattern, 2, 21, 68, 74 export pattern, 10, 21±22, 33, 61±62 import pattern, 11, 21±22, 33, 77 Trade sanctions, 43, 56 Transaction costs, xiii, 30, 61 Transition, xiv initial conditions of transition countries, 108 structural reforms, 108 transition reforms, 9 Transition Report, 108, 189±90, 192 Treisman, D., 192 Trust lack of trust, 11±12, 109, 163, 179 Turnkey contracts. See Buyback Tying of trade ¯ows, 15, 20, 22, 33, 78, 91 Ukraine, 2, 4, 8, 105±106, 109±10, 125, 133, 135, 137, 153 Uncertainty, 51 export revenue uncertainty, 32, 51, 54, 56, 73 stochastic export revenues, 32, 51, 54, 73 UNCTAD, 5 United Nations Commission on International Trade Law (UNCITRAL), 37 Uzbekistan, 4 Venables, A. J., 192 Verdier, T., 108 Verzariu, P., 38 Virtual economy, 111±12, 118±23, 163 cross-subsidization, 114 manufacturing sector, 114, 118, 122 natural resource sector, 112, 114, 118± 19, 122 pricing behavior of sectors, 116±17 Vishny, R. W., 109, 192
206
Walsh, P. P., 108, 191 Warner, A. M., 192 Weingast, B. R., 8 Welt, L. G. B., 22, 38 Welter, R., 38 Williamson, O., 34, 78, 190 Wood, P. R., 38±39 Woodruff, C., 109, 185, 190±91 World Bank (see World Debt Tables World Business Environment Survey, 2, 185 World Debt Tables, 57, 69, 185 Wright, B., 44 Wright, R., 61 Xu, C., 179, 183±84 Yarbrough, B. V., 185 Yarbrough, R. M., 185 Yeung, B., 23, 30, 187 Yugoslavia (former), 18 Zambia, 19
Index