U* A W O R L D B A N K P O L I C Y R E S E A R C H R E P O R T PRIVATE CAPITAL FLOWS TO DEVELOPING COUNTRIES T,HE ROAD TO FINANCIAL INTECRAT+ISII I Private Capital Flows to Developing Countries A World Bank Policy Research Report Private Capital Flows to Deve oping Countries The Road to Financial Integration Published for the World Bank OXFORD UNIVERSITY PRESS Oxford University Press OXFORD NEW YORK TORONTO DELHI BOMBAY CALCUTTA MADRAS KARACHI KUALA LUMPUR SINGAPORE HONG KONG TOKYO NAIROBI DAR ES SALAAM CAPE TOWN MELBOURNE AUCKLAND and associated companies in BERLIN IBADAN C 1997 The International Bankfor Reconstruction and Development! THE wORLD BANK 1818 H Street, N W Washington, D.C 20433, US.A. Publzshed by Oxford University Press, Inc. 198 Madlison Avenue, New York, N. Y 10016 Oxford is a registered trademark of Oxford University Press. All rights reserved No part of thts publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, record- ing, or otherwise, without the priorpermission of Oxford University Press. Manufactured in the United States ofAmerica FirstprintingApril 1997 Cover photographs. clockwise, from top right Josef Po/leross/The Stock Market; Nadia Mackenzie/Tony Stone Worldwide; Julian Calder/Tony Stone Worldwide, Jane Evelyn Atwood/The Stock Market, D. StoeckleinIThe Stock Market The World Bank does notguarantee the accuracy ofthe data included in this publication and accepts no responsibility whatsoever for any consequence of their use. Library of Congress Cataloging-in-Publication Data Private capitalflows to developing countries . the road to financial integration P. cm Includes bibliographical references (p.) ISBN 0-19-521116-2 1. Investments, Foreign-Developing countries. 2. Capital movements-Developing countries. L World Bank. HG5993.P747 1997 332.67'3'091724-DC21 97-11785 CIP ISSN 1020-0851 e Textprinted onpaper that conforms to theAmerican National StandardsforPermanence of Paperfor Printed Library Materials. Z39.48-1984 Contents Foreword xi The Report Team xiii Acknowledgments xv Data Notes and Abbreviations xvii Summary 1 1 The Main Findings 9 A New Age of Global Capital 13 Benefits and Risks of Financial Integration 22 Policy Challenges and Emerging Lessons 32 Notes 72 2 The New International Environment 75 The Structural Character of New Private Capital Flows 80 The Structural Forces Driving Private Capital Flows to Developing Countries 85 The Outcome of Structural Changes: Growing Investment in Emerging Markets 104 The Prospects for Private Capital Flows 111 Volatility Arising from the International Environment 124 Conclusion 140 Annex 2.1. Key Deregulations, Financial Innovations, and Technological Advances 144 Notes 148 3 The Benefits of Financial Integration 153 Integration and Growth 153 Diversification Benefits 166 Conclusion 168 Notes 169 v eLQWS To DEVELOPING COUNTRIES 4 Challenges of Macroeconomic Management 171 Capital Inflows and Overheating: The Country Experience 174 Policy Regimes and Vulnerability 192 Macroeconomic Management with Growing Financial Integration 196 Lessons for Macroeconomic Management 214 Annex 4.1. Monetary and Fiscal Policies during Capital Inflow Episodes 219 Annex 4.2. Can Sterilization Be Effective? 223 Notes 226 5 The Effects of Integration on Domestic Financial Systems 229 Bank Lending and Macroeconomic Vulnerability 236 Bank Lending and Financial Sector Vulnerability 249 Financial Sector Reforms in the Context of Integration: Policy Lessons 257 Annex 5.1. Capital Inflows, Lending Booms, and Banking Crises: Country Episodes 287 Annex 5.2. Government Guarantees, High Real Interest Rates, and Banking Sector Fragility 289 Annex 5.3. Risk-Adjusted Capital-Asset Ratios: The BIS Classification of Risky Assets 293 Annex 5.4. Crisis Management in a Constrained Setting: The Case of Currency Boards 296 Notes 298 6 Preparing Capital Markets for Financial Integration 303 The Main Issues 304 Capital Markets and Financial Integration 313 Improving Market Infrastructure 323 The Regulatory Challenges of Financial Integration 340 Summary and Conclusions 368 Annex 6.1. G-30 Recommendations and Suggested ISSA Revisions 376 Annex 6.2. Market Microstructures 378 Annex 6.3. Construction of the Emerging Markets Index 382 Notes 385 Bibliography 391 Boxes 1.1 Measuring Financial Integration 17 vi '?~~~~M 1.2 What Is Constraining Foreign Capital Inflows to Sub-Saharan Africa? 34 1.3 Best Practice in Market Infrastructure 56 1.4 Principles of the Self-Regulatory Model 58 1.5 International Regulatory Cooperation and Coordination 64 2.1 Are Private Capital Flows a Cyclical and Temporary Phenomenon? The Early Answers 81 2.2 Are the New Private Capital Flows Cyclical or Structural? Recent Empirical Evidence 82 2.3 Pension Funds in Emerging Markets 108 2.4 Pension Reforms and Their Implications for Investments in Emerging Markets 118 2.5 Defined Contribution Pension Plans and International Diversification 122 2.6 Is the Current Investor Base Prone to Herding? 128 2.7 Do Investors Overreact to Changes in Brady Bond Prices? 132 2.8 The Form in Which Foreign Investors Participate in Emerging Markets Will Vary with the Extent of Financial Integration 134 3.1 The Empirical Link between Integration and Deepening 160 4.1 Capital Controls in Chile, Colombia, and Malaysia 179 4.2 Objectives of Capital Controls and Conditions for Effectiveness 198 4.3 Effectiveness of Capital Controls in Chile, Colombia, and Malaysia 201 4.4 Recent Country Experience with Exchange Rate Bands 204 4.5 The Mexican Tesobono Crisis 213 5.1 The Costs of Banking Crises 234 5.2 International Financial Integration and Boom-Bust Cycles 238 5.3 Excessive Bank Lending Tends to Increase the Current Account Deficit 242 5.4 Excessive Bank Lending Can Lead to Underinvestment and Overconsumption 244 5.5 Bank Lending Booms and Overheating: The Importance of Fiscal Policy 247 5.6 Good Bankers Make a Safe and Sound Banking Industry 264 5.7 Containing the Lending Boom: Taxation 274 5.8 Orthodox versus Heterodox Bank Regulations 279 5.9 Infeasible Policies in Banking Crisis Management 282 6.1 Investors' Viewpoint: Risks and Transaction Costs in Emerging Markets 310 6.2 Corporate Governance in India 318 6.3 Clearance, Settlement, and Depository Functions in Thailand 325 6.4 Developing Capital Market Infrastructure in Asia and Latin America 334 vii VS TO DEVELOPING COUNTRIES 6.5 The Anatomy of Bolsa de Mexico's Clearance, Settlement, and Depository System 340 6.6 The Regulator and Market Development 345 6.7 Promoting Domestic Institutional Investors 350 6.8 Effective Governance of sRos in the United Kingdom and the United States 362 6.9 The Role of the State in Capital Markets in Asia and Latin America 365 6.10 Developing Local Bond Markets 374 Text figures 1.1 Private Capital Flows to Developing Countries, 1975-96 10 1.2 Composition of Net Private Capital Flows to Developing Countries, 1980-82 and 1995-96 11 1.3 Concentration of Private Capital Flows, Selected Developing Countries, 1990-95 12 1.4 Composition of Net Overall Private Capital Flows by Region 13 1.5 Large Reversals in Net Private Capital Flows 28 1.6 Sources of Volatility in Emerging Markets 29 1.7 Changes in Volatility of Reserves: 1980-89 and 1992-96 30 1.8 Banks' Share of Financial Intermediation, Selected Developing and Industrial Countries, 1994 46 1.9 Health of Banking Systems 47 1.10 Capital Flows, Lending Booms, and Potential Vulnerability 49 1.11 Official and Private Gross Capital Flows to Low-Income Countries (excluding China and India), 1970-95 67 1.12 Official and Private Gross Capital Flows to the 12 Largest Recipients of Private Capital, 1970-95 69 2.1 Structural Forces Driving Private Capital to Developing Countries 76 2.2 Effects of Pohcy Reforms in Developing Countries Receiving Large Private Capital Flows 87 2.3 Output and Export Growth, Selected Regions, 1986-90 and 1991-95 88 2.4 Potential Benefits from Portfolio Risk Diversification 90 2.5 Correlations of Returns among Emerging Markets 91 2.6 Institutionalization of Savings 98 2.7 International Diversification of Institutional Investors, Selected Countries, 1990 and 1994 100 2.8 Entry Restrictions for Foreign Investors in Emerging Stock Markets, 1991 and 1994 103 2.9 International and Emerging Market Assets of U.S. Open-End Mutual Funds, 1990-95 107 2.10 Emerging Market Investments of Industrial Country Mutual Funds, 1993-95 108 viii 2.11 International and Emerging Market Assets of U.S. Pension Funds, 1990-94 110 2.12 Developing and Industrial Countries' Growth, 1980-2005 112 2.13 Demographic Structures of Developing and Industrial Countries, 1995 113 2.14 Elderly Dependency Ratios, Selected Industrial Countries, 1990-2030 114 2.15 Pension Fund Assets and International Diversification, Selected Countries 121 2.16 International Interest Rates and Portfolio Flows to Emerging Markets, 1993-96 127 2.17 Excess Volatility in Emerging Markets, Selected Countries and Years 131 2.18 Impact of the Mexican Crisis on Stock Market Prices in Emerging Markets, 1994-95 138 3.1 Initial Financial Depth and Future Growth 158 3.2 Portfolio Equity and Capital Market Development, Selected Countries 159 3.3 Growth in Domestic Trading in Selected Emerging Equity Markets 162 4.1 Change in the Current Account Deficit and the Real Effective Exchange Rate during Inflow Episodes, Selected Countries 187 4.2 Change in Growth and the Composition of Absorption during Inflow Episodes, Selected Countries 190 4.3 Change in the Composition of Absorption and the Real Effective Exchange Rate during Inflow Episodes, Selected Countries 191 5.1 Bank Lending to the Private Sector during Inflow and Pre-inflow Periods, Selected Countries and Years 240 5.2 Stock Prices during Inflow Periods, Selected Countries 248 5.3 Change in Nonperforming Loans 255 5.4 Change in Foreign Currency Exposure 255 5.5 Change in Net Interest Margin 255 5.6 Change in Capital-Asset Ratios 255 5.7 Change in Loan Loss Provisions 255 5.8 Exposure to Sectoral Risks 255 5.9 Macroeconomic and Financial Sector Vulnerabilities during Capital Inflow Episodes, Selected Countries and Years 258 6.1 Factors Affecting Volatility of Asset Prices in Emerging Markets, 1990s 321 6.2 Excess Volatility and Market Development 322 6.3 Settlement and Postsetdement Efficiency, Selected Emerging and Developed Markets, 1995 336 6.4 Bid-Ask Spreads in Emerging and Industrial Country Equity Markets, 1996 372 6.5 Major Types of Trading Systems 379 ix fIZS 1~LWS_TO DEVELOPING COUNTRIES Text tables 1.1 Net Private Capital Inflows to 20 Developing Countries, 1990s 27 1.2 Volatility of Private Capital Flows from the United States, 1980s and 1990s 31 1.3 Volatility of Capital Flows, Selected Countries, 1980s and 1990s 31 1.4 Surges in Private Capital Flows to 20 Countries: Policy Responses and Outcomes 36 1.5 Impact of the Mexican Crisis on Selected Developing Countries 41 2.1 Volatility in Various Stages of Financial Integration 136 3.1 Change in Domestic Activity in Selected Emerging Equity Markets 161 3.2 Bank Operating Ratios, OECD Economies, 1980s 162 3.3 Potential Savings from Adopting Best Practice, Selected EU Countries 163 3.4 Spread between Short-Term and Deposit Rates: France, Germany, and Spain, 1980-89 164 3.5 Cross-correlation of Consumption Growth, Selected Regions, 1960-87 167 4.1 Net Private Capital Inflows to 21 Developing Countries, 1988-95 175 4.2 Main Policy Responses to Surges in Capital Inflows, 1988-95 178 4.3 Disposition of Private Capital Inflows during Inflow Episodes 180 4.4 Macroeconomic Performance during Inflow Episodes 186 4.5 Composition of Absorption during Inflow Episodes 189 4.6 Vulnerability to the Mexican Crisis, Selected Countries, 1988-93 194 5.1 Indicators of Financial Intermediation, Selected Countries 230 5.2 Domestic and External Financial Sector Reform: Policy Recommendations 260 5.3 Containing Lending Booms: Policy Options 268 5.4 Banking Crisis: Policy Options 276 6.1 Stock Market Growth in Selected IFC Index Countries, 1985-96 305 6.2 Estimates of Foreign Presence in Emerging Stock Markets, 1995 306 6.3 Issues of American and Global Depository Receipts, 1989-95 308 6.4 Conformity with the Group of 30 Recommendations 337 6.5 Investment Restrictions in Emerging Equity Markets, 1995 366 6.6 Legal and Regulatory Initiatives in Emerging Markets 368 6.7 Capital Market Development in Emerging Markets, 1995 371 x Foreword F INANCIAL MARKETS AROUND THE WORLD ARE RAPIDLY INTE- grating into a single global marketplace, and developing coun- tries are increasingly part of this process. The process is being driven by advances in communications and information technology, deregulation of financial markets, and the rising importance of insti- tutional investors that are able and willing to invest internationally. The good news is that developing countries are attracting private capi- tal flows by improving macroeconomic policy and by establishing institutions and regulatory regimes that have increased creditworthi- ness and promise a more stable environment. Moreover, investors are also becoming more sophisticated in differentiating among countries and their economic fundamentals. Finally, after the Mexican crisis of 1994-95, the international community has realized that more should be done to reduce volatility and risks in international financial mar- kets by improving market disclosure and strengthening coordination among national authorities. Nevertheless, there remain reasons for concern. First, for the twenty or so countries that have been the major recipients, the management of private capital flows has not proved to be easy. It is not just the volume of flows, but the speed at which such investment pours in-and can be withdrawn-that present particular challenges to these economies. Governments need to build the kind of macroeconomic, regulatory, and institutional environments that channel this private capital into broad-based and sustainable growth. Second, the overwhelming major- ity of developing countries, in particular the smaller low-income economies, still need to create the conditions to attract private capital and must depend on declining official flows. This report makes a serious and timely contribution to the analysis of these issues. It explores the nature of the changes that are leading to the integration of developing countries in world financial markets, and it analyzes the policy challenges these countries face in attracting and managing private capital flows. It concludes, for example, that coun- tries receiving large capital inflows should avoid using them to finance xi I' TO3?'O TO DEVELOPING COUNTRIES large fiscal deficits or consumption booms. The report also includes specific recommendations and warnings on regulatory design that may be useful to developing countries as they seek to maximize the positive contribution of capital inflows while minimizing their potentially dis- ruptive effects. This book, therefore, will be highly useful to policymakers in devel- oping countries and, more generally, to all development specialists. But it will also be essential reading for members of the global financial com- munity. Investors have seen in recent years how dynamic-and some- times volatile-emerging markets can be. Understanding these opportunities and challenges is critical for everyone. The report also comes at an important time for the World Bank. The challenge for the Bank and other development agencies is to create strategies to help developing countries leverage pr.vate capital flows so that all benefit. The research presented in this book is an important step in constructing such strategies. Like previous volumes in the Policy Research Report series, Private Capital Flows to Developing Countries is designed for a wide audience. It is a product of the staff of the World Bank; the judgments made in the report do not necessarily reflect the views of the Board of Directors or the governments they represent. Joseph E. Stiglitz Senior Vice President and Chief Econormist The World Bank April 1997 xii The Report Team T HE REPORT TEAM WAS LED BY AMAR BHATTACHARYA AND comprised Pedro Alba, Swati Ghosh, Leonardo Hernandez, Peter Montiel, and Holger Wolf. Substantial background input was also provided by Andrea Anayiotos (chapter 2), Soonhyun Kwon (measurement of financial integration), Maria Soledad Martinez-Peria (chapter 5), and Michael Pomerleano (chapter 6). The main research assistance was provided by Jill Dooley, Matthew Anderson, and Jos Jansen. Bakhodir Atahodjaev, Madeleine Li-Chay- Chung, and Alok Garg also provided research assistance. Deborah Davis edited the first draft of the study, and Anatole Kaletsky provid- ed editorial input on the summary and chapter 1. K. Anna Kim pro- vided logistical and secretarial support for the overall study, and Hong Vo for chapters 2 and 5. The study was initiated by and carried out under the direction of Michael Bruno and Masood Ahmed. Joseph Stiglitz provided guidance in the dosing stages of the study. The editorial production team was led by Anthony Pordes. The team was assisted by Paola Brezny, Amy Brooks, and Audrey Heilig- man. Joyce Gates provided production support. xiii Acknowledgments T HE REPORT HAS BENEFITED FROM DISCUSSIONS WITH AND comments and contributions from many persons outside and within the World Bank. Special thanks are due to those who have provided guidance and feedback from the inception to the com- pletion of the study. These include Gerard Caprio, Stijn Claessens, Richard Frank, Morris Goldstein, Sarwar Lateef, Diana McNaughton, Gary Perlin, D. C. Rao, Jed Shilling, and Paulo Vieira da Cunha. Special thanks are also due to Mariano Bengoechea, Claudia Morgenstern, Susan Pascocello, Ester Saverson Jr., and Robert Strahota for their extensive feedback on chapter 6. The report also benefited from discussions with Ismail Dalla, Jack Glen, Ross Levine, Millard Long, and Andrew Sheng. We would like to acknowledge support from staff and management of the International Monetary Fund; Sunil Sharma participated in a mission to prepare a back- ground paper, and Donald Mathieson, Anthony Richards, and Michael Spencer provided valuable comments. The report would not have been possible without the willingness of the private sector to share its perspective on investing in emerging mar- kets. The report has drawn on interviews with many fund managers and pensions funds, information provided by global custodians, and discussions with many others in the financial community. Although they are too numerous to mention, we would like to make special note of the support received from the Frank Russell Company; Citibank, N.A.; the Bank of New York; INTERSEC Research Corporation; and the International Federation of Stock Exchanges (FIBV). The report drew on background papers or inputs by Reena Aggarwal, Philip Brock, Jose de Gregorio, Persephone Economou, Ian Giddy, Morris Goldstein, Naoko Kojo, Elizabeth Morrissey, Gabriel Perez-Quiros, Maria Scattaglia, and Charles Seeger. The study has also drawn on a number of country studies that we intend to include in a subsequent volume. The Asian component of the review of country ex- periences was undertaken collaboratively with the Asian Development Bank (ADB). The study has also benefited from a joint ADB-World xv = , WS TO DEVELOPING COUNTRIES Bank workshop on "Private Capital Flows: Implications for Capital Markets in Asia," and we would like to thank the participants in that workshop. Many individuals in the International Economics Depart- ment provided data needed for the study; we would like to make spe- cial acknowledgment of the support received from ]'unam Chuhan and Himmat Kalsi. Many Bank staff members provided valuable comments, including Noritaka Akamatsu, Suman Bery, Francis Colaco, Uri Dadush, Asli Demirgiiu-Kunt, Cevdet Denizer, Hinh Dinh, Daniela Gressani, Nor- man Hicks, Akira lida, Ronald Johannes, Johannes Linn, Will Martin, Vikram Nehru, Ngozi Okonjo-Iweala, Guillermo Perry, Thierry Pujol, Zia Qureshi, Jean-Fran,ois Rischard, Hemant Shah, Robert Shakotko, C. K. Teng, John Williamson, and Roberto Zagha. Finally, we have benefited from comments received at a Board seminar on the draft report. xvi Data Notes and Abbreviations Data Notes H ISTORICAL DATA IN THIS BOOK MAY DIFFER FROM THOSE in other World Bank publications if more reliable data have become available, if a different base year has been used for constant price data, or if countries have been classified differently. * Billion is 1,000 million. * Trillion is 1,000 billion. * Dollars are current U.S. dollars unless otherwise specified. Abbreviations and Acronyms ASEAN Association of Southeast Asian Nations BIS Bank for International Setdements C&S Clearance and settlement (securities) CPI Consumer price index FDI Foreign direct investment IFC International Finance Corporation IFCI IFC Emerging Markets Index IMF International Monetary Fund IoSco International Organization of Securities Commissions ISSA International Society of Securities Administrators GDP Gross domestic product GNP Gross national product SRO Self-regulatory organization xvii Summary HE WORLD'S FINANCIAL MARKETS ARE RAPIDLY integrating into a single global marketplace, and ready or not, developing countries, starting from different points and moving at various speeds, are being drawn into this process. If they have adequate institutions and sound policies, developing countries may proceed smoothly along the road to financial integration and gain the considerable benefits that integration can bring. Most of them, however, lack the prerequisites for a smooth journey, and some may be so ill prepared that they lose more than they gain from finan- cial integration. Developing countries have little choice about whether to follow this path, because advances in communications and new developments in finance have made the course inevitable. They can, however, decide how they wish to travel, choosing policies that benefit the economy and avert potential shocks. This volume describes the forces that have created and that sustain this road, analyzes the benefits and problems likely to be encountered on it, and examines the experiences of those who are farther along on the journey to see what can be learned from them. The Changing Financial Environment While the cyclical downturn in global interest rates provided an im- portant initial impetus for the resumption of private capital flows to developing countries in the 1990s, these flows have now entered a new phase, reflecting structural forces that are leading to progressive finan- cial integration of developing countries into world financial markets. The two primary forces that are driving investor interest in developing 1 1$ TO DEVELOPING COUNTRIES countries are the search for higher returns and opportunities for risk diversification. Although these forces have always motivated investors, the responsiveness of private capital to cross-border opportunities has gained momentum as a result of internal and external financial deregu- lation in both industrial and developing countries and major advances in technology and financial instruments. This process of financial integration is still unfolding. The pace of change will be especially rapid for developing countries, given their more insulated financial markets. Even in the more regulated economies, growing economic sophistication means that financial inte- gration will increasingly not be a choice for governrnents to make. Mar- kets are making the choice for them. As a result, the financial integration of developing countries is expected to deepen and broaden over the coming decade against a background of increasing global financial inte- gration. As part of this process, gross private capital flows may be ex- pected to rise substantially, with capital flowing not only from industrial to developing countries but also, increasingly, among developing coun- tries themselves and from developing to industrial countries. Given the continuing decline in investment risks, the higher ex- pected rates of return in developing countries, and the underweighting of emerging markets in institutional portfolios, net private capital flows to developing countries in aggregate are likely to be sustained. The rate of growth, though, will inevitably diminish. There will undoubtedly also be considerable variation among countries, depending on the pace and depth of improvements in macroeconomic per3formance and credit- worthiness. Such basic factors as domestic politics, the availability of re- sources, and the level of development that has been attained are bound to affect the flow of capital as well, so the process of financial integration can take many courses. In fact, in countries where economic and policy fundamentals are quite weak, the initial manifestation of growing finan- cial integration may take the form of net outflows of private capital. With changes in the international financial environment, there are likely to be considerable year-to-year fluctuations in private capital flows to developing countries, even in aggregate. These nations are, and will continue to be, highly susceptible to both domestic shocks and changes in the international environment, such as in global interest rates. Never- theless, flows to developing countries in the aggregate are unlikely to suffer from major reversals as long as the probability of abrupt changes in the international environment remains low. The main risks of volatil- 2 ity and large reversals lie at the individual country level and stem from the interaction of domestic conditions and policies with international factors. And as markets become more discerning, contagion effects of the kind seen after the Mexican crisis are not likely to be long-lasting. Winning and Losing in an Integrating Market The experience of nations that have successfully managed financial inte- gration suggests that the benefits of this process are likely to be especially large for developing countries. The direct advantages are twofold: these countries can tap the growing pool of global capital to raise investment, and they can diversify risks and smooth the growth of consumption and investment. The more important benefits of financial integration, how- ever, are likely to be indirect. These include knowledge spillover effects, improved resource allocation, and strengthening of domestic financial markets. In addition, the increasing safety of financial operations in de- veloping markets can support a shift to higher-return investments, with gains for both developing and industrial nations. As the Mexican peso crisis has so forcefully demonstrated, however, these benefits are by no means assured. In fact, there are large potential costs if integration is not carefully managed. There are two reasons for this. First, although international investors are becoming more discern- ing, market discipline tends to be much more stringent when investor confidence is lost-a fact that can lead to large outflows-than during the buildup to a potential problem. Second, and more important, many developing countries lack the preconditions needed to ensure the sound use of private capital and manage risks of large reversals. Finan- cial integration can magnify the effects of underlying distortions and institutional weaknesses in these countries and thereby multiply the costs of policy mistakes. The challenge, therefore, for developing countries is how to exploit the growing investor interest in their markets and so to enter a virtuous cycle of productive financial integration rather than a vicious cycle of boom and bust. In a virtuous cycle, integration and access to external private capital lead to increased productive investment, momentum for policy and institutional reform, and greater resilience to potential in- stability. In contrast, when the necessary macroeconomic fundamentals are lacking, banking systems are weak, and domestic distortions are pervasive, countries may experience capital flight rather than capital in- 3 :b t~V~~S TO DEVELOPING COUNTRIES flows, or they may be unable to use inflows efficiently-with very high costs in terms of growth and instability. The Lessons from Evolving Expeience This report provides perspectives on how developing countries can re- spond to these challenges, drawing on the evolving experience of the more rapidly integrating countries. Although the agenda confronting policymakers is necessarily broad and complex, ranging from macro- economic issues to the so-called plumbing of markets, a number of strategic themes emerge from this study. * Given thegrowingtrend towardfinancial integrat,ion, developingcoun- tries need to vigorously pursue policies that will enable them to benefit from global capitalflows and avoid the associated dangers. There is broad consensus-based on lessons from country experi- ence and the considerable literature on the sequencing of reforms- that the most important prerequisites for suaccessful financial integration are a sound macroeconomic policy framework, in particu- lar a strong fiscal position, the absence of large domestic price distor- tions (for example, those arising from import protection), a sound domestic banking system with an adequate supervisory and regulatory framework, and a well-functioning market infrastructure and regula- tory framework for capital markets. All these are key elements of the broader policy agenda that developing countries rLeed to adopt in any case, and financial integration only makes their pursuit more urgent, for several reasons. First, progress on these prerequisites will help improve a country's creditworthiness and attractiveness to foreign investors. Second, these preconditions will encourage capital flows (such as foreign direct in- vestment) based on long-term fundamentals rather than short-term re- turns. Third, attainment of these preconditions will ensure that capital inflows are well used, ultimately determining whether countries can reap the benefits from financial integration and avoid its risks. Fourth, the more robust a country is with regard to these preconditions, the greater will be its latitude in responding to surges and volatile flows. * How countries respond to the initial surge of capital inflows, which is often associated with the openingphase of integration, will largely deter- 4 :-, oT mine their success in dealing not only with overheatingpressures but also with potential vulnerability. Countries have typically used a combination of policies to respond to large surges of capital inflows. Capital controls, when combined with other policies, appear to have been at least partially successful in reducing the magnitude of inflows and altering their composition. Sterilized intervention has been the most widely used instrument and has been generally successful as an initial response in curbing the growth of base money and building up reserves. Most important, coun- tries that have resisted real exchange rate appreciation by placing greater emphasis on fiscal tightening have tended on average to have lower current account deficits, a mix of absorption oriented toward in- vestment, and faster economic growth. The main lesson for macroeconomic policy from recent country ex- perience is that a heavy reliance on fiscal policy, supported by steriliza- tion and some nominal exchange rate flexibility-and in the more extreme cases by temporary taxes or controls on inflows-can be an ef- fective response to overheating and can reduce the likelihood of vul- nerability to large reversals of private capital flows. More generally, countries are likely to suffer a loss of investor confidence when the real exchange rate is perceived to be out of line, the government's debt obligations are large in relation to its earning capacity and external re- serve position, fiscal adjustment is perceived to be politically or admin- istratively infeasible, or the country's growth prospects are bleak. * There is much merit in curbing lending booms associated with capital in- flows while addressing the underlying weaknesses in the banking system. The banking system plays a dominant role in the allocation of capital in a developing country, and the health of this system largely determines whether a country will be able to exploit the benefits of financial integration and avoid its pitfalls. In many developing countries, banking systems have only recently been deregulated, incentives for banks are distorted toward ex- cessive risk taking, banks are poorly capitalized, and adequate prudential regulation and supervision capabilities have not yet been established. Addressing the underlying weaknesses of the banking system be- comes more urgent in a globally integrated environment because banks can increase lending more easily and incur greater risks. The standard 5 3EFWS TO DEVELOPING COUNTRIES tools for bank monitoring and supervision are rendered less effective. Institution building, removing incentive distortions (for instance, in the form of excessive insurance), and strengthening bank supervision capabilities are therefore crucial. Since reforms of the banking system will take time to implement, it will probably be necessary to curb the lending booms associated with capital inflows by using macroeconomic policies, as well as more tar- geted restrictions, such as raising reserve requirements or adopting risk- weighted capital adequacy requirements. This will help alleviate overheating pressures resulting from surges in capital inflows and will reduce the vulnerability of the banking system. * Development of well-functioningcapital markets will reduce risks ofpoten- tial instability as well as attract the growingpool ofjortfolio investment. Investors are concerned with the unreliability of emerging markets in three main areas: market infrastructure (where the consequences in- clude high transaction costs, frequent delays in settlement, and out- right failed trades); protection of property rights, in particular those of minority shareholders; and disclosure of market and company infor- mation and control of abusive market practices. Unfortunately, there are no simple solutions to preparing capital markets for financial inte- gration, which requires concerted action across a broad array of areas to improve market infrastructure and the regulatory framework. International standards for market infrastructure provide excellent medium-term benchmarks for emerging markets, although they need to be tailored to fit individual country circumstances. The experience of the more advanced emerging markets, especially those in Asia, indi- cates that it is possible to improve market infrastructure in a relatively short period by leapfrogging to state-of-the-art systems. This experi- ence, however, together with the not infrequent wveaknesses in emerg- ing market financial intermediaries, also suggests two cautionary notes. First, improving the speed of settlement and custody functions should not be achieved at the expense of reliability. Second, despite the impor- tance of promoting competition among financial intermediaries, mem- bership standards in key capital market institutions should be set high to bolster market safety and improve investor confidence. A regulatory model based on disclosure and self-regulation is gain- ing wide acceptance in emerging markets because it has strong advan- 6 tages over direct government regulation. But government regulation and oversight are still essential, and the state can play a crucial role in capital market development in partnership with the private sector, pro- viding the basic legal structures, for example, and fostering vital market institutions. Given the weaknesses in the regulatory systems of many emerging markets and their susceptibility to "reputational risk," the tradeoff between market development and effective regulation re- quired to develop and maintain market confidence is less pronounced than is sometimes thought. Emerging markets should also promote the development of domes- tic institutional investors. By mobilizing significant amounts of re- sources, these investors can serve as a counterweight to foreign investors and thereby assuage fears of excessive foreign presence; they also reduce the vulnerability of domestic capital markets to foreign in- vestor herding, and their presence may reassure foreign investors about the nation's respect for corporate governance and property rights. All these policy and institutional initiatives to attract foreign in- vestors and contain the potential negative impact of financial integra- tion on capital market volatility will significantly help the development of domestic capital markets. In turn, there is increasing evidence that well-functioning capital markets make an important contribution to the overall growth process. * Developing countries need to build better shock absorbers and develop mechanisms to respond to instability because they will remain highly vulnerable to economic disturbances for some time. Growing financial integration may require three types of shock ab- sorbers. First, the level of international reserves needs to be established in relation to the variation in the capital account, rather than in terms of months of imports, since the level of gross flows is higher following integration. For countries where investor confidence is less firm, there is a case for an even larger cushion of reserves. International reserves can also be buttressed with contingent lines of credit, as Argentina has done recently. Second, financial integration heightens the need for fis- cal flexibility, which in turn will depend on the level of public debt, among other things. Third, there is strong merit in building up cush- ions in the banking system. Authorities should use periods of credit boom to increase bank capitalization and provisioning requirements as 7 a4Y WS_ TO DEVELOPING COUNTRIES a way to promote sound banking practices and increase the resilience of banks. Even with these shock absorbers, countries need to have well-delin- eated mechanisms that enable policymakers to deal with crises promptly and effectively. This is particularly impori ant for the banking system, where delaying actions intended to contain a crisis will only in- crease its cost. * International cooperation between regulators and adequate disclosure of information at all levels are increasingly important to ensuring safe and efficient markets. The globalization of financial markets, along with new forms of in- vestment and the growing prevalence of financial conglomerates, in- creases the number of channels that transmit systemic shocks across borders and sectors and the speed at which these shocks travel. At the same time it reduces the transparency of the marketplace. Despite the worldwide integration of financial markets, the author- ity of regulators has remained mainly national in scope. In a global marketplace, it is difficult to assess the risk exposure of financial inter- mediaries, and such complex operations as derivatives trading make risk evaluation even more uncertain. To address these problems, regu- latory authorities from industrial countries are increasingly cooperating and coordinating with one another. In addition, to reinforce market discipline, these regulators are emphasizing the supervision of the qual- ity of risk management by financial firms (rather than the position of a firm at a particular moment in time) and improved disclosure practices. In this new environment, reducing information asymmetries across borders will have a significant payoff for emerging markets. At the macroeconomic level, more accurate and timely disclosure of country information would decrease the likelihood of cross-country contagion of financial shocks. Adequate disclosure of the risk profile and financial status of financial intermediaries would increase the effectiveness of market discipline and facilitate supervision by regulators. And for reg- ulators, there will be a large payoff in coordination and information- sharing agreements, with both industrial and other emerging markets, in particular with those likely to share financial institutions and sources of financial shocks. 8 The Main Findings T HERE CAN BE NO DOUBT THAT DEVELOPING countries have entered a new age of globalized pri- vate capital. Consider these facts: *Net private capital flows to developing countries exceeded $240 billion in 1996 ($265 billion if the Republic of Korea is included), nearly six times times greater than they were at the start of the decade, and almost four times more than the peak reached during the 1978-82 commercial bank lending boom (figure 1.1).' * Private capital flows now dwarf official flows in terms of relative importance. They are now five times the size of official flows. This is remarkable, since only five years ago official flows to de- veloping countries were larger than private flows. * Developing countries are now a much more important destina- tion for global private capital. Their share of global EDI flows is now almost 40 percent, compared with 15 percent in 1990, and their share of global portfolio equity flows is now almost 30 per- cent, compared with around 2 percent before the start of the decade. * The importance of private flows has also increased markedly in the economies of developing countries, from 4.1 percent of do- mestic investment in 1990 to almost 20 percent in 1996. * There has been a remarkable broadening in the composition of private capital flows (figure 1.2). Whereas traditional commercial bank lending used to account for more than 65 percent of all pri- vate flows, FDI has now emerged as the most important compo- nent of private capital flows, and starting from a negligible level 9 StOW TO DEVELOPING COUNTRIES Figure 1.1 Private Capital Flows to Developing Countries, 1975-96 (net long-term internationalprivate capitalflows) Billions of dollars Percent 300 - 100 --Net private capital flows 250 . - i Percentage of total 80 flows 80 -h-- Percentage of fixed investment 200-_l---6 - 40 100 - -. < . } t ,; } - ~~~~~20 Private capital flows to developing countries have surged during the 1990s. o__ I '__X 1975 1978 1981 1984 1987 1990 1993 1996 Source World Bank data. in 1989, portfolio flows-both bonds and equities-have in- creased sharply so that they now account for more than a third of total private capital flows. * The flow of private capital has also shifted on the recipient side away from governments to the private sector. Borrowing by the public sector now accounts for less than a fifth of total private flows, so that the bulk of these flows to developing countries is passing through market channels. * Private capital flows to developing countries have proved to be surprisingly resilient. Despite the increases in U.S. interest rates in 1994 and the Mexican crisis in 1995, private flows increased a further 55 percent during the past three years. Just as private capital is flowing into developing economies, the fi- nancial markets that channel this tide of investment are becoming in- creasingly integrated. More than a century ago, the great scholar and 10 '4 Figure 1.2 Composition of Net Private Capital Flows to Developing Countries, 1980-82 and 1995-96 1980-82 199596 Portfolio Bank and trade-related Portfolio Foreign direct onds Portfolio equity As FDI and portfolio flows have grown, bank and trade-related lending have declined in relative importance. Bank and trade-related Foreign direct lending investment Source: World Bank data practitioner of finance Walter Bagehot noted the beginnings of this trend toward integration, observing that "the same instruments which diffused capital through a nation are gradually diffusing it among na- tions." Prescient, he went on to warn that while "the effect of this will be in the end much to simplify the problems of international trade ... for the present, as is commonly the case with incipient causes whose ef- fect is incomplete, it complicates all it touches" (Bagehot 1880, p. 71). The current wave of financial integration, however, differs from its counterpart of the late nineteenth century in at least three respects: its geographic scope, the extent and depth of integration, and the speed with which markets can now react. Developing countries vary considerably in the rate at which they are integrating with global financial markets as well as in the amounts of private capital they have attracted. Only a dozen countries-all of which are rapidly integrating with international financial markets- accounted for about 80 percent of net private flows to developing na- tions during 1990-95, and less than a score accounted for 95 percent (figure 1.3). In these countries, private capital flows have been very large in relation to the size of their domestic economies, constituting as much as 5 percent of gross national product (GNP) in some surge years. In contrast, most developing countries are just starting the process of financial integration and have so far received much smaller amounts of private capital. In fact, 140 of the 166 developing nations account for 11 = S0WS TO DEVELOPING COUNTRIES Figure 1.3 Concentration of Private Capital Flows, Selected Developing Countries, 1990-95 (net long-term internationalprivate capitalflows) Billions of dollars Percent 180 - - - - X - - - - 100 160 _ X Net private capital flows, 1990-95 ' - 90 Cumulative share of private capital flows to developing countries ,80 140 - so - L 100 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~1 Chmna Mexico Brazil Korea Malaysia Argentina Thailand Indonesia Russia India Tuirkey Hungary Source World Bank data A! dozen countries accounted for about less than 5 percent of private capital flows to such countries. These dif- 80 percent of net international private fierences, however, become less stark when scaled by GNP and when pri- c.apital flows. vate transfers and other unrecorded private flows are taken into account (figure 1.4) . In the case of the Middle East and North Africa and South Asia, such flows are more important than recorded private interna- tional capital flows, suggesting that integration is proceeding through several channels. The chapter briefly describes the setting in which the process of fi- nancial integration takes place, and chapter 2 examines this back- ground in greater detail. In the next section, we outline the benefits of financial integration, a topic covered more thoroughly in chapter 3, as well as the risks and potential costs of integration. The last part of this chapter explores the lessons from developing nations' recent experience with global financial flows and provides recommnendations for sound policies and institutions, summarizing the material covered in depth in chapters 4 through 6. 12 Figure 1.4 Composition of Net Overall Private Capital Flows by Region Percentage of GNP …E- ; _v_ _ . __ .. ._< .< v Other private capital Private transfers * International private - capital flows ---- - ---- -- < . _ i .............. _ _' _ .... _But when economic size and all types Sub-Saharan East Asia and Latin America North Africa South Asia Europe and of private flows are taken into Africa Pacific and Caribbean and Middle Central Asia account, private capital does not flow as disproportionately to a few Source IMF, International Financial Statistics data base; World Bank data, developing countries. A New Age of Global Capital T WO PRIMARY FORCES ARE DRIVING INVESTOR INTEREST IN developing countries and leading to their increased integra- tion: the search for higher returns and the opportunity for risk diversification. Although these underlying forces have always moti- vated investors, the responsiveness of private capital to cross-border opportunities has gained new momentum as a result of internal and external financial deregulation in both industrial and developing countries and major advances in technology and financial instru- ments. Figure 2.1 in chapter 2 summarizes the process under way. In industrial countries two key developments have increased the re- sponsiveness of private capital to cross-border investment opportuni- ties. First, competition and rising costs in domestic markets, along with falling transport and communications costs, have encouraged firms to 13 S*~L-O-S TO DEVELOPING COUNTRIES look for opportunities to increase efficiency by producing abroad. This is leading to the progressive globalization of production and to the growth of "efficiency-seeking" FDI flows. Second, financial markets have been transformed over a span of two decades from relatively insu- lated and regulated national markets toward a more globally integrated market. This has been brought about by a mutually reinforcing process of advances in communications, information, and financial instru- ments, and by progressive internal and external deregulation of finan- cial markets. An important facet of this globalization of capital markets has been the growing importance of institutional investors who are both willing and able to invest internationally. For exanple, total assets of pension funds at the global level are estimated to have increased from $4.3 trillion in 1989 to $7 trillion in 1994, while, at the same time, the share of nondomestic investment in their portfolios rose from around 7 percent in 1989 to 11 percent in 1994. Together, these two factors have resulted in an increase in total international investments by pension funds from $302 billion in 1989 to $790 billion in 1994. In developing countries, the environment is also changing rapidly. Since the mid-1980s, several countries have embarked on structural re- form programs and increased openness of their markets, through the progressive lowering of barriers to trade and forcign investment, the liberalization of domestic financial markets and removal of restrictions on capital movements, and the implementation of privatization pro- grams. There have also been major improvements in fiscal performance and the sustainability of external debt. Although the perceived risks of investing in emerging markets re- main relatively high, the more stable macroeconomic environment, growth in earnings capacity (both output and exports), and a reduction in the stock of debt in many countries (following the implementation of the Brady Plan) are leading to a decline in such risks and an increase in the expected rates of return in the major recipient countries.2 The second force behind the structural trend in private capital flows is investors' desire to diversify portfolio risks. Investors can benefit from holding emerging market equities because returns in emerging markets tend to exhibit low correlations with industrial country returns-that is, they tend not to move in tandem with those of industrial countries. By holding an asset whose returns are not correlated with the returns of another asset, investors can raise the overall return on their portfolios without a commensurate increasing in risk (variarnce). 14 Developing countries have begun to offer investors significant op- portunities for risk diversification that arise from the low correlation be- tween rates of return in emerging markets and industrial countries. This opportunity for portfolio diversification in emerging markets is a rela- tively recent phenomenon. The combined market capitalization of the 18 major developing countries that constituted the International Fi- nance Corporation (IFC) Emerging Markets Index (IFCI) in 1996 was, at $1.4 trillion, 14 times higher than it was in 1985. Although still much lower than that of industrial countries as a ratio of gross domestic prod- uct (GDP), the average market capitalization of these countries rose from 7 to 40 percent of GDP during the period. Turnover ratios reveal the same trends, increasing approximately twofold between 1985 and 1994. As a result of this growth, financial markets in developing countries are now beginning to provide foreign investors with significant opportuni- ties to diversify into investments that have good prospective returns and low correlations with their investments in industrial country markets. Growing Investments and Financial Integration As a result of these changes, developing countries have seen a strong surge in private capital flows of all kinds. FDI has responded most vig- orously to the improving economic environment in developing coun- tries. The driving factor for FDI has been the sustained improvement in domestic economic fundamentals. The nature of the FDI flows has also changed. In the 1970s and early 1980s, resource extraction and import substitution were the primary motives for FDI in developing countries. In contrast, a high proportion of FDI flows now going to developing countries can be characterized as being efficiency seeking, associated with the globalization of production. Commercial bank lending, which accounted for the bulk of private flows in the late 1970s and early 1980s, has also made a strong comeback. But most striking has been the growth of portfolio investment flows. Whereas developing countries attracted barely any portfolio flows a decade ago, in the past five years they have received almost 30 percent of all the eq- uity capital moving across national borders worldwide. Given its growing importance as a channel of financial integration, we focus primarily on portfolio investment in the analysis of international capital flows. The driving force behind the portfolio flows of the 1990s has been in- stitutional investors. The initial impetus for the globalization of institu- 15 ;,ys TO DEVELOPING COUNTRIES tional investment was probably the cyclical decline in global interest rates. But as improvements in economic fundamentals and creditworthiness began to take hold, and as the new investors became more familiar with emerging markets, the emphasis has shifted to long--term rates of return and opportunities for portfolio risk diversification. As a result, institu- tional investors have become an increasingly important segment of the in- vestor base in emerging markets and their influence seems certain to grow. Among institutional investors, mutual funds led the surge in invest- ments in emerging market equities. In 1986 there were 19 emerging market country funds and 9 regional or global emerging market funds. By 1995 there were over 500 country funds and nearly 800 regional and global funds. The combined assets of all closed- and open-end emerging market funds increased from $1.9 billion in 1986 to $10.3 billion in 1989 to $132 billion at the middle of 1996. International and global funds are also allocating increasing proportions of their portfolios to emerging markets, leading to a large increase in the total exposure of mutual funds to these markets. These trends have meant that emerging markets are accounting for a rising proportion of international investment by mutual funds- more than 30 percent of new international investments by U.S. mutual funds went to emerging markets during 1990-94. Since international investment itself has been rising, the share of emerging market assets in total mutual fund assets has risen quite sharply. The exposure of U.S. open-end mutual funds to emerging markets rose from just $1.5 billion in 1990 to $35 billion in 1995, or 14 percent of international exposure (figure 2.9, chapter 2). Pension funds have followed suit, investing through mutual funds or directly on their own account. Even though they began to invest in emerging markets more recently, allocations of U.S. pension funds to emerging markets are now comparable with those of mutual funds, with some of the bigger pension funds investing considerably larger proportions (figure 2.11, chapter 2). Indeed, U.S. pension fund invest- ments in emerging markets, including investmerLts made on their be- half by mutual funds, have been an important factor in propping up investment flows to emerging markets during 1994 and 1995. As with mutual funds, most pension fund investments in emerging markets are in the form of portfolio equities. The rapid growth and changing composition of private capital flows reflect the increasing financial integration of developing countries with 16 THE MAIN international financial markets. In the 1980s, private sector lending was mostly in the form of bank lending to sovereign governments. The bulk of private capital flows now is to the private sector and is increas- ingly taking place through channels that link markets for capital in de- veloping countries with their international counterparts. Since direct measures of financial integration for developing coun- tries are not available, we have constructed an indirect measure, using three separate elements: a country's access to international financial markets, its ability to attract gross financing, and the breadth of access to financial markets (box 1. 1).3 These indexes are then used to evaluate Box 1.1 Measunrng Financial Integration TO ASSESS THE DEGREE TO WHiCH COUNTRIES The third measure looks at the level of diversifi- are financially integrated, we have used several mea- cation of a country's financing, based on the com- sures to construct an overall index of integration. position of flows. The composition matters because We computed the index for the periods 1985-87 different components of flows have different effects and 1992-94 in order to evaluate the degree of inte- on financial integration: FDi benefits local suppliers, gration that has taken place since the mid-1980s. who may turn to foreign equity investors to obtain The first measure looks at a country's access to financing for expansion; banking inflows, which international financial markets. This is provided by lead to a deepening of the financial system, may en- the country risk ratings of the Institutional Investor hance the liquidity of stock markets and thus in- Survey. Ratings of less than 20 are categorized as crease the latter's attractiveness to foreign investors; low, ratings ofgreater than 50 as high, and ratings in and so forth. A significant participation in all three between as medium. forms of flows is desirable for balanced integration. The second measure looks at a country's ability Countries receiving a minimum of 5 percent of to attract private external financing computing the total flows for each type of flow were categorized as ratio of private capital flows to GDP. Since financial high. Countries that received a minimum of 5 per- integration implies the linking of markets, however, cent in two types of flows were categorized as not all flows were weighted equally in this measure. medium, and the remaining countries were catego- Thus, a country receiving private flows primarily in rized as low. the form of FDI is likely to be less integrated with These measures were then converted to an overall world financial markets than is a country that re- index of financial integration. As box table 1.1 ceives flows in the form of portfolio and commercial shows, developing countries as a group have become bank lending. Portfolio flows were therefore given a more integrated since the mid-1980s. The number weight of 5, bank flows a weight of 3, and FDI a of countries in the relatively highly integrated cate- weight of 1. Using these multipliers, countries gory has increased sizably, as has the number of whose flows amount to less than 10 percent of GDP countries in the medium category. were categorized as low, countries whose flows ex- A similar conclusion is suggested by the declining ceed 20 percent were considered high, and countries usage of convertibility restrictions and multiple ex- in between were categorized as medium. (Box continues on thefollowingpage.) 17 ,_1ITAL FLOWS TO DEVELOPING COUNTRIES Box Table 1.1 Change in Degree of Financial Integration 1985-87 1992-94 Korea- High Suriname Low Thailand High Tunisia Medium - Malaysia High Togo Low Trkey High Ecuador Medium - Thailand Medium + Tanzania Low Brazil High Kenya Medium - Cameroon Medium + Swaziland Low Argentina High Cameroon Medium - india Medium South Africa Low Korea Higb Egypt Medium - Colombia Medium Peru Low Indonesia High Togo Low Niger Medium Paraguay Low Malaysia Hzgh Mauritania Low Kenya Medium Sierra Leone Low Mexico High Myanmar Low Papua New Senegal Low Hungary High Nicaragua Low Guinea Medium Dominican Rep. Low Ghana High Tanzania Low Indonesia Medium Costa Rica Low Chile High Senegal Low Mexico Medium Congo Low Pakistan Htgb Sierra Leone Low Egypt Medium Ghana Low Philipptnes Higb Venezuela Low Chile Medium Gabon Low Mauritius Medium + Niger Low Sri Lanka Medium - El Salvador Low Panama Medium + Nigeria LOW Philippines Medium - China Low Colombia Medium + Zambia Low Cote d'Ivoire Medium - Benin Low Jamaica Medium + Guatemala Low Ecuador Medium - Bangladesh Low India Medium + Burkina Faso Low Mauritius Medium - Algeria Low Peru Medium Guyana Low Morocco Medium - Burkina Faso Low Papua New Guinea-Bissau Low Pakistan Medium- Brazil Low Guinea Medium Gabon Low Nigeria Medium - Bolivia Low Morocco Medium Costa Rica LOW Turkey Medium - Lesotho Low Zimbabwe Mediumn Congo Low Tunisia Medium - Jamaica Low C6te d'Ivoire Medium El SAvador LOW Uruguay Medium - Madagascar Low Uruguay Medium Dominican Rep. Low Zimbabwe Mediun - Mauritania Low China Medium Haiti Low Argentina Medium - Mali Low Sri Lanka Medium - Madagascar Low Trinidad and Hungary Low Suriname Medium - Lesotho Low Tobago Low Guyana Low Swaziland Medium - Mal Low Panama Low Guinea-Bissau Low Honduras Medium - Algeria LOW Zatnbia Low Guatemala Low Paraguay Medium - Bangladesh LOW Nicaragua Low Honduras Low South Africa Medium - Bolivi-a LOw Venezuela Low Haiti Low Trinidad and Benin Low Myanmar Low Tobago Medium - chanBe rates. While in 1985 some 60 percent of In- restrictions. In the same period, the use of capital ac- ternational Monetary Fund (iMF) memnbers imposed count restrictions decreased from 80 to 70 percent restrictions on current account transactions, by of IMF members, and the use of multiple exchange 1995 less than a third of members employed such rates declined from 30 to 16 percent of members. 18 THE MAI the degree of financial integration of various countries and changes that have taken place between the two periods 1985-87 and 1992-94. The index of financial integration in box 1.1 provides only a broad indication of the degree and change in financial integration. Neverthe- less, two main conclusions emerge from the results: * Developing countries as a group have become more integrated since the mid-1980s. The number of countries that are classified as highly integrated increased from 2 in 1985-87 to 13 in 1992-94, whereas the number of countries classified as moder- ately integrated increased from 24 to 26. * However, most developing countries are still in the very early stages of financial integration. The Prospects for Private Capital Flows The structural changes underlying capital flows to developing countries are continuing to unfold at both the domestic and the international levels, and there are several reasons to believe that they will spur grow- ing financial integration and the expansion of international capital movements in the years ahead: * Structural reforms taking place in developing countries are, in some cases, just beginning to produce positive results. * Technological change and financial innovation will continue to re- duce transaction costs and make distant markets more accessible. * Financial deregulation and competitive pressures will lead to greater emphasis on maximizing returns and diversifying risks. * Demographic shifts are becoming a major factor behind the long- term movement of capital from the industrial to the developing world. Although it is difficult to speculate on the course of future innova- tion and technological change, competitive pressures and increasing in- tegration will probably continue to reduce transaction costs and make distant markets more accessible to small as well as large investors. Such innovations will make policy-induced barriers less effective, spurring even more deregulation and competition. The pace of change will be particularly rapid for developing countries, given their nascent finan- cial systems. 19 s A CAPITAL FLOWS TO DEVELOPING COUNTRIES Developing country markets are indeed becoming more open and attractive to foreign investors as these nations inmplement macro- and microeconomic policy reforms. These reforms, in turn, are likely to in- crease the productivity of investment. On the expectation that devel- oping countries will continue to strengthen theii policies and that the external environment will remain favorable, developing countries are likely to grow at almost double the rate of industrial countries, provid- ing significant opportunities for investors (figure 2.12, chapter 2). As financial deregulation gathers speed worldwide, a far larger uni- verse of investors will seek to diversify their portfolios by investing in emerging markets. Despite the recent increases, the share of emerging markets in the portfolios of institutional investors remains small-well below their share in world market capitalization. Even in the United States, where institutional investors have increased their exposure to emerging markets more rapidly, mutual funds and pension funds are es- timated to hold an average of only 2 percent of their portfolios in emerg- ing markets. For most other industrial countries these shares are much smaller. Given that the share of emerging markets in world market cap- italization is currently around 10 percent and is expected to rise, the scope for benefits from further portfolio diversification is substantial. As a hypothetical illustration, given the current correlations of re- turns between emerging markets and an international portfolio, in- vestors could simultaneously increase expected rates of return and reduce the risks (variance) in their overall portfolio by increasing their holdings of emerging market assets up to 41 percent of their interna- tional holdings-compared with their current allocations of around 12 to 14 percent (figure 2.4, chapter 2). Although such a large reallocation represents an upper bound of possible gains ancL is not even feasible under present market conditions, it nonetheless illustrates the potential for further risk diversification.4 The demographic shift under way in industrial countries is an im- portant new factor that will provide further impetls to these underlying trends. Industrial countries now have a pronounced bulge in their de- mographic structures, reflecting the aging of the baby boom generation and declining birthrates (figure 2.13, chapter 2). This will lead to a steady rise in the proportion of elderly to active population in all indus- trial countries, although the pace of this increase is most pronounced in Japan. This is in sharp contrast to the situation in developing countries, whose clearly pyramidal structure reflects a much younger population. 20 THE MAIN There are three broad implications of this difference in demographic patterns. First, the aging of populations in industrial countries could lead to an increase in savings in the short to medium term. Second, aging and the associated slowing of labor force growth are likely, other things being equal, to exert downward pressure on the rate of return to capital relative to that of labor in industrial countries.5 Given the de- mographic structure prevalent in developing countries, the reverse can be expected there. Thus, differences in demographics are likely to rein- force the differentials in the expected rates of return to capital between industrial and developing countries. Both these factors should provide a stimulus for capital to flow to emerging markets. Third, and perhaps most important in terms of its impact, the aging of population in in- dustrial countries is leading to pressures for pension reform. The growing recognition of the burden that pension obligations under the current system will place on fiscal positions in industrial countries and the need to address the problem during the next 10 to 15 years is beginning to spur pressures to reform the existing pension sys- tems. In particular there are pressures to (a) move away from pay-as- you-go (PAYG) systems and toward more funded systems; (b) privatize pension schemes, through an increasing reliance on private employer and individual schemes-the so-called second and third pillars of old age security (World Bank 1994)-as well as through greater private sector management of public pension schemes; and (c) deregulate the investment allocations of pension funds to enable them to earn higher returns on their investments. All three factors are likely to result in greater response to investment opportunities in emerging markets. The United Kingdom and the United States have been at the fore- front of these shifts, but even in the United States, prospective changes in Social Security and corporate pension plans could have a large im- pact on asset allocation. Japan and most of continental Europe are also beginning to establish market-oriented schemes. For instance, in Japan, the government has begun over the past three years to implement a se- ries of reforms that is expected to lead to greater international diversifi- cation and greater demand for emerging market equities. Although the pace of change in individual countries is uncertain, the direction is clear. Global pension assets are likely to rise substantially and with a growing trend toward international diversification. Hold- ings of international equities, in particular, are expected to increase from their very low levels in most countries (figure 2.15, chapter 2). 21 ; `APITAL FLOWS TO DEVELOPING COUNTRIES Pension funds, therefore, are likely to be a major force in the demand for portfolio equities from developing countries. Today pension funds hold about $70 billion of emerging market assets. This amount could rise very considerably over the next decade. What do these trends imply for future flows to and financial inte- gration of emerging markets? * First, financial integration of developing countries is bound to deepen and broaden over the coming decade. Indeed, given the changes that are taking place at the international level, the pro- gressive financial integration of developing countries appears to be inevitable. * Gross private capital flows to developing countries will rise sub- stantially, with capital flowing not only from industrial to devel- oping countries but, increasingly, among developing countries themselves, and from developing to industrial countries. * There is likely to be considerable variation among countries, de- pending on the pace and depth of improvements in macroeco- nomic performance and creditworthiness. Indeed, in countries where economic and policy fundamentals are very weak, an ini- tial manifestation of growing financial integration may take the form of net outflows of private capital. * Fourth, given the growing importance of institutional investors, a larger proportion of these flows than in the past will be in the form of portfolio flows-especially equities. Benefits and Risks of Financial Integration L IKE FREE TRADE, FINANCIAL INTEGRATION HOLDS SUBSTAN- tial benefits for individual countries and for the global econo- my as a whole. And as in the case of trade, the gains for devel- oping countries are disproportionately large. Large, too, are the risks and potential costs that accompany financial integration. The Gains from Financial Integration The benefits of financial integration arise in two main ways. First, fi- nancial integration can boost growth, by raising the level of investment 22 T H E MtAIN- N and by improving the returns on investment, through knowledge spillover and market efficiency effects. Second, integration allows indi- viduals to insure themselves against adverse developments in their home economies by diversifying their assets and tapping global markets to smooth temporary declines in income. These effects are summarized below and are considered in detail in chapter 3. Financial integration can boost investment in developing countries by severing the link between local savings and investment. The potential gains from higher investment vary from country to country depending on the relative profitability of investment opportunities and on the dif- ference between the domestic and the world cost of capital before inte- gration. A notional indication of the magnitude of these gains can be seen from the following illustration. If capital inflows reach 3 to 4 percent of GDP (a typical figure for current large capital importers), the growth rate, based on typical capital shares and capital output ratios, would in- crease by about half a percentage point (Reisen forthcoming). This tem- porary growth effect is by no means negligible-an increase in the growth rate from 1.5 to 2 percent reduces the period required for output to double by 12 years-yet it is dwarfed by the range of growth rates among developing economies, showing that the successful tapping of for- eign capital markets is only one of many ingredients in a successful growth strategy. Integration may also boost growth by shifting the investment mix to- ward projects with higher expected returns because of the improved abil- ity to diversify the higher risks typically entailed in higher-return projects. The quantitative significance of a switch to higher-return investments in the wake of integration depends on three factors: the technological lag of a particular country, risk tolerance on the part of the country, and the speed with which existing capital can be reallocated. Simulation studies (Obstfeld 1992) suggest that the gains are very large. The size of gains re- flects the magic of compound interest: even small increases in growth have first-order-level effects over time. For the economies lagging farthest be- hind current best practice, the gains comprise both catch-up and the global efficiency gain, while the leading economies benefit from only the latter change. Yet even for these relatively advanced economies the bene- fits of higher output and hence consumption growth rates are very signif- icant: a well-managed integration is a win-win situation. A third and more immediate channel through which financial inte- gration can raise growth is through FDI. There is empirical evidence to 23 -,CAPITAL FLOWS TO DEVELOPING COUNTRIES suggest that a dollar of FDI raises the sum of domestic and foreign in- vestment by more than a dollar; thus FDI complements rather than sub- stitutes for domestic investment. More important, cross-country evidence suggest that FDI is substantially more efficient than domestic investment. Overall, the growth benefits come primarily through this latter channel, the quality rather than the quantity of investment. Integration enhances the depth and efficiency of the domestic fi- nancial system, with important positive feedback to investment and growth. Greater openness of the banking system can yield significant efficiency gains, even for industrial countries. And the gains from knowledge spillovers, deepening, enhanced competition, and the stim- ulus given to improvements in the institutional and supervisory frame- work are likely to be much greater for developing countries. Even more than in banking, foreign investment has been a driving force of change in the development of capital markets in developing countries. Most directly, financial integration enhances the role of cap- ital markets through its effects on depth and liquicdity. This positive as- sociation can be clearly seen in figure 3.2, chapter 3, which shows that countries that have received the largest portfolio inflows have experi- enced the largest increase in market capitalization. The scope for fur- ther improvements in the efficiency and functionin-ig of capital markets in many emerging markets is very large. Bid-offer spreads are much higher in developing countries than in industrial countries, suggesting that these indirect benefits associated with financial integration could go substantially farther in reducing the cost of capital to firms and re- ducing risks to investors. In addition to these spillover effects on the banking system and capi- tal markets, integration can promote better macroeconomic policies. Al- though integration increases the costs of policy mistakes in the short term, and increases the constraints on the conduct of macroeconomic policy, the market discipline that comes with integration can be a power- ful force in promoting prudent and stable macroeconomic policies, with large benefits over the longer term. For instance, Indonesia's decision to open its capital account almost three decades ago has been an important element underpinning its track record of prudent arid responsive macro- economic policies. Apart from boosting growth through these direct and indirect chan- nels, the increased opportunities for risk diversification with integration can bring gains in the form of higher and less variable consumption. In- 24 THE MAIN- tegration reduces the volatility of consumption by allowing a better di- versification of portfolios and by permitting international borrowing and lending to offset temporary income movements. Integration can also boost consumption by permitting a shift toward a portfolio with higher expected returns-the mirror image of the shift toward invest- ment projects with higher expected returns. The empirical evidence sug- gests that while there is substantial scope for risk diversification, the primary gains do not come through the reduction of risk per se, but rather through the ability to raise average portfolio returns (and hence consumption growth rates) as a consequence of diversification. The Risks of Financial Integration Despite the large potential benefits, there is widespread concern among policymakers that growing financial integration and increased reliance on private capital flows might render emerging markets more suscepti- ble to volatility-including large reversals in capital flows. The decline in portfolio flows, following the interest rate increases in early 1994 and the Mexican peso crisis, was seen as a stark reminder of the poten- tial for such volatility. Such fears are not without foundation, especially for countries with weak fiscal policies, badly managed or overprotected banking systems, and highly distorted domestic markets. International capital flows can act like a magnifying glass on the domestic economy, multiplying the benefits of well-structured reform programs but also increasing the costs of poor economic fundamentals and policies that are unsound. At the aggregate level, private capital flows to developing countries are less likely to suffer from major reversals. As noted, private capital flows to developing countries have demonstrated a remarkable degree of resilience. Portfolio flows have also shown an impressive recovery from the aftershocks of the Mexican crisis. Despite the large drop in the first half of 1995, portfolio flows recorded only a small decline for the year as a whole, and a significant increase in 1996 to levels that exceed the peak reached in 1993. The sustained increases in private capital flows in the face of recent shocks suggests that markets have entered a more mature phase (see World Bank 1997). Governments have demonstrated an awareness of and ability to respond promptly and aggressively to changes in market conditions. And markets have become more able to discriminate 25 KLCITAL FLOWS TO DEVELOPING COUNTRIES among countries on the basis of their underlying fundamentals. This does not mean that there will not be year-to-year fluctuations in private flows in response to changes in international financial conditions. But private flows to developing countries in the aggregate are less likely to suffer from a widespread or prolonged decline of the kind seen after the debt crisis. However, at the individual country level, volatility of flows and po- tential vulnerability to reversals remain a serious concern. There are three separate elements of this concern: large surges in the inflow of capital in the early stages of integration; susceptib ility to large reversals, as recently experienced by Mexico; and, more generally, the increase in volatility as a country becomes integrated. Surges. Virtually all major recipients of privatz capital flows saw a surge in private capital flows during the 1990s, although there has been considerable variation among countries in the timing, duration and magnitude of the surge (table 1.1). These surges have been extremely large in relation to the size of the economies; more than half of the countries shown in table 1.1 received annual inflows aver- aging more than 4 percent of GDP. Chile, Malaysia, Thailand, and until 1994, Mexico, experienced the earliest, and cumulatively among the largest, surges of private capital flows in the 1990s. At the other end, South Asian countries typically experienced the surge after 1992-generally in a smaller form. Eastern Europe also experienced the surge after 1992, but the magnitude of the inflows has been very large, even compared with flows to the largesr recipients in Latin America and East Asia. While low international interest rates were a common factor underpinning the surge in privalte capital to develop- ing countries during 1990-93, discrete changes in perceptions of creditworthiness and the resulting stock adjustment in international portfolios mean that surges in private capital are a likely feature of the early stages of integration. Reversals. With the growing importance of private capital flows has come the threat of major reversals of these flows. A number of coun- tries have seen such reversals (figure 1.5). Turkey and Venezuela were the first to experience major capital flow reversals in the 1990s, in both cases following a loss of investor confidence in government poli- cies. The reversal of private capital flows in the case of Mexico was notable in two respects: its size and its triggering of reversals in several other countries, most markedly in Argentina and Brazil. These 26 THE MA144 l- Table 1.1 Net Private Capital Inflows to 20 Developing Countries, 1990s (net long-term internationalprivate capital as a percentage of GDP) Cumulative Maximum inflowslGDP annual Country Inflow episodea at end of episode inflo w Argentina 1991-94 9.7 3.8 Brazil 1992-95 9.4 4.8 Chile 1989-95 25.8 8.6 Colombia 1992-95 16.2 6.2 Hungary 1993-95 41.5 18.4 India 1992-95 6.4 2.7 Indonesia 1990-95 8.3 3.6 Korea 1991-95 93 3,5 Malaysia 1989-95 45.8 23.2 Mexico 1989-94 27.1 8.5 Morocco 1990-95 18.3 5.0 Pakistan 1992-95 13.0 4.9 Peru 1990-95 30.4 10.8 Philppines 1989-95 23.1 7.9 Poland 1992-95 22.3 12.0 Sri Lanka 1991-95 22.6 8.2 Thailand 1988-95 51.5 12.3 Tunisia 1992-95 17.6 7.1 Turkey 1992-93 5.7 4.1 Venezuela 1992-93 5.4 3.3 a. The period during which the country experienced a significant surge in net private capital inflows. Source: World Bank data; IMF, World Econeomc Outlook data base; IMF, Internzaonal Financial Statistics data base. episodes of capital flow reversals in the 1 990s were typically not larger in absolute magnitude, or in relation to GNP, than similar episodes during the debt crisis. What is common to most of these reversals (except for Malaysia during 1993-94, where it was attributable to the authorities' attempts to restrain very large inflows) is that they were triggered by a lack of confidence in domestic macroeconomic policies. Volatility and herding. Apart from the vulnerability to major reversals, countries have been concerned about increased volatility associated with private capital flows, especially portfolio flows. Integration can give rise to greater volatility in two ways: first, it can expose econ- omies to new sources of shocks in the international economy, and sec- ond, it can magnify the effects of domestic shocks (figure 1.6). These 27 S:PAITAL FLOWS TO DEVELOPING COUNTRIES Figure 1.5 Large Reversals in Net Private Capital Flows (short-term and long-term internationalflows) Mexico, 1981-83 12 Mexico, 1993-95 6 Turkey, 1993-94 10 Argentina, 1982-83 20 Argentina, 1993-94 4 Malaysia, 1993-94 Venezuela, 1992-94 Venezuela, 1988-89 Chile, 1990-91 Several develloping countries have . experienced major reversals of Chile, 1981-83 private capital flows. 7 0 5 10 15 20 25 30 35 Billions of dollars Note The figure to the right of each bar indicates the magnitude of the reversal as a percent- age of GDP Source IMF, loternatoonal Financial Statistics data base; World Baiik data concerns become all the more important because the degree of policy autonomy declines with growing financial integrotion. The main international sources of volatility are changes in asset re- turns (interest rates and stock market returns) and potential investor herding and contagion effects. Shifts in asset returns affect capital flows by altering the fundamentals, whereas investor herding and contagion effects may change investment in a country even if fundamentals are un- changed. We find that although portfolio flows to emerging markets re- main sensitive to changes in international interest rates, there has been some decline in the degree of sensitivity as country-specific factors have become more important. There is also some evidence to suggest that emerging markets are prone to foreign investor herding, but the experi- ence of countries that began the process of integration some time ago suggests that the peak of such herding behavior is relatively short-lived. 28 THE M AI EIW - Figure 1.6 Sources of Volatility in Emerging Markets Characteristics of emerging Changes in Foreign markets that interact with and international investor magnify volatility interest ~behavior: rates anW herding and other asset contagion returns, and effects terms-of- trade Weaknesses in financial markets Factors in the international environment Weaknesses in capital market infrastructure Emerging markets are marginal in international VOLATILITY investors' Factors in the domestic portfolios environment Asymmetric information prob- lems between for- Domestic Domestic eign and domestic real policy investors Both domestic and international shocks shocks factors can precipitate volatility in an integrated market. Finally with regard to contagion, the experience of countries following the Mexican crisis suggests that pure contagion is a relatively brief phe- nomenon; international markets were able to differentiate among emerging markets, so that those countries with the strongest economic fundamentals saw a resumption in flows quite quickly. On the domestic side, developing countries are more susceptible to real and policy shocks than industrial countries, and these shocks, in turn, will result in greater volatility of capital flows and asset prices and returns in a more integrated environment. In addition, there arc also several characteristics of emerging markets that could possibly magnify international and domestic shocks in the early stages of integration. Fi- nancial and capital markets in developing countries suffer more from incomplete and asymmetric information, and from other institutional weaknesses, than do industrial country markets. In this environment, the potential for investor herding is greater, and domestic investors may be influenced by foreign investors, leading to even greater volatility. Emerging markets are also still quite marginal in international in- 29 CA1JTAL FLOWS TO DEVELOPING COUNTRIES vestors' portfolios, making them more susceptible to fluctuations in international financial conditions. All of this suggests that developing countries are likely to be subject to a greater degree of volatility in the early stages of financial integration. However, overall volatility of reserves and private flows has shown no systematic increase in the 1990s (figure 1.7). This is probably because developing countries' exposure and vulnerability to real shocks may have come down during this period. What is striking, however, are the significant differences in volatility between countries and types of flows (tables 1.2 and 1.3). Asia generally shows less volatility, and has reduced its volatility by more, than Latin America, suggesting that country con- ditions remain the primary determinant of volatility. Of course this does not necessarily suggest that the observed volatility of flows is detrimen- Figure 1.7 Changes in Volatility of Reserves, 1980-49 and 1992-96 (coefficient of variation based on quarterly data) 1980-89 1.0 0.8.E/ 3 decrease 0 0.6 < ~~~~~~~~~~~~iricrease 0.6 U~ 0 0.4 The volatility of reserves has not 0 increased systematically in the * 1990s. 0.2 ---- -------------- ------ * Latin America *Other L*Asia 0 0.2 0.4 0.6 0.8 1.0 1992-96 Source IMF, International Financial Statistics data base 30 THE MA Table 1.2 Volatility of Private Capital Flows from the United States, 1980s and 1990s (based on quarterly netflows) 1980s 1990s Billions of dollars Billions of dollars Average Standard Coeffiient of Average Standard Coeffi cient of level deviation variation level deviation variation Latin America Total private capital flows 4,177 8,013 1.92 6,685 13,943 2.09 Direct investment 421 1,495 3.55 3,216 2,084 0.65 Foreign securities -285 397 1.39 1,923 3,393 1.76 Nonbank -74 1,359 18.41 449 5,906 13.15 Bank 3,896 8,366 2.15 1,116 11,223 10.06 Africa andAsia Total private capital flows 1,031 2,558 2.48 3,550 5,285 1.49 Direct investment 411 816 1.99 1,570 875 0.56 Foreign securities -203 505 2.48 1,226 1,793 1.46 Nonbank -12 193 15.87 408 938 2.30 Bank 1,049 2,093 2.00 367 4,493 12.24 Source: U.S Department of Commerce, Survey of Current Business; U.S. Treasury Department data. tal in all cases, since this volatility could be in response to, and therefore could offset, real shocks such as a terms-of-trade deterioration. In terms of different types of flows, FDI is the least volatile. Portfolio flows show a higher degree of volatility, but the level of volatility has tended to come down. The volatility of bank flows has gone up relative to Table 1.3 Volatility of Capital Flows, Selected Countries, 1980s and 1990s (coefficient of variation of quarterly flows) Overall capital account Foreign direct investment Portfolio investment Country 1980s 1990s 1980s 1990s 1980s 1990s Argentina 1.28 1.09 1.01 1.26 43.60 2.96 Indonesia 0.93 0.70 0.82a 0.39 3.38 1.61 Korea 334.24 0.92 1.22 0.59 4.55 0.87 Mexico 3.85 0.66 0.47 0 62 3.34 1.25 Pakistan 0.71 0.65 0.68 0.30b 1.39 0.76 Philippines 1.21 0.75 1.79 0.99 0,68 0.30 Sri Lanka 0.55 0.61 0.36 0.76b 0.68 0.30 Thailand 0.78 0.33 0.83 0.23b 0.68 0.30 a. 1981-88. b. 1990-93. Source. IMF, International Finrancal Statisucs data base. 31 ?ITAL FLOWS TO DEVELOPING COUNTRIES the 1980s, in part because net flows have become relatively small. These differences can be explained by the fact that FDI investments are more costly to reverse, and hence are more responsive to Long-term fundamen- tals, than short-term fluctuations. Portfolio flows, on the other hand, are quite sensitive to short-term differentials in rates of return, and hence to changes in interest rates. Moreover, unlike FDI ir[vestors, portfolio in- vestors can divest themselves easily of their stocks of equities or bonds. Finally, it is worth noting that while volatility-meaning variation around an average level-has tended to come down, or at least not in- crease, in the majority of countries, the absolute magnitude of variation is now much greater, since the average level of flows is higher. Conse- quently, the potential impact of such volatility on] the domestic econ- omy is also much greater than it was during the 1980s. In sum, factors in the international environment do contribute to potential volatility in flows and asset prices in emerging markets. Since emerging markets are still quite marginal in foreign investors' portfo- lios, these markets remain quite susceptible to cychlcal conditions in in- dustrial countries. And since foreign investors are still relatively unfamiliar with emerging markets, these markets are also more prone to investor herding than are industrial countries. The main risks of volatility and large reversals, however, lie at the individual country level and come from the interaction of domestic policy and economic con- ditions with international factors. As argued below, most developing countries lock strong macroeco- nomic, banking sector, and institutional underpitinings, and hence re- main vulnerable to potential instability and reversals of flows. While international investors are becoming more discerning, market disci- pline tends to be much more stringent when investor confidence is lost-triggering large outflows-than during the buildup to a potential problem. Financial integration can therefore magnify shocks or the costs of policy mistakes, leading to greater instability. Policy Challenges and Emerging Lessons M OST DEVELOPING COUNTRIES ARE STILL A LONG WAY from establishing the rigorous preconditions needed for suc- cessful financial integration. The importance of sound macroeconomic policies as the key prerequisite for external financial 32 THE MAI.X liberalization has been recognized for some time but cannot be stressed enough. The need to eliminate large price distortions has also been recognized, leading to the recommendation that the current account should be liberalized before the capital account. There is now also growing awareness of the importance of financial sector preconditions and their interaction with the macroeconomic environment as a cen- tral element of successful integration. Developing countries also face a daunting institutional agenda of banking and capital market reform, since a much higher proportion of capital flows will be intermediated through these markets. All these are key elements of the broader poli- cy agenda that developing countries need to adopt in any case, and financial integration makes their pursuit more urgent. For many countries that are not attracting private capital flows, the main task is to make rapid progress in establishing these preconditions and eliminating other constraints that may be impeding foreign capital flows. The experience of Sub-Saharan Africa shows that countries which have made more progress in establishing stable macroeconomic conditions, adopting outward-oriented policies, and addressing other constraints to private investment are beginning to attract significant private capital flows (box 1.2). Since policies to attract private capital flows overlap with the broader agenda for private sector development, on which there is now broad consensus, we do not dwell on it in this study. Instead, the remaining sections focus on the challenges that de- veloping countries are likely to face in the early stages of financial inte- gration, given that it will take time to establish all the rigorous preconditions needed for financial integration. Macroeconomic Management in an Integrating World While increased financial integration may provide substantial macro- economic benefits for developing countries, the integration process also carries with it some difficult macroeconomic challenges. Policymakers in these countries have been concerned with three types of problems: * The potential for macroeconomic overheating, in the form of an excessive expansion of aggregate demand as a consequence of cap- ital inflows. * The potential vulnerablnity to abrupt and large reversals of capital flows because of changes in creditor perceptions. 33 -IdTAL FLOWS TO DEVELOPING COUNTRIES Box 1.2 What Is Constraining Foreign Capital Inflows to Sub"aharan Africa? THE RECENT SURGE IN INTERNATIONAL PRIVATE much higher in Sub-Saharan Afiica than in other capital Flows to developing countries has largely by- developing regions. During 1990--94, rates of return passed Sub-Saharan Africa (excluding South Africa), on FDI in the region averaged 24 to 30 percent, com- but the region's small share masks significant differ- pared with 16 to 18 percent for all[ developing coun- ences among countries and the types of capital flows tries. This suggests that risks are perceived to be they receive. higher in Sub-Saharan Africa than in other regions Countries with positive per capita growth (Ppc or that foreign investors face greater impediments countries) received larger flows than countries with there than elsewhere. negative per capita growth (NPc countries). Grow- Experience in other regions has shown that in- ing economies also showed an improving trend, es- vestors choose countries with stable political and pecially when private transfers are taken into economic environments. Open markets, minimal account. The CFA countries (the members of the regulation, good infrastructure f'acilities, and low African franc zone) suffered larger and more sus- production costs are also key factors in attracting tained declines in private flows during the 1980s and holding foreign investment. Bringing these fac- than d d non-CFA countries. In contrast, private tors together has proved difficult for many countries flows (including transfers) to the non-CFA countries in Sub-Saharan Africa. Specifically, the main con- recovered significantly in the second half of the straints that have impeded the flow of FDI to Sub- 1980s and increased further during 1993-95. Saharan Africa include the prevalence of civil strife, Underlying the aggregate trend in private flows macroeconomic instability, low economic growth to Sub-Saharan Africa are also quite marked differ- and the small size of domestic markers, inward ori- ences in trends for different types of flows. A de- entation and burdensome regulations, slow progress tailed examination of these different types of flows on privatization, poor infrastructure, and high wage yields irisights into the factors constraining private and production costs. Conversely, countries that capital inflows to Sub-Saharan Africa. have addressed these weaknesses have begun to see a Foreigl direct investme. FDI has shown a signifi- payoff in increased FDI. For instance, several coun- cant increase since the late 1980s for the non-CFA tries that have seen an end to civtl conflict (such as countries and the PPC countries in the factors con- Angola, Mozambique, Namibia, South Africa, and straining private capital inflows to Sub-Saharan Uganda) have benefited from large increases in PDI Africa. For some of these countries (such as Angola, inflows during the 1 990s. Similarly, countries that Botswania, Ghana, Mozambique, and Uganda), FDI have made progress in reducing macroeconomic in- as a percentage of GDP in 1994-95 compares favor- stability have enjoyed some success in attracting pri- ably with that of recipient countries in Asia and vate FDI inflows. Latin America. Private loans. Unlike other de-veloping regions, Despite these promising trends, most countries where commercial bank lending has shown a sharp in Sub-Saharan Africa have received very modest turnaround in the 1990s, such lending has remained amounts of FDr. This has been the case despite the either negative or at very low levels for most Sub- fact that rates of return on FDP have generally been Saharan African countries. In part, this has occurred 34 THE MAA1t because most African countries have not yet restored of international custodians are important elements their access to financial markets. In contrast to cred- of the financial infrastructure that is needed to at- itworthiness ratings in other regions, which have tract foreign investors. Many investors say that cor- shown a marked improvement in the 1 990s, credit- ruption in the public sector, including the judiciary, worthiness ratings for Sub-Saharan African coun- not only increases transaction costs but also as acts as tries have remained much lower, on average, and are a deterrent in its own right. The supply of assets is only just beginning to see some improvement. The still very limited, and in addition to the public com- main factors believed to have contributed to Sub- panies already listed on stock exchanges, the number Saharan countries' generally low levels of creditwor- of private firms listed needs to be increased. In some thiness are high political risk, weak growth and cases, privatization of public assets offers the best av- export performance, macroeconomic instability, enue for increasing the supply of assets in the econ- and high levels of indebtedness. omy and attracting foreign investors- Portfolho equity flows. Although portfolio invest- Challenges ahead. With many Asian and Latin ment flows to Africa (with the notable exception of American countries growing rapidly and moving far South Africa) are still extremely small compared with ahead of most African countries in terms of putting flows to other emerging markets, there are encourag- in place the financial infrastructure needed to absorb ing signs of growing investor interest. Since 1994, foreign capital efficiently, most African countries more than 12 Africa-oriented funds, with a total size of will have to undertake speedy policy and structural more than $1 billion, have been set up. Initially, the reforms to attract private flows. Market discipline is focus of these funds was primarily the South African likely to be severe in the initial stages, and countries market, but the base has been broadening to include a that backtrack on reform will find that their access growing (though still limited) number of other to international capital is lirnited and that such cap- African countries, including Botswana, C6te d'Ivoire, ital will be provided on costlier terms. Ghana, Kenya, Mauritius, Zambia, and Zimbabwe. While microeconomic factors-poor infrastruc- This growing pool of portfolio investment is already tures, shaky banking systems, thin financial markets, perceived as bringing important benefits, including weak regulatory frameworks, dearth of human capi- improved liquidity, greater incentives for privatiza- tal, slow pace of privatization, and corruption-are tion, and increased pressure for policy reforms and im- difficult to quantify, the macroeconomic factors provement of the financial infrastructure. used in the empirical analysis we have carried out A nunber of factors are, however, still seen as (see Bhattacharya, Montiel, and Sharma 1996) constraining portfolio investment: investors view yielded clear-cut conclusions. In Sub-Saharan political instability and weak macroeconomic fun- Africa, economic characteristics like output growth, damentals as the most important impediments. openness, relative stability of real effective exchange They also see many structural weaknesses that in- rates, low external debt, and high investment rates hibit investment and believe that a reduction in have encouraged private capital flows. The first three transaction costs is critical. The establishment of an of these have been crucial in attracting FD}, the last efficient securities trading system and the presence two in attracting foreign private loans. 35 +CAPTAL FLOWS TO DEVELOPING COUNTRIES * The more general, longer-term implications of financial integra- tion for the conduct of macroeconomic policy. As integration ad- vances further, policymakers will have to manage the increased macroeconomic volatility that may prevail when the economy is more exposed to external shocks. In addition, they will need to face these and other shocks with reduced policy autonomy. Policies to avert overheating. The initial manifestation of growing financial integration in many developing countries was a large one- Table 1.4 Surges in Private Capital Flows to 20 Countries: Policy Responses and Outcomes Response andoutome Magnitude of the surge (net private capital flows as percentage of GDP)a 2.8 3.0 5.8 4.8 14.8 1.8 1.8 2.3 Policy response Reduce net infldws of foreign exchange Controls on inflows x x x x x Liberalize capital outflows x x x x Trade liberalization x x x x x Reduce official borrowing x x Float/appreciate exchange rate x x x Reduce impact on monetary aggregates Sterilized intervention x x x x x x Higher reserve requirements x x x x Reduce impact on aggregate demand Fiscal contraction x x x x x Outromes Average annual GDP growth (percent) 9A 3.1 5.7 1.6 7.5 -0.7 2.2 -2.5 Consumption (percentage of GDP) 4.4 3.6 -8.5 4.1 6.4 -1.7 -5.2 1.1 Investmnent (percentage of GDP) 0.6 -2.0 10.2 0.9 1.6 -1.3 5.7 4.7 Currenrt account deficit (percentage of GDt) 1.8 0.6 -4.9 4.9 9.8 -1.2 0.2 5.0 Reservce .ccumuhation (percent) 23 75 -1,970 -22 29 1,820 80 -11 Average annual inflation (percent) -801.1 -93.5 -4.1 4.8 -5.5 0.1 1.3 0.8 Real exclange rate (percent) 91.7 7.4 -25.5 14.7 18.9 -30.8 -29.4 4.4 a. This is the annual average for the surge period b. All figures represent the change in the surge period compared with the preceding period of equal length, except reserve accumulation, which is expressed as a percentage of the capital account surplus. SQusc. eWrld Bank data. 36 THE MA1N Ij. way net flow of capital, magnified by favorable cyclical conditions in world financial markets. Since such flows have been large relative to the size of the economies of recipient countries (table 1.4), macroeco- nomic overheating has been a potentially serious problem. In contrast to what happened during the commercial lending boom of the 1980s, however, virtually all recipient countries in the 1990s have resisted allowing the current account deficit to widen by the full magnitude of the incipient inflows. Countries have taken three broad approaches to addressing the potential overheating problem: 9.4 5.3 3.5 3.6 6 4 4.3 6.5 5.5 9.9 5.1 3 2.7 x 4.0 2.9 -3.3 -2 3 3.3 2.2 8.5 2.0 3.9 U.5 1.4 -5.0 ^-1.8 6.7 0.8 -2.0 3.1 6.1 11.3 -L.9 -11.2 -1.4 -0.5 6.8 4.8 2.4 -1.1 1.0 -4.0 1.7 -11.1 2.2 13.4 2.6 1.3 6.8 2.9 7.1 0.1 0.9 1.4 0 7 3.9 -0.2 2.3 3.7 i.4 14.6 49 0 93 59 67 79 39 105 59 -28 31 -85 1.4 -74.4 0.1 1.7 -79.1 -3.1 -146.7 -2.2 -1.1 -2.2 5.0 -2.7 -24.5 20.0 -6.5 -9.0 120.9 -10.7 37.9 0.6 -18.9 0,6 1.0 9.8 37 'FN?TAL FLOWS TO DEVELOPING COUNTRIES * Reducing the net inflows offoreign exchange. Countries have de- ployed a number of policies toward this end, including controls on capital inflows, the liberalization of capital outflows, trade liberal- ization, reduced official borrowing, and appreciation of the nom- inal exchange rate. The majority of countries attempting to cope with the effects of integration have eschewed direct controls on capital inflows, but a significant minority have continued to main- tain capital account restrictions, and new capital controls in some form have been adopted by several countries, in particular those receiving the largest flows (for example, Brazil, Chile, Malaysia, and Thailand) or those that were constrained in their ability to use other instruments to contain overheating (such as Indonesia). No country made wholesale changes in its foreign exchange regime. Bands were introduced by Chile, Colombia, and Mex- ico, but only those of Chile and Colombia seemed designed to accommodate some nominal appreciation in response to capital inflows. More recently, some of the Southeast Asian economies, notably Indonesia, Malaysia, and Thailand, have moved toward greater flexibility in their exchange rate regirnes but have resisted significant nominal exchange rate appreciation. Among the other measures listed, reduced official borrowing, trade liberal- ization, and liberalization of capital outflows were adopted by several countries during the surge period, but these reforms have been of lesser importance. The liberalization of capital outflows may, in some cases, actually have abetted inflows by enhancing credibility. * Reducing the impact of reserve accumulation on monetary aggregates. Since most countries have maintained an officially determined exchange rate, the central banks have had tc, intervene to defend the parity in the face of large inflows. Virtually all recipient coun- tries, therefore, have had to contend with the monetary impact of a substantial buildup of external reserves. In order to curb the cre- ation of high-powered money, virtually all countries undertook sterilized intervention in the foreign exchange market. A few countries (Chile, Colombia, the Philippines, and Sri Lanka) also raised reserve requirements to reduce the potential multiplier ef- fect on the expansion of broader monetary aggregates. * Reducing the impact of inflows on aggregate demand. In addition to tightening monetary policy, a majority of countries appear to have 38 THE MAIN H tightened fiscal policy in response to inflows, but to very different degrees. In Latin America, Argentina, Chile, and Colombia relied most strongly on fiscal adjustment, whereas virtually all East Asian countries used the tightening of fiscal policy as a cornerstone of their macroeconomic policy response to capital inflows. Given the scale of the inflows and the accumulation of reserves in the 1990s, most recipient countries have been remarkably successful in curbing inflationary pressures. A majority of countries managed to re- duce inflation during this period, and where there was an increase, it was relatively modest. This reasonably favorable outcome was achieved despite an acceleration in GDP growth in most of these countries. The current account deficit widened in most countries in the post- surge period, as was to be expected, but there were substantial differ- ences in the magnitude of these increases. Hungary, Mexico, and Venezuela saw the largest increases in the current account deficit as a ratio to GDP, whereas Chile actually saw an improvement in its current account balance, with other countries falling in between. There were also substantial differences in the composition of absorption. For some countries, notably Chile and the East Asian countries, a widening in the current deficit (where it occurred) was primarily reflected in an in- crease in investment. This orientation toward investment was associ- ated in these countries with faster economic growth. In contrast, the composition of absorption was more skewed toward consumption in Argentina, Colombia, Hungary, Mexico, Peru, and Poland. This pat- tern of absorption was strongly correlated with the degree of real exchange rate appreciation. In particular, an increase in the consump- tion-to-GDP ratio was associated with real appreciation (figure 4.3, chapter 4). The differences in the composition of absorption and the apprecia- tion of the real exchange rate between these two groups of countries appear to have been due largely to fiscal policy. Countries that signifi- cantly tightened fiscal policy-including Chile, Indonesia, Malaysia, and Thailand-also reduced their share of consumption in GDP during the inflow period. Moreover, all of these countries actually achieved a substantial depreciation in the real exchange rate during the inflow pe- riod. Chile and Indonesia also curbed their current account deficits. While the current account deficit increased in Malaysia and Thailand, these two countries also experienced very substantial rises in the share 39 -AZEECAPITAL FLOWS TO DEVELOPING COUNTRIES of investment in GDP, as well as in their rates of economic growth, in association with the arrival of capital inflows. In sum, developing countries have been able to date to counter most of the symptoms of overheating by adopting a wide range of policies. There have been some common elements in their policies, as well as some important differences that have produced different macroeco- nomic outcomes. In particular: * Several major recipients have employed new capital controls in the face of large inflows, with at least some degree of effectiveness in reducing the magnitude of inflows and altering their composition. * Sterilized intervention was used by virtually all countries to offset the impact of reserve accumulation on monetary aggregates. Ster- ilization appears to have been generally successful in curbing the growth of base money, but less effective in preventing asset infla- tion. Beyond this, there were significant differences in the policy mix. • At one end were a group of countries that resisted real exchange rate appreciation and, toward this end, placed greater emphasis on fiscal tightening. These countries were rot only able to avoid substantial real appreciation but tended on average to have lower current account deficits, a mix of absorption more oriented to- ward investment, and faster economic growth. • At the other end were a group of countries that used the nominal exchange rate as a nominal anchor, with relatively greater reliance on monetary rather than fiscal policy. These countries typically experienced both consumption booms and larger real exchange rate appreciation. The major policy lesson to be derived from cross-country experience in dealing with surges in capital inflows is that the policy mix employed to combat overheating also has a major effect on the performance of the real economy and its ability to benefit in the lo rig term from capital flows. In particular, a heavy reliance on fiscal policy, supported by ster- ilization and some nominal exchange rate flexibility-and in more ex- treme cases by temporary taxes or controls on inflows-can be an effective response to overheating and can also improve the balance be- tween domestic investment and consumption and reduce the risks of an overappreciated real exchange rate and an unsustainable deficit in the current account. 40 THE MAIN TB 1 Policies to limit vulnerability. In addition to responding in a particular way to inflows of private capital, a country can be more or less vulner- able to large reversals of private capital flows, depending on its under- lying economic and political conditions. The factors affecting investor confidence, and hence the creditworthiness of countries in the new international setting, are not yet well understood, in part because investors are still in the process of discovering emerging markets. This study uses the Mexican crisis to examine the relationship be- tween macroeconomic performance and vulnerability, by assessing how the incidence and severity of "the tequila effect" (contagion) emanating from Mexico was linked to macroeconomic performance in other de- veloping countries. Our key findings are that some of the initial inter- pretations of the crisis, which attributed Mexico's vulnerability solely to its high current account deficit, are not well supported by the cross- country evidence. Instead, as shown in table 1.5, the countries that were less vulnerable to the Mexican crisis tended to be those that had avoided real appreciation and imposed a tight fiscal policy during their inflow period, thereby achieving a larger share of investment and more rapid growth than they would have done under different policies. The Mexican crisis and its effect on other countries also highlight at least Table 1.5 Impact of the Mexican Crisis on Selected Developing Countries Indicator More affected group,a ess affected groupb Change in stock pricesc -22.5 -3.5 Current account balance (percentage of GtYP)d -3.5 -3.2 FDiltotal private flowae 39.8 36.3 GDP growth (percent)e 3.4 7.7 Inflation (percent)e 361.2 11.4 Central government balance (percentage of GDP)e -4.1 0.6 Debtfexporrsd 273.8 123.5 Debt/reservesd 879.0 296.8 a Countries in this group indude Argentina, Brazil, India, Mexico, Pakistan, Turkey, and Venezuela. b. Countries in this group include Chile, Colombia, Indonesia, Korea, Malaysia. and Thailand. c. Between December 1994 and the end of March 1995. d. 1993. e. Average for 198-93 period. Source: IMF, World Economtc Outlok data base. World Bank data. 41 TTtAL FLOWS TO DEVELOPING COUNTRIES four important conditions in the initial environrnent that could affect subsequent vulnerability: * Inflation and inflationary expectations at the time of integration can constrain the choice of policy mix. In particular, concerns about inflation often have led governments to adopt a nominal exchange rate anchor, which in almost all cases has led to sub- stantial real exchange rate appreciation. * The magnitude of external and domestic debt has a direct bearing on investor confidence and also affects the room for fiscal ma- neuvering and monetary tightening in the face of balance-of-pay- ments pressures. * As the Mexican crisis has so forcefully demonstrated, the matu- rity and composition of public debt is as important as the size of the debt. The inability to roll over debt couLd lead to large capital outflows, especially when a country is vulnerable to shocks, so excessive reliance on short-term debt is risky. The currency com- position of debt also matters, because with foreign-currency- denominated debt, the government assumes the risk of an exchange rate devaluation. * The health and resilience of the banking system are crucial to en- suring the sound use of capital inflows and avoiding vulnerability, as discussed below. A weak banking system can increase vulnera- bility through too much and too risky domestic lending and by severely limiting the authorities' ability to raise interest rates in the face of a loss of reserves-as happened in Mexico in early 1994. The most important lesson for policymakers is that creditors (both domestic and foreign) can be expected to take their capital out of a country when they think that a policy change could impair the value of their investments. Thus, vulnerability will arise when the perception is created that a devaluation, nonpayment of public sector debt, or the imposition of restrictions on capital outflows is about to occur. Such expectations are likely to arise when the real exchange rate is perceived to be out of line, the government's debt obligations are large in relation to its earnings capacity and external reserve position, fiscal adjustment is perceived as politically or administratively infeasible, or the country's growth prospects are bleak. From the perspective of creditors, there- fore, a high share of investment in absorption and a strong record of 42 THE MAIN growth, a low stock of government obligations coupled with demon- strated fiscal flexibility (in the form of small deficits and low inflation), and a real exchange rate broadly perceived to be in line with funda- mentals all augur well for future debt service. From a policy perspec- tive, countries need to have an exchange rate policy that avoids substantial appreciation of the real exchange rate and responsible fiscal policies. For countries that are vulnerable to shocks, it is particularly important to have fiscal flexibility as well as a debt structure with longer maturity and an adequate cushion of reserves. A final macroeconomic policy lesson is that governments need to move forcefully in the face of shocks and loss of investor confidence. Governments therefore should monitor warning signals and take prompt measures to prevent a large loss of reserves; these measures include tight- ening monetary policy and adjusting the exchange rate. Countries with nominal exchange rate anchors face special constraints. They must either take draconian fiscal measures or allow interest rates to climb to very high levels, with adverse impact on the budget and investment. Macroeconomic management in the long term. The two macroeconomic challenges of overheating and vulnerability are likely to be particularly important during the transition to financial integration, when inter- national investors are adjusting their portfolios and when policy credi- bility is not yet well established in the capital-importing countries. However, new macroeconomic challenges will continue to arise even when the process of integration is well advanced. In particular, finan- cially integrated developing countries will find themselves operating in a very different macroeconomic environment-one in which capi- tal movements are highly sensitive to changes in prospective foreign and domestic rates of return. The macroeconomic consequences of domestic shocks (including changes in policies) will be altered, and the country will also be faced with new types of shocks arising in international financial markets. This new environment may thus be characterized by ongoing volatility, to which policymakers will need to respond, but with a re- duced degree of policy autonomy. Developing country experience to date provides only limited insight about these longer-term implications for macroeconomic management, since most countries are still in the early stages of financial integration. There is, however, a growing body of analysis and empirical evidence from more integrated economies that provides relevant lessons for developing country policymakers. 43 Ve :APiTAL FLOWS TO DEVELOPING COUNTRIES First and foremost is the importance of respond ible and flexible fiscal management. Financial integration places a premium on short-run fis- cal flexibility, which provides an effective instrument for stabilizing do- mestic aggregate demand in response to external shocks in an environment where it is increasingly difficult to restrict capital move- ments and hence rely on monetary policy. Fiscal policy can also substi- tute for exchange rate policy as a stabilizing device. Equally important, a stable perception of fiscal solvency becomes crucial for preventing the domestic public sector itself from becoming an important source of macroeconomic shocks (through perceived changes in solvency) and for providing the latitude for fiscal policy to play a short-run stabilizing role. The second important policy lesson is that exchange rate management re- mains possible and that crawling pegs, in particular, remain an exchange rate option for developing countries that become closely integrated with world capital markets. However, the scope for deviation from fun- damentals, as well as from commitment to the announced regime, is much reduced under these circumstances. Managed rates are certainly capable of generating macroeconomic volatility in this environment, but they generally have not done so. Moreover, an active exchange rate policy does not imply a loss of control over the domestic price level. Countries that have managed the exchange rate with the objective of maintaining competitiveness-and thereby achieving stable or depreci- ating real exchange rates-have not exhibited cleteriorating inflation performance in the context of capital inflows. Third, the ability to conduct independent monetary policy becomes in- creasingly difficult if the government wishes to target the nominal exchange rate. Nevertheless, the experience to date suggests that there is scope for sterilized intervention as a short-run stabilizing tool despite increased financial integration. However, sterilization does carry a fiscal cost, es- pecially where credibility is not well established. Fourth, the evidence on capital controls suggests that they have been able, temporarily, to affect both the level and the composition of capital in- flows in the early stages of integration. In the early stages of integration, recourse to capital controls may be warranted for two reasons. The first is on the grounds of prudential restrictions on external borrowing and lending by domestic financial institutions, especially when it is not pos- sible to remove underlying financial sector distortions that lead to ex- plicit or implicit guarantees. The second is as a transitional device to 44 THE MAIN FI l preserve monetary autonomy pending the reforms required to permit the use of fiscal policy as a stabilization instrument. But experience suggests that capital controls become progressively less effective as a country becomes more integrated and that they have not been able to prevent either large capital outflows or inflows when there is the po- tential for significant arbitrage profits. Managing the Financial Sector in a More Integrated World A country's ability to maintain the health of its banking system and capital markets will determine, to a large extent, whether it will be able to realize the benefits of financial integration and avoid its pitfalls. Fi- nancial integration makes it imperative for developing countries to move much more aggressively to reform the domestic financial sector, because risks and losses are potentially greater in an integrated environ- ment. However, because of the time needed for such reforms, and the high costs of financial distress, there is a strong case for trying to con- tain lending booms associated with the early stages of financial integra- tion when the banking sector is weak and for establishing a mechanism for crisis management. Banking system reform. The risks of distress in the banking system increase with financial integration in three ways. First, increased for- eign competition can reduce the franchise value of banks sharply in the short run, inducing domestic banks to invest in a risky manner. Second, banks are able to expand lending much more quickly and thereby incur risks in their portfolios that they may not be able to manage. Third, circumstances can change much more quickly with financial integration, creating macroeconomic and financial sector distress that carries high economic and social costs. In addition to the direct costs of bailing out failed banks, the sudden loss of liquidity in the banking sector can amplify economic downturns. Such distress can severely set back reforms, as was the case with several Latin American countries following the debt crisis of 1982. Recent difficulties faced by the banking systems in Argentina and Mexico have highlighted the importance of a sound banking system in ensuring macroeconomic and financial stability in a more integrated en- vironment. In this regard two contrasting facts stand out. First, banking systems in developing countries play an even more dominant role in fi- nancial intermediation than in industrial countries (figure 1.8). Second, 45 X-APITAL FLOWS TO DEVELOPING COUNTRIES Figure 1.8 Banks' Share of Financial internediation, Selected Developing and Industrial Countries, 1994 (assets of the banking sector as a percentage of assets of allfinancial institutions) 100 E Developing countries p 80I taIndustrial countries i 60 40 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~10 Note~ ~ ~ ~ ~~~poe cosierbl alnge tahat caia inlw ilb nfiinl n lareroleinfinnciliteneiaton teriesd(igured (directly many instanes,y ndta the banking systemshv onlyle tha baks n idutril cunties cntly beetina dreguate, therb inotivel framewrkming dsothed bnftsowar ecessital rskw btaigbnkar po orlalymamym caializeconmi and th frdninan- reglaetionrn sulnerabisioncpbltithvyotytbe.etbihd Increased4openness6of the financial system si 20 T~~~~~osscnieal dne htcpta nlw ilbe inefcenl n termediated(directly ad indirecty) and thatthe bankingsystem wil incur additioal risks, threby not onlyunder55i5nin th eeiso capital inflows but potentially magnifying macroeconomicadfn - cial sector vulnerability.~~~~~~~~~~~~~~~55 46~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~55 THE MAIW` Figure 1.9 Health of Banking Systems (percentage of total rated banks) 30 25- 20- 10 5 Banks in developing countries are [O Z ~ ~~~~~~~~~~~~~~~~~generally weaker than banks in industrial countries. A B+ B C+ C D+ D E+ E U Industrial countries D Developing countries Source Moody s Bank Financial Strength Ratingsfor Selected Countries, May 1996. Since the process of financial integration is usually associated in its early stages with economic reforms and improvement in prospects, it is often accompanied by a surge in capital inflows. The resulting increase in aggregate demand, rise in asset prices, and easier access to funds can trigger a rapid expansion in bank credit. In turn, the surge in bank credit can reinforce these effects, leading to potential macroeconomic and financial sector vulnerability. This potential vicious cycle occurs because the increase in bank lending finances an increase in aggregate expenditures, accelerating output growth. The increase in output growth, in turn, reinforces economic agents' expectations and induces both firms and households to borrow and spend more. The increase in aggregate expenditures also appreciates the real exchange rate, aggravat- ing macroeconomic vulnerability. Meanwhile, domestic residents' nominal wealth rises because of the appreciation of the real exchange rate and the increase in asset prices. The increase in nominal wealth adds to the surge in households' expenditures. 47 A TAL FLOWS TO DEVELOPING COUNTRIES A weak and inadequately reguLated banking sectoi aggravates this process by lending for consumption and speculative purposes, such as portfolio in- vestment, or by financing a construction boom. By granting more credit for portfolio investments or real estate, the banking sector helps to push asset prices upward, increasing the value of nominal weaLth. This can be a self- perpetuating process if banks use inflated assets as co]lateral in granting new credits. Consequently bank portfolios may become increasingly exposed to sectors prone to boom-bust cycles, such as real estat,z, and may suffer large losses in case of a deflation in asset prices. This increased exposure to fluctu- ations in asset pnces, or to sectors highly sensitive to variations in interest rates, increases the vulnerability of the banking sector as a whole. The fragility of the banking system can also become an important constraint for macroeconomic policy. For instance, in both the Mexi- can and the Venezuelan financial crises, authorities were reluctant to raise interest rates in the face of a loss of international reserves, because of concerns about the effects of such action on the health of the banking system. A weak banking sector can also amplify economic downturns as banks are forced to caLl in loans or widen spreads to cover losses. In the extreme case of a banking crisis, there may be large fiscal repercussions, since the costs are typicaLly absorbed by the public sector. Unsettled macroeconomic conditions, in turn, can have severe repercussions on the domestic banking system. Developing countries are more prone to suffer from large macroeconornic swings, given that they are more susceptible to real and financial shocks and that these shocks tend to be large in relation to the size of thelir economies. But in- appropriate macroeconomic management and pclicy mix can also con- tribute to macroeconomic and financial sector vulnerability. For example, an overvalued exchange rate that raises the expectation of de- valuation can undermine the health of the banking system by leading to high real interest rates and an increase in the stock of nonperform- ing loans, as was the case in Chile and Mexico in the early 1980s. Sim- ilarly, excessive reliance on monetary restraint in the implementation of a stabilization program can also adversely affect the banking system. Fiscal policy also plays a critical role, not only in ensuring macroeco- nomic resilience but in avoiding excessive taxation of forced lending by the banking system. The importance of macroeconomic conditions for the banking sector is underscored by the fact that most major banking crises have been preceded by a deterioration in the macroeconomic en- vironment (Kaminsky and Reinhart 1996). 48 THE MAII There is growing recognition that it is the dynamic interactions be- tween macroeconomic conditions and the banking sector that lead to boom-bust cycles and the risks of macroeconomic and financial sector vulnerability. These risks increase in an integrating environment, since, as noted, banks can expand lending more quickly and in larger amounts, circumstances can change more swiftly, and markets react faster. The likelihood that the process of financial integration in its early stages will lead to a vicious cycle depends on the initial conditions in the recipient country and the policy response to the surge in inflows. Weak initial conditions in the banking sector, including the regulatory framework, and an accommodating macroeconomic policy stance in- crease the probability that macroeconomic and financial sector vulner- ability will increase with the surge in inflows (figure 1.10). An important challenge facing developing country policymakers, there- fore, is how to manage the joint process of external financial liberaliza- tion and domestic financial sector reform, especially where macroeconomic conditions are still unsettled. To distill lessons for banking sector reform in the face of growing fi- nancial integration, this study draws on a number of different country episodes during the 1980s and 1990s, when countries received sub- stantial capital inflows as a result of integration.6 A number of impor- tant conclusions emerge from our review of country experience: Figure 1.10 Capital Flows, Lending Booms, and Potential Vulnerability Capital inflows Maciroeconomic WelhicessBankinig sector tar 1 e M.acroeconomic Lendin vulnerability boomr Asset prices and real Booms in Capital inflows can lead to a vicious exchange rate consumption and appreciate Bank portfolios real estate circle that increases economic vulnerabilities. 49 kTITAL FLOWS TO DEVELOPING COUNTRIES First, countries that received substantial capital inflows as part of the process offinancial integration typically experienced lending booms. For the country episodes analyzed, bank lending to the private sector as a share of GDP increased from an average of about 29 percent during the years prior to the inflows, to an average of about 41 percent during the years of large capital inflows (figure 5. 1, chapter 5). These lending booms are typically associated with an expansion in aggregate demand, significant asset infla- tion, and, in several cases, an appreciation of the real exchange rate. Second, countries that had the largest lending bo6ms typically saw a sig- nfficant increase in macroeconomic vulnerability, measured by a widen- ing of the current account deficit, above what would be expected on the basis of the size of the inflows; a consumption boom relatively larger than what would be expected given the size of thc inflows; and a lower level of investment than would be expected given the size of flows.7 It is striking that every one of these country episodes with symptoms of increased macroeconomic vulnerability ended in a banking crisis. Third, lending booms are often associated with an increase in financial sector vulnerability, despite the fact that booms tend to improve bank profitability and hence offer the opportunity for [he banking sector to strengthen its resilience to shocks. Yet in the majority of country episodes, banks showed a deteriorating shock absoi ption capacity during the lending boom period. Countries that did not improve conditions in their banking sectors during the inflow period, but rather allowed bank lending to increase without addressing underlying weaknesses, generally experienced a banking crisis later on. Only in three of the country episodes, namely, Chile, Colombia, and Malaysia during the early 1990s, do we find that banks consistendy improved their resilience to shocks as measured by, for example, improved liquidity, higher capitalization rates and loan loss provisions, lower stock of nonperforming assets, and re- duced exposure to foreign exchange risk. Other Latin American coun- tries during the early 1990s, namely, Argentina Brazil, Mexico, and Venezuela, have shown an overall increase in banking sector vulnerability, while Indonesia, the Philippines, and Thailand have shown a deteriora- tion in some indicators and an improvement in others. Fourth, all of the countries in which lending booms were associated with an increase in financial sector vulnerability, and in which macroeconomic vulnerability increased or remained constant, experienced banking crises (figure 5.9, chapter 5). Countries with the largest lending booms were typically the ones that saw the most significant increase in vulnerability, 50 T HE MAII especially of the financial sector. In contrast, countries that took action to restrain the lending boom, or that pursued prudent macroeconomic policies, fared well and avoided major crises. Furthermore, countries that improved their macroeconomic stance-such as Indonesia, the Philippines, and Thailand-managed to avoid or at least delay banking crises despite some worsening in banking sector vulnerability. This review of country experience, therefore, suggests that lending booms in the early stages of financial integration are often a manifesta- tion of underlying weaknesses in the banking system that, if unchecked, can lead to increased macroeconomic and financial sector vulnerability. These vulnerabilities, in turn, negatively reinforce one another and in al- most all cases lead to a costly banking crisis. A strong macroeconomic policy stance can help, at least temporarily, to avert a banking crisis, even with a lending boom and weaknesses in the financial sector. Based on these findings, the report draws the following conclusions for managing the banking sector in a more integrated economy: Extremely weak initial conditions in the macroeconomy and the finan- cial sector may warrant a cautious approach toward externalfinancial lib- eralization. As the recent experience in Venezuela vividly illustrates, weaknesses in the banking sector combined with poor macroeconomic fundamentals can, in an integrated environment, quickly lead to a very costly banking crisis (exceeding 10 percent of GDP). Financial integration puts a premium on vigorous reform of thefinan- cial sector. Although there is still some debate about the sequencing of financial sector reform, there is broad consensus about its urgency in a more integrated setting and about the importance of undertaking a comprehensive reform of three main elements: financial infrastructure and institution building, the incentive structure, and the regulatory and supervisory framework.8 The main institutional building blocks are sound accounting stan- dards and practices, auditing, improved disclosure, and the legal framework. But special attention must also be devoted to bankers' management and risk assessment skills. These skills are often eroded after domestic financial liberalization but become all the more impor- tant in an integrated environment. The incentive structure becomes particularly important to discour- age excessive risk taking by bank owners and managers, and more broadly to encourage sound bank practices and close monitoring by large depositors. For instance, as shown by the Chilean experience of 51 TTAL FLOWS TO DEVELOPING COUNTRIES the early 1980s, and more recently by Mexico, explicit guarantees for foreign exchange can induce banks to increase exposure to foreign ex- change risk. Another common distortion comes from explicit or im- plicit deposit insurance, which can lead to excessive risk taking by depositors and banks. Speedy elimination of these distortions will en- sure that market discipline complements the monitoring of banks by domestic supervisory agencies, which may often be imperfect in the early stages of financial integration. The regulatory and supervisory framework needs to address five as- pects that are particularly important with financial liberalization: (a) strengthening banks' profitability and capitalization, (b) restricting connected lending, (c) improving public disclosure, (d) careful screen- ing of bank applicants, and (e) tightening asset classification and loan loss provisions. Virtually all developing countries have now adopted capital requirements that meet or even exceed the minimum standards under the Basle Accords. However, most developing countries have ra- tios in the same range as industrial countries, despite facing a much higher degree of risk and a more volatile environment (Goldstein and Turner 1996, Hausmann and Gavin 1995). Sustained efforts to strengthen supervision along these lines will not only reduce the possi- bilities for unsound lending but will also help build a shock absorber in the banking system so it can weather macroeconomic swings that may be associated with increased financial openness. In regard to the timing of reforms, the inflow period is likely to be associated with lower interest rates and an upturn in growth, and there- fore is an opportune time to push ahead with financial sector reform. A strong case can be made for containing the lending booms associated with the early stages offinancial integration. Lending booms can be curbed through a combination of macroeconomic and banking sector policies. As already noted, a combination of tighE fiscal policy and an exchange rate policy directed toward maintaining competitiveness has proved to be the most effective means of restrainin-ig aggregate demand and shifting its composition away from consumption and toward in- vestment, particularly in the tradable sector. While monetary policy can play an important complementary role in targeting inflation, a tight monetary policy can induce additional capii:al inflows, especially where fiscal policy is less stringent. Possible policies targeted at the banking sector to contain the lend- ing boom include (a) increasing banks' capitalization requirements, 52 THE MkAT, weighted by the risk of bank assets, as in Chile; (b) raising provisions against loan losses, as in Chile in the 1980s and Mexico more recently; (c) raising reserve requirements, as in Malaysia; (d) imposing ceilings on commercial bank lending; and (e) imposing restrictions on com- mercial banks' external borrowing, as in Indonesia and Thailand. The first two policies can be implemented more easily during a lend- ing boom, when banks' profits increase, but in order to be effective they need to be implemented within a consistent incentive and regulatory framework. This is so because bank managers tend to hide their financial problems by underprovisioning nonperforming loans-and, therefore, overstating their banks' capital-especially during periods of financial distress. The last three policies have been used very effectively to contain lending booms, but by implicitly taxing the banking system, they intro- duce distortions that lead to disintermediation in the medium term. Countries should establish mechanisms that enable policymakers to deal with a banking crisis promptly and effectively. Delaying actions intended to contain the crisis will only increase its cost. Many countries start reshaping their institutional framework-a lengthy and difficult process-at the same time that they are integrating with the rest of the world. During this period, domestic financial markets will experience strains that, if not properly managed, may lead to systemic crisis and large economic losses. One important aspect of bank supervision and regulation, there- fore, is crisis preparedness and management. Given the greater volatility and responsiveness of financial markets in an integrated environment, and the potential for larger losses, the need to detect and contain a bank- ing crisis at an early stage becomes even more urgent. For this, there is a need for good and reliable information, to allow for prompt decision- making and swift corrective action by bank supervisors. In particular, bank supervisors need to carefully monitor off-balance and offshore op- erations and the extent to which these are used as a device to hide non- performing assets and bypass domestic regulations. The practice of lending to third parties abroad-who then relend the same funds to a related domestic party-has been used in some instances to bypass do- mestic regulations against self-lending. This occurred, for example, in Chile during the early 1980s and Venezuela during the 1990s. More broadly, there is a need for better information about banks' risk assess- ment and management, self-lending, portfolio concentration, foreign exchange exposure, and currency mismatches. Such information can help policymakers take preemptive actions and may provide the basis for 53 FITAL FLOWS TO DEVELOPING COUNTRIES contingency planning, which, in turn, can help reduce delays and also trial-and-error policies in the event of a crisis. Crisis management involves the allocation of losses and the manage- ment of failing institutions-that is, whether and how to intervene in ailing banks. In an integrated environment, policymakers need to in- tervene promptly and decisively in ailing institutions in order to con- tain the crisis and impose losses on the parties responsible: the bank owners and managers. Although rigid rules can he]p policymakers resist the opposition of interest groups affected by the intervention, they can also force the intervention and liquidation of viab le institutions (Gold- stein and Turner 1996). Thus, good judgment and discretion are needed in managing a banking crisis, and in many instances policy- makers may choose to follow a more heterodox approach rather than a strict application of rules and pure orthodoxy. The allocation of losses will most likely be biased toward the less protected and less informed groups-small depositors, small share- holders, and taxpayers-in an integrated setting, since large depositors, bank owners, and managers will often be able to move funds offshore. Thus, the need to intervene promptly is even more urgent if the objec- tive is to impose maximum losses on the largest market participants, those who have greater incentives to be well informed and for whom it is cheaper-proportional to their investment-i o move funds from onshore to offshore locations. Finally, a cautious macroeconomic stance becomes even more im- portant when managing a banking crisis under the more stringent con- ditions imposed by integration, especially in the presence of a currency or quasi-currency board. Preparing Capital Markets for Financial Integratioin Given the changing investor base, a growing proportion of flows to de- veloping countries is being channeled through their capital markets in the form of portfolio equity capital. These investments, which repre- sent an important opportunity for developing countries, have been ac- companied by a spectacular increase in activity in their equity markets. Table 3.1 in chapter 3 illustrates the spillover efFects of portfolio in- vestment on domestic trading activity. But while the improvements in many emerging markets have been noteworthy, most of these markets are still in the early stages of development and reed to close the gap 54 THE MA separating them from more advanced capital markets to be able to compete in an integrating world. And they need to compete for new is- sues and investors not only with the more advanced markets but also increasingly with other emerging markets. To this end, developing countries face two main tasks: * First, they must implement policy reforms and strengthen insti- tutions to make their markets more attractive to foreign investors and reduce the risks of capital market instability. While investors are attracted by the potential for rapid growth and high returns, they are discouraged by operating inefficiencies, by the lack of re- liability of market institutions and infrastructure, and by regula- tory frameworks that increase transaction costs and reduce transparency.9 Improvements that increase the attractiveness of emerging markets for foreign investors also serve to reduce volatility and risks. * Second, authorities in developing countries need to deal with the new regulatory concerns resulting from globalization. These con- cerns, which also relate to the banking sector, are reviewed in the last section of this chapter. This policy agenda is not new, in that it includes many of the policy and institutional reforms essential for developing capital markets in a more closed economy. Domestic and foreign investors generally wel- come the same institutional and policy improvements, which also di- minish volatility and risks irrespective of whether they originate from domestic or external sources. Improving market infrastmcture. As cross-border flows increased during the 1980s, there was considerable debate on best practice in market infrastructure and how to harmonize systems across countries. 10 The landmark efforts in this area were the Group of 30 (G-30) initiative in 1989 and the 1995 workshops organized by the International Society of Securities Administrators (ISSA). The G-30 recommendations and suggested ISSA revisions are summarized in box 1.3. Partly in response to foreign portfolio investment, emerging markets have made major strides in improving their infrastructure and now meet many of the G-30 standards (table 6.4, chapter 6). Twelve out of the 16 markets included in table 6.4 now have central depositories, and the remaining four-India, Indonesia, Pakistan, and the Philippines- are all scheduled to have them by mid-1997. Many emerging markets, 55 4tIAT 1TAL FLOWS TO DEVELOPING COUNTRIES Box 123 Best Practice in Market Infraisbucture Matching. The matching system should be integrated with the dear- ance and settlement (C&S) system, and all trades by direct and indirect (institutional) participants should be matched at most by T+l (one day after trade). Clearance and settlement Settlement should be accomplished by a delivery versus payment system of good quality with same-day funds (that is, final delivery of securities takes place only if final payment is made); there should be real-time gross setdement or a netting system that meets stringent risk control standards, depend- ing on the characteristics of the market; and the rolling setdement system should have final setdement occur by T+3. Depositoy. There should be one independent central depository managed for the benefit of the industry, broadly defined, and if more than one depository exists, they should be interlinked; there should be an independent registry or registries; and immobilization and dematerialization should be encouraged, ancl the legal frame- work revised, if necessary, to permnit this. especially in Asia, have also adopted computerized trading systems to increase market transparency and the capacity of the market to handle the surge in activity that accompanies financial integration. These im- provements are remarkable. A key lesson that can be drawn from recent country experience, especially in Asia, is that it is possible to achieve rapid progress and in some areas leapfrog to stare-of-the-art systems with a sustained and well-organized effort. At the same time, country experience also suggests that the G-30/ ISSA benchmarks are objectives to be attained over time and tailored to country circumstances. For example, the shortening of the settlement cycle to the G-30 benchmark of T+3 (the third day after the trade) has been a key focus of this effort. But the emphasis on speed may not be fully justified. Foreign investors care less about G-30 speed standards than about reliability. Malaysia, one of the more successful emerging markets, illustrates this point very well: although settling only at T+5, it has been able to achieve one of the more reliable settlement systems in the developing world. Similarly, developing countries need to give greater emphasis to re- ducing systemic risks in their securities markets. With respect to inter- 56 THE M AI national benchmarks, least progress has been made in achieving deliv- ery versus payment in the settlement process. With central deposito- ries, the delivery side of the equation is generally working well. It is on the payment side that delivery versus payment seems to fail, perhaps because of weaknesses in the domestic banking and payment systems. Market authorities in developing countries should also consider setting strict membership standards for clearinghouses and other focal points of transmission of systemic risks. High entry standards are particularly important until other risk reduction systems become fully operational at the clearinghouse and until the brokerage industry strengthens. Despite the major improvements in market infrastructure and per- formance in recent years, there is still a gap between emerging and ma- ture markets. As shown in figure 6.3, chapter 6, industrial countries are more efficient by a factor of 10 to one regarding the average number of days it takes to collect dividends, 20 to one regarding the average num- ber of days to register, and two or three to one regarding trades that set- tle with a delay. The width of the gap varies widely among emerging markets. For some countries, such as India, the gap is very striking. Other markets, such as Korea and Mexico, have a performance close to or even better than that of industrial markets. Legal and regulatory framework. Constructing and reinforcing the regula- tory framework is essential for emerging markets to attract foreign investors and reduce systemic risk. Investors are most concerned with protection of property rights (including minority shareholder rights) and transparency. For example, investors want both macro data on economic prospects and micro data on corporate performance, to be able to make informed investment choices. Improving disclosure will not only address investor concerns but will also reduce the susceptibility of the market to volatility resulting from incomplete or asymmetric information. Unlike market infrastructure, however, there are no clear-cut criteria for a sound regulatory framework, and institutional structures and practices in industrial countries vary considerably depending on histor- ical antecedents. There is, nevertheless, a growing convergence toward the so-called self-regulatory organization (sRo) model, one based on public disclosure, on self-imposed market and industry discipline, and on better internal risk management by financial firms themselves." The key elements of this model are discussed in box 1.4. The SRO model faces a number of potential problems in emerging markets: the institutional and human capital may be insufficient to en- 57 C6AAPITAL FLOWS TO DEVELOPING COUNTRIES Box 1.4 Principles of the Seif-Regulatory Model FIRST, THE PURPOSE OF THE REGULATOR IS NOT prone to contagion. An exception to this rule regards to substitute for the market in helping investors to clearance and settlement arrangements, discussed make investment decisions, but to ensure that the above. However, with the increasing integration of incentives and structure of the market are consistent banks and securities businesses, there is a much with efficiency, fairness, and safety. The regulatory stronger rationale for intensifying prudential rules framework should ensure that timely and accurate and oversight of the latter. information is available, so that investors can judge Fourth, the protection of investor's assets from the mei its of alternative investmnents, hence the reg- loss and the insolvency of investment firms is an- ulatory emphasis on disclosure and eradicating other rationale for prudential rulks and oversight of fraud. The regulator should prohibit insider trading investment firms. Investment firms are often re- for fairness reasons, and because such practices have quired to meet capital adequacy standards, segregate the negative externality of discouraging investors investor's assets, and meet minimum internal opera- and savers from participating in capital markets. tional standards. In addition, the legal system should Second, market participants, rather than govern- provide the basis for protecting investor's assets. ment, should have the main responsibility for estab- Disclosure to investors of the level and type of risks lishing and enforcing market rules and regulations. to which their assets are exposed and the financial The rationale is that participants have an interest in status of investment firms is essential for market dis- ensuring that markets are fair and efficient and are cipline to play a role in reducing ihese risks. better able to judge how to make them so. In practi- Fifth, financial innovation, in particular the de- cal terms, this means that self-regulatory organiza- velopment of a wide variety of derivative products, is tions (SROs) such as the exchanges, broker-dealer leading to some important changes in how financial associations, and accounting and auditing associa- firms are regulated. Although der[vative trading still tions will bear much of the responsibility for the reg- involves price risk, the speed witht which these risks ulation and surveillance of securities rnarkets and can be transformed and the opaci[ty of the transfor- auxiliary supporting services. For the system to work mation process make it difficult for regulators to as- correctly, the official regulator must have sufficient sess the degree of exposure of financial firms. The regulatory oversight to ensure that the SROs enforce regulatory response to this lack of transparency has securities regulations as well as their own market con- been to focus on the ability of each firm to manage duct rules, and that they act to minimize potential these risks, and to create incentives for financial conflicts of interest and restraints on competition. firms to put in place appropriate risk control proce- Third, concerns about systemic risk in capital dures. This is a relatively new xrea of concern in markets do not justify prudential regulations and emerging markets, since derivative products are not monitoring as intense as those in the banking sector, widely used (except in Brazil and Malaysia) or may As a general rule, investment firms are less vulnera- actually be proscribed. However, as financial inte- ble than banks, on both the asset side and the liabil- grarion intensifies, domestic firnancial firms may ity side, to liquidity and solvency crises, and are less trade in these instruments overseas. sure that two pillars of the system-self-regulation and disclosure-re- sult in fair and efficient markets, the structure of iEhe securities markets is imperfect because of limited competition, and there may be short- 58 T]HE MAIN N s term tensions between the regulatory and market development objec- tives of the authorities or the SROs. Despite these problems, however, the self-regulatory model is the best alternative, and emerging markets appear to be converging toward it. Emerging markets have also made progress in developing their framework of law and statutory regulation. All the major emerging markets now have in place the basic building blocks of such a frame- work; for instance, they have enacted securities laws (Russia most recently) and established independent securities commissions.12 Ac- counting standards of many emerging markets have also strengthened considerably, although accounting practices are often impeded by the lack of qualified accountants and auditors, a common problem in many emerging markets. Finally, most major emerging markets now have regulations defining disclosure standards and listing requirements. Where the gap between developing countries and their industrial counterparts is still wide is in the more detailed, but still critical, aspects of a sound regulatory framework. For instance, about half of the emerging markets in a sample of 16 (see table 6.6, chapter 6) have not established the legal and regulatory basis for compensation funds, takeovers, and insider trading.13 Many emerging markets have not yet instituted the legal and regulatory basis for domestic institutional in- vestors. More broadly, there is a danger of overregulation in Asia, which may stifle market development and discourage foreign investors. In Latin America, the danger seems to be underregulation or lack of effec- tive enforcement. In Eastern Europe and the former Soviet Union, the main task is to establish the basic legal and regulatory framework for capital market development. The remaining agenda. Developing countries show considerable varia- tion in the capital market attributes needed for financial integration. The most dynamic emerging markets, where progress has been partic- ularly intense during the last five years, include most of high-growth Asia (Korea, Malaysia, and Thailand, with Indonesia and the Philip- pines not far behind) and two markets in Latin America (Chile and Mexico, with Brazil also ranking well). The East Asian markets stand out for their depth and liquidity, and because of efforts undertaken in the 1990s their market infrastructures are now equal to those in Latin America. The lagging emerging markets in the sample are in South Asia (India, Pakistan, and Sri Lanka) and China. Generally, these countries need to continue to improve their market infrastructures, as 59 _ ITAL FLOWS TO DEVELOPING COUNTRIES well as their institutional development. But even in the most advanced markets, the outstanding agenda is large. In the lnfrastructure area, for example, about 20 percent of all securities trades in Malaysia and Thailand do not settle on the contractual settlement date-four times more than in the United States. To close the gap, emerging markets should pursue the following pol- icy agenda: * Infrastructure. Emerging markets should implement well-syn- chronized comparison, clearance and settlerment, and central de- pository systems, with the goal of meeting (G-30/issA guidelines, though not at the expense of reliability. Emerging markets should also pay close attention to reducing systemic risk by developing sound links with the banking and payment systems. In addition, the risk control procedures of a central clearmng agency, if such an agency is required, should meet Bank for International Settle- ments (BIS) guidelines, and a central deposiitory should be estab- lished early in the integration process. • Property rights. The legal and regulatory framework should in- clude two basic principles of shareholder governance: fair treat- ment for all shareholders and shareholdLer approval of key corporate decisions. In transition economies, it is essential to es- tablish an independent registry of equity ownership so that records cannot be manipulated by management or shareholders. * The regulatory framework. Emerging markets should adopt inter- national best practice on disclosure (including accounting) and self-regulation to local conditions, and improve enforcement of these rules. Government regulatory functions (starting with over- sight of trading activities) should be devolved to SROs as quickly as practicable but should take into account potential conflicts of interest and the capacity of the sRos. Finally, emerging markets should also promote the development of domestic institutional investors, who can serve as a counterweight to foreign investors and thereby assuage fears of excessive foreign pres- ence. Domestic institutional investors can ensure that a large pool of dedicated money will be available for bottom fishing and value picking, which will reduce the vulnerability of domestic capital markets to a rapid liquidation of assets by foreign investors. In addition, domestic institutional investors will increase the depth and liquidity of domestic 60 THE M A I capital markets, enabling the markets to absorb the benefits that inte- gration can produce. New Regulatory Challenges and the Need for International Cooperation A basic problem that financial integration creates for regulators is that although financial institutions and transactions are increasingly global, their activity and authority remain mainly national in scope. Given this limited scope, regulators would not be able to insulate their domestic financial systems from overseas volatility and shocks unless they cooperated and coordinated with foreign supervisory and regula- tory authorities. In other words, while the domestic policy and insti- tutional improvements described previously are critical to prepare for financial integration, they need to be complemented by international initiatives. The new risks and challenges. Financial integration can affect the nature and magnitude of the underlying market, counterparty, and systemic risks in domestic financial markets. In particular: * As domestic financial firms engage in cross-border activities, they may incur new market risk from open positions in exchange- and interest-rate-sensitive foreign assets and liabilities. * Domestic residents, including financial institutions, engaging in cross-border transactions will be exposed to new counterparty and credit risks with foreign investors and firms. * Systemic risks also increase with financial integration, since it subjects the home country to more sources of external distur- bances and increases the speed at which these disturbances are transmitted.14 Systemic risk also rises because increasing volume and the different hours of operation among payment systems delay settlement and increase settlement risks. These activities and associated risks are creating new and interrelated regulatory challenges in two broad areas. First, the most obvious chal- lenge is to contain these new sources of systemic and counterparty risk at the same time that globalization is rendering the regulatory environ- ment more complex. Second, globalization is also magnifying the importance of regula- tory issues that arise from cross-country differences in regulations and the possibility that financial intermediaries will engage in regulatory ar- 61 CAPJTAL FLOWS TO DEVELOPING COUNTRIES bitrage-that is, seek to take advantage of differences in regulations and their coverage among countries. Two other trends in international financial markets are interacting with globalization, significantly magnifying the regulatory challenge: * First, financial firms are increasingly becoming financial con- glomerates, combining traditional banking with securities opera- tions and other nonbank financial activities. * Second, financial innovation has resulted in a vast array of new derivative instruments that can change the risk profile of financial firms very quickly and in a very complex manneL In summary, globalization, combined with financial innovation and conglomerates, is increasing the channels and speed of transmission of systemic shocks across borders and sectors while reducing transparency in the marketplace. This lack of transparency is undermining the abil- ity of the market to police itself and of the author ties to regulate. The international regulatory response. National regulatory authorities have tried to address the risks and challenges of globalization through stronger international cooperation. One clear objective of this cooper- ation has been to control systemic risk, since national regulatory frameworks in this area obviously have an impact across borders. Indeed, there seems to be growing concern among national authori- ties, heightened by the failure of Barings bank in 1995, regarding the dangers of unregulated or underregulated financial activities of global firms.i5 There has also been international cooperation in the area of investor protection, with long-standing efforts to harmonize regula- tions across countries and promote international cooperation in enforcement. In the securities markets, these efforts were undertaken in part to reduce transaction costs and increase market efficiency, and are motivated by pressure from institutional investors and internation- al issuers. With regard to banks, these efforts were primarily directed at achieving a level playing field for banking institutions from differ- ent countries. And for both banks and securities markets, these efforts were also directed at addressing the fears that competition between financial centers would result in a regulatory race to the bottom, and that banks and investors would practice regulatory arbitrage. Recent international initiatives show a gradual convergence toward regulatory systems more based on disclosure and market discipline that increase the incentives for improved internal methods of risk control. 62 THE MAIN t There are several reasons for this. First, this convergence illustrates the growing global financial clout of institutional investors who strongly prefer a system based on strong disclosure. Second-and more impor- tant, as noted above, in today's sophisticated global financial markets- the risk exposure of financial firms is increasingly hard to quantify and, in any case, is changing very quickly. In addition, it is hard for regula- tors to keep up with the rapid evolution of financial instruments and techniques. Hence, a consensus has emerged that regulators would do better by supervising the quality of risk management (rather than the position of a firm at a particular moment in time) and by ensuring ad- equate disclosure for market discipline to work effectively. The focus of international cooperation has been the Basle Commit- tee on Banking Supervision for banking regulators, and the Interna- tional Organization of Securities Commissions (losco) for securities regulators. Both have played a key role in these endeavors to reduce transaction costs and systemic risk. The IMF has also played an impor- tant role, in particular in the area of ensuring adequate disclosure of macroeconomic information. The main international initiatives are de- scribed in box 1.5. The main efforts at international collaboration in banking supervi- son have focused on developing prudential standards, which in early stages were designed primarily for international banks in industrial countries. The rash of severe crises in developing countries over the past 15 years has, however, recently rekindled interest in developing more comprehensive guidelines. Indeed, international supervisory and regula- tory authorities are expected to issue such guidelines in the near future. Other efforts at international collaboration in financial markets have focused on clearance, settlement, and payment systems and on regulat- ing financial conglomerates. In the payment area, settlement and sys- temic risk have been reduced by harmonizing regional payment standards and increasing the overlap of hours of operation of different payment systems. With regard to financial conglomerates, securities, banking, and insurance regulators have been collaborating in develop- ing principles for the supervision of financial conglomerates and for collaboration among the three different types of regulators.i6 Finally, the private sector itself has been the source of initiatives on best prac- tice, as well as a key source of advice and criticism for official initiatives. The G-30 (for example, in the area of capital market infrastructure), the Institute of International Finance (in macroeconomic information 63 ITAL FLOWS TO DEVELOPING COUNTRIES standards), the International Federation of Stock Exchanges (in con- trolling systemic risk in equity markets), and mariy others have played important roles. Policy implications for emerging markets. The above discussion suggests several implications for policymakers responsible for emerging securi- ties markets. First, reducing information asymmetries across borders has a high payoff both for reducing systemic risk and for improving the efficiency of financial markets. At the macroeconomic level, more accurate and timely disdosure of country macroeconomic data would increase the informational content of the capital flows to developing countries and reduce the likelihood of cross-country contagion of financial shocks. More emphasis on disclosure of accurate information Box 1.5 Intemational Regulatory Cooperation and Coordination THE MAIN EFFORTS OF THE BASLE COMMITTEE tant aspects of this agreement are the use of and IO',CO are as follows: the concept of "value at risk" and permitting the more sophisticated banks to use their own * Among the more successful efforts in interna- internal risk management systems to deter- tional regulatory cooperation is the so-called mine their own capital adequacy require- Basle system for banking supervision and regu- ments.1 The new guidelines, however, lation. It consists of a series of agreements over continue to be the subject of much debate. the period 1975-92 among banking regulators Among the most important areas of concern fi-om the main industrial countries. These in- are the apprehensions of securities regulators clude understandings on the allocation of regui- that banks and securities firms may develop latory responsibilities, on collaboration models that reduce capital requirements arrangements including information sharing rather than measure market risk adequately, and enforcemenet, on supervisory standards, and the impact of the guidelines on the com- and on the harmonization of minimum capital petitive positions of international banks. adequacy standards. The 1988 agreement on a In the securities area, to promote regulatory capital adequacy is perhaps the most important convergence and reduce transaction costs, of these achievements (see annex 5.3). IOSCO has issued many reports and resolu- * T'he Basle Committee has been working in tions on best practice, including such areas as the 1990s on refining capital adequacy stan- curbing and punishing securities fraud, dis- dards so as to take into account the growing closure and accounting standards, clearance ijmportance of market risk in bank portfolios, and settlement, and investor protection including the rapid increase in trading in fit- (which is also important in reducing systemic trares and options. After extensive debate, the rnsk). In addition, given the strong representa- committee issued its final recomnmendations tion of emerging markets at IOSCO, its resolu- in December 1995. Two of the most impor- tions, reports, and accompanying discussions 64 T HE M{AlP4X of the financial status and risk profile of financial intermediaries would increase the effectiveness of market discipline. And for regula- tors, there is a high payoff in information sharing and enforcement agreements, both between emerging and industrial country markets and among developing countries themselves. In particular, although authorities in the home country bear the main responsibility for cross- border supervision of banks' activities, authorities in host countries can help by monitoring specific aspects, such as banks' liquidity, and, most important, by removing impediments to the exchange of rele- vant information between supervisory authorities. Second, as financial integration deepens, the experience of the BIS member countries suggests that it would be highly beneficial for devel- have served as a conduit for the transfer of ex- in the number of these bilateral agreements pertise and experience between countries. these last few years, many of these under- * In the area of systemic risk, especially during standings have been between industrial coun- the 1990s, a key focus of both the B3asle Corn- tries, some between industrial and developing mittee and IOSCO has been to contain the sys- countries, and only a few between developing temic risks arising from derivative activities. countries. Recently, in response to the Wind- In this regard, they collaborated in several re- sor Declaration, both exchanges, clearing- ports during 1994-95 on best practice in dis- houses, and regulators of the major global closure of both qualitative and quantitative futures and option markets have signed infor- information of derivative activities (for exam- mation-sharing agreements. pie, Basle Committee and IOSCO 1995a and 1995b), and in the internal management and 1. Value at risk is an estimate of the maximum loss in the control of ris3k. IoSCo is also fleshing out the value of a portfolio, including positions in fitures and options, contrl of isk. OSCO s als flesing ot the over a certain time period and at a certain level of confidence. recommendations made in May 1995 by the The confidence level refers to the probability that the loss will regulators of most major futures and options be lower than a presepecified maximum. See MF (1995) for a markets (the Windsor Declaration), including discussion of the concept and estimation procedures. protection of customer positions and assets, 2. These memoranda usually include all or a subset of and the strengthening of default procedures. sharing ofc routine sharing of general information, shrn fcertain information on firms operating in the * Given that its recommendations are advisory two markets, access to official information held by the and nonbinding to its members, IOSCO has counterpart regulator that may help in an investigation or strongly encouraged coordination among an enforcement action, approval of certain investigative members, in particular through bilateral powers of the domestic regulator in the counterpart coun- agreements (memoranda of understanding) try, and the obligation to report to the other party that a firm is experiencing financial difficulties. regarding information sharing and enforce- Source: Dale (1996), imE (1995 and 1996), and Gold- ment.2 While there has been a sharp increase stein (1996). 65 ICAPiTAL FLOWS TO DEVELOPING COUNTRIES oping countries to become more active members in international arrangements regarding bank supervision. Unlike IoSCO, the Basle Com- mittee and other international institutions in this area (for example, the European Union) include mainly Organisation for Economic Co-opera- tion and Development (OECD) countries. It is also in the best interest of the industrial countries to extend the improvements in regulation and su- pervision in international financial markets to the larger and systemically important developing countries. However, there is still no clear view on the best way to achieve this extension of the multilateral framework of bank surveillance. In addition, many developing countries would have to improve significantly their prudential standards and supervisory practices to meet the minimum standards of the Basle Committee. The Role of Official Finance The large increase in private capital flows to developing countries raises the issue of the role that official finance plays in an age of global private capital. Although private flows now surpass official flows by a margin of five to one, the vast majority of developing cDuntries receive rela- tively little private capital. Low-income countries (excluding China and India) received only 3.4 percent of total private capital flows to de- veloping countries during the past three years, and all of Sub-Saharan Africa (excluding South Africa) received only 1.5 percent of the total. Indeed, these countries remain largely dependent on official financing (figure 1.1 1). The challenge for these nations, and many middle- income countries, is to establish the prerequisites for attracting private capital on a sustainable basis. There is now compelling evidence that success in attracting private flows depends on macroeconomic fundamentals, outward orientation and a market-friendly environment, and progress on infrastructure and human resource development (box 1.2 and box 2.2). The experience of countries that have been most successful in attracting private capital suggests that official finance has played an important facilitating role in helping them establish these necessary conditions. These countries have typically relied on significant official fino nce during the two decades preceding the surge in private flows in the 1990s to support the buildup of physical and human capital stocks. Perhaps even more significant, for a majority of these major recipi- ents, the surge in private capital flows was preceded by concerted policy 66 THE MAI1 Figure 1.11 Official and Private Gross Capital Flows to Low-income Countries (excluding China and India), 197095 Percentage of GNP 10 9 | Official flows 8 -+ Private flows 7 6 S 4 3 2 Except for China and India, low- 1 income countries depend more on official financing than on private o capital flows. 1970 1975 1980 1985 1990 1995 Source World Bank data lending by the IMF and the World Bank, sometimes with an increase in the overall level of official assistance. For some countries, the surge was preceded by the successful completion of a Brady debt reduction opera- tion. Policy advice and conditionality associated with this lending and the Brady operations played an important facilitating role, by helping these countries attain macroeconomic stability and adopt market-ori- ented policies, and by signaling these improvements to private markets. An analysis of the determinants of private capital flows for a sample of 73 developing countries shows that net lending from multilateral sources and the completion of Brady operations have both played com- plementary roles in facilitating private capital flows (Corbo and Hernandez 1997). The challenge for official finance in the case of the vast majority of countries that are not yet attracting private capital is to foster change that would raise private returns (through policy reforms and sound pub- 67 JJA-PITAL FLOWS TO DEVELOPING COUNTRIES lic investments) and to help countries address the two main impedi- ments to private capital flows-high risks and limited information. The principles that govern the role of official assistance apply as well to developing countries that are already attracting substantial amounts of private capital. Despite significant progress, most of the major de- veloping country recipients of private capital still have limited macro- economic track records, do not yet have robust banking systems, suffer from deficiencies in infrastructure and human resource development, and more generally lack the institutional structures needed for inte- grated financial markets. Accordingly, while the role of official finance has appropriately declined for this group of countnes (figure 1.12), it can continue to play a valuable complementary role to private flows even in these countries for several important reasons: * Official assistance can help sustain improvements in the policy and institutional framework. One area of paiticular importance is the strengthening of banking systems. Financial systems that are not performing their critical functions (resource mobilization and allocation, monitoring investments, facilitating risk manage- ment, and providing payments mechanisms) are more likely to misallocate capital and contribute to increased vulnerability in the face of growing financial integration. But strengthening the financial sector is a complex and long-term task, and the differ- ences in countries' financial systems and the dynamic nature of fi- nancial sector development mean that no one formula will work in all systems. International agencies can br a source of knowl- edge of what has worked and what has failed, and why, and can also help countries with institutional development and expertise in effecting reform programs. Recognizing the important role that the financial sector plays, multilateral 3nstitutions-includ- ing the World Bank, the IMF, other multilateral development banks, the BIS, and the Group of Ten (G-10)-are expanding their efforts in this area. * Although markets are becoming more discerning, and more countries are adopting the needed policy reforms, developing countries that are attracting private capital flows will remain vul- nerable to shocks and large reversals for some time to come. The IMF, the World Bank, and other official lenders can reduce the likelihood and costs of such reversals by helping to improve the 68 THE MAIN-'fl Figure 1.12 Official and Private Gross Capital Flows to the 12 Largest Recipients of Private Capital, 1970-95 Percentage of GNP 7 -U-Official flaws 6 Private flows 4 3 2 Within the last few years, private capital has rapidly outstripped official financing in the 12 developing countries that received the largest amounts of private 0 investment. 1970 1975 1980 1985 1990 1995 Note The countries are Argentina, Brazil, Chile, China, India, Indonesia, Hungary, Mataysia, Mexico, Russia, Thailand, and Turkey Source World Bank data. availability of information about these markets, and by helping countries get back on track in the face of balance-of-payments difficulties. The IMF has recently formalized the procedures used in the Mexican crisis as a mechanism for emergency financial sup- port, and has established New Arrangements to Borrow that will make available resources of up to SDR 34 billion (special drawing rights, about US$48 billion) for such support. The IMF has also drawn up standards for the dissemination of data, and some countries have started to provide the financial markets with bet- ter information. * Official lenders can also finance and help improve the quality of investments in human resources and infrastructure, and support adequate social and environmental protection. These are all areas where private returns will often be less than the social returns, and where public investments play a vital role in enhancing the 69 -t-APITAL FLOWS TO DEVELOPING COUNTRIES productivity of private investment. Official assistance can also promote increased private participation in the infrastructure and human resources sectors, by supporting improvements in the pol- icy and institutional framework and by assisting with privatiza- tion programs. In addition, support for poverty reduction and better social safety nets will not only promote human develop- ment-an important determinant of long-term growth-but re- duce risks of social and political instability as well. This complementary role of official financing is being reflected in the changing composition of lending by official agencies. For in- stance, lending for environment, social sectors, and infrastructure now constitutes 80 percent of total World Bank investment lend- ing, compared with 60 percent in the early 1 980s. * International financial institutions may also be able to catalyze private investment by providing selective guarantees for risks that they may be better able to mitigate-such as political risks and risks arising from shifts in the regulatory regime-or for promot- ing pioneering investments with large potential spillover benefits. Although third-party guarantees complicate negotiations on the cost of financing, guarantees by multilatera. agencies can play a useful role in reducing the uncertainties sutrounding the future policy regime. Even if the government establishes a schedule of changes for the telephone industry, for example, or commits to a formula for electricity tariffs, private lenders may still fear that such decisions will be reversed. Agencies such as the World Bank can support transactions by placing their influence and financial resources behind the government's commicment (World Bank 1997). * Finally, official institutions can work directly with private part- ners to expand and improve private capital flows (de Larosiere 1996).17 In addition to the use of guarantees, the experience of the lFc and the European Bank for Recovery and Development (EBRD) shows that multilateral agencies can work directly with private companies in a number of ways. They and other multilat- eral agencies have helped set up joint ventures for activities such as leasing, venture capital, infrastructure dev,2lopment, and bank- ing, which have not only brought benefits [n terms of loan and equity financing, but have transferred know-how, encouraged fi- nancial innovation, spurred privatization, provided demonstra- 70 THE MAfl40, tion effects of good management and professionalism, and pro- moted institutional development. The catalytic role of official finance outlined above is greater than ever before because of the increased responsiveness of private capital and the shift to market-oriented policies in developing countries. Re- search shows, however, that aid is not effective when the framework for good economic policies-especially sound macroeconomic policies- and governance is lacking. Indeed, increased official borrowing can have an adverse effect on private flows if the returns are not commen- surate with the cost of borrowing (by reducing tax-adjusted expected rates of return) and if it sustains policy distortions and fiscal laxity. Hence policy performance is a key prerequisite and should be the driv- ing force determining the allocation of all types of aid. Official assis- tance also needs to be redirected to those areas where it can be most effective and generate the greatest leverage-and this will vary accord- ing to country situations. While official finance can continue to play a complementary role in countries that have been more successful in attracting private capital, official concessional assistance is especially important for many low- income countries. There is an urgent need for continued concessional assistance to support reform programs in these countries and address their large needs in human, social, and infrastructure development, be- cause without such assistance they would fall even farther behind. The infusions of private capital that financial integration is channel- ing to developing countries can-and will-do much to transform these nations, but whether the transformation is beneficial or detri- mental depends, to a large extent, on these countries' ability to develop the institutions and implement the policies that will allow integration to succeed. Unfortunately, they must accomplish these tasks in an en- vironment that affords restricted scope for independent action, and they must start down the road to financial integration before they are fully ready for the journey. Even so, it is within the power of these na- tions to realize substantial gains in the new age of global private capital. 71 -, CAPITAL FLOWS TO DEVELOPING COUNTRIES Notes 1. Korea has been classified as a high-income country episodes in which the consumptilon boom was larger since 1996. than expected comprise Argentina 1992-93, Brazil 1992-94, Finland 1987-94, Mexico 1989-94, Norway 2. Such r.sks include not only sovereign risk-which 1984-89, Sweden 1989-93, and Venezuela 1975-80. is still quite high in developing countries given their rela- Finally, the country episodes in which investment was tively short policy track record-but also higher invest- lower than expected were Argentina 1979-82, Brazil ment risks, such as legal and custodial risks, settlement 1992-94, Chile 1978-81, Finlandl 1987-94, and Mex- and operatienal risks, information and regulatory risks, ico 1989-94. and nonmarket risks. Hence part of the differential in ex- pected rates if return between industrial and developing 8. These banking sector reforms need to be comple- countries reflects these higher risks in the latter. mented by supportive macroeconomic policies, given the importance of macroeconomic stability for sustained re- 3. Direct measures of financial integration based on forms and the possibility of increased macroeconomic the equalization of expected returns are difficult to con- swings in an open environment. In particular, the threat struct for developing countries, given that forward mar- of large capital flow reversals and their consequent im- kets or surveys of exchange rate expectations exist for pact on the banking system (through loss of deposits, only a small proportion of developing countries. Under higher interest rates, and downturn in economic activity) managed exchange rates, potential "peso problem" diffi- makes macroeconomic stability an even more vital pre- culties make it hard to construct a measure of integration requisite in an integrated setting. based on arbitrage. 9. To date, foreign investors have been more attracted 4. In fact, correlations between returns in emerging to emerging market debt issued in international or in- markets and industrial countries would likely rise in the dustrial country markets than to domestic issues in de- very process of such a massive reallocation (as emerging veloping countries. Hence the report focuses on markets would become more integrated). improvements in emerging stock markets, although many development issues in the areas of market infra- 5. Technological innovations, which increase the rate structure and the regulatory framework are common to of return to capital, could mitigate some of the adverse both debt and equity markets. effects of an eventual decline in the labor force on the rel- ative rate of -eturn to capital. 10 Market infrastructure comprises the systems and institutions that facilitate the trade and custody of securi- 6. The country episodes we analyze are Argentina ties These functions can be subdivided into matching 1979-82 an(d 1992-93, Brazil 1992-94, Chile 1978-81 buyers and sellers, determining pr[ce, exchanging securi- and 1989-94, Colombia 1992-94, Finland 1987-94, ties for good funds, registering securities to the new own- Indonesia 1990-94, Malaysia 1980-86 and 1989-94, ers, and collecting dividends and oi:her custody functions. Mexico 1979-81 and 1989-94, Norway 1984 89, the Philippines 1978-83 and 1989-94, Sweden 1989-93, 11. Strahota (1996) describes how this model could Thailand 1978-84 and 1988-94, and Venezuela be applied in emerging markets. 1975-80 andi 1992-93 12. The latter is a relatively recent phenomenon in 7. Among the countries that experienced an exces- many Asian countries, where independent commissions sively large current account deficit were Chile 1978-8 1, replaced ministries of finance andi central banks as the Malaysia 1980-86, Mexico 1979-81 and 1989-94, the pnmary regulators of securities markets only in the early Philippines 1978-83, and Sweden 1989-93 Country 1990s. 72 THE MAIN F, 13. Compensation funds protect investors from 16. This includes the report of the so-called Tripartite losses arising from the failure of broker-dealers (not Group and the ongoing work by the Joint Forum. See market risks), while takeover rules protect the rights of IMF (1996). minority shareholders in a company targeted for a takeover. 17. "The basis of MDB collaboration with the private sector is straightforward. The funds, instruments, mde- 14. Most directly, the effects of a collapse or the in- pendence, and experience of the MDBs are combined with solvency of a financial firm in another country may be the know-how, management capabilities, and capital of transmitted to domestic markets, either directly, if the the private sector. Provided it takes account of the char- firm has established itself in the domestic market, or acteristics of the pro)ect, the market environment, and through the payment and settlement systems. the needs of the partners, this combination is a powerful force for private sector development." (Jacques de 15. The regulatory implications of the failure of Larosikre, the Per Jacobsson Lecture, IMF, Washington, Barings are discussed in Dale (1996) and IMF (1995) D.C., September 29, 1996.) 73 I TER 2 The New International Environment T _ NHE CYCLICAL DOWNTURN IN GLOBAL INTEREST rates in the early 1990s provided an important initial impetus for the resumption of private capital flows to developing countries. The fact that private capital flows to developing countries have persisted even after the upturn in interest rates in industrial countries suggests, however, that these flows have entered a new phase, reflecting structural forces that are leading to the progressive integration of devel- oping countries in world financial markets. The two primary forces that are driving investor interest in developing countries and leading to their increased integration are the search for higher returns and opportunities for risk diversification. Although these underlying forces have always motivated investors, the responsiveness of private capital to cross-border opportunities has gained new momentum as a result of internal and external financial deregulation in both industrial and developing coun- tries and major advances in technology and financial instruments. Fig- ure 2.1 summarizes the process. In industrial countries two key developments have increased the re- sponsiveness of private capital to cross-border investment opportuni- ties. First, competition and rising costs in domestic markets, along with falling transport and communications costs, have encouraged firms to look for opportunities to increase efficiency and returns (that is, prof- its) by producing abroad. This is leading to the progressive globaliza- tion of production and to the growth of "efficiency-seeking" FDI flows. Second, financial markets have been transformed over a span of two decades from relatively insulated and regulated national markets to- ward a more globally integrated market. This has been brought about by a mutually reinforcing process of advances in communications, in- 75 gn CAPITAL FLOWS TO DEVELOPING COUNTRIES Figure 2.1 'Structural Forces Driving Private Capital to Developing Countries Changes in the enabling environment in industrial countries (factors that increase the magnitude and speed of response of flows) Increased responsiveness of financial markets *greater internationalization and linking of national Increased responsiveness of firms financial markets gieater international investments because of greater internationalization of investor portfolios piessures to increase effictency and reduce costs because institutional investors ars more willing and able to invest internationally Higher long-term expected rates Opportunities for portfolio risk of return Primary diversification * better growth prospects factors * capital market deepening * improving creditworthiness * low correlation of returns tetween developing and industrial countries Increased market accessiblity * trade liberalization * Investment deregulation * privatization * opening of capital markets Domestic ani international Changes in the enabling environment in developing countries structural folrces (factors that increase the magnitude and speed of response of flows) are driving private capital to develop- ing countries. 76 THE NEW INTERNATIONAL ENVf IN formation, and financial instruments, and by progressive internal and external deregulation of financial markets. An important facet of this globalization of capital markets has been the growing importance of in- stitutional investors who are both willing and able to invest internationally. In developing countries, the environment is also changing rapidly. Since the mid-1980s, several countries have embarked on structural re- form programs and have increased the openness of their markets, through the progressive lowering of barriers to trade and foreign invest- ment, the liberalization of domestic financial markets and removal of restrictions on capital movements, and the implementation of privatiza- tion programs. Although perceived risks of investing in emerging markets remain high, these reforms have led to improvements in country creditworthi- ness risks, declines in investment risk, and increases in expected rates of return. As developing countries' securities markets have broadened and deepened, and as their market accessibility has increased, these coun- tries have also begun to offer investors significant opportunitites for risk diversification, which arise from the low correlation between rates of return in emerging markets and industrial countries. As a result, for- eign investors, who initially turned to emerging markets largely because of the cyclical downturn in interest rates and stock market returns in industrial countries in 1990, have begun to consider these markets on a more long-term basis. These changes both in the international setting and in developing countries have meant that developing countries have seen a strong surge of private capital in the 1990s. Of all the types of capital, FDI has responded most vigorously to the improving economic environment in developing countries. Commercial bank lending, which accounted for the bulk of the flows in the late 1970s and early 1980s, has also made a strong comeback. What is striking, however, is the growth of portfolio bond and equity flows. Whereas developing countries attracted barely any portfolio flows a decade ago, in the past five years they have re- ceived almost 30 percent of global equity flows. This growth was first stimulated by mutual funds, which were at the forefront of emerging market investments. More recently, pension funds have followed suit, investing either through mutual funds or directly on their own behalf. Consequently, institutional investors now form a very important part of the investor base in emerging markets, and this chapter focuses pri- marily on these investors and their (portfolio) investments. 77 -ZA -TAL FLOWS TO DEVELOPING COUNTRIES Together, these developments have resulted in a wider range of in- vestors and a broader composition of flows to dleveloping countries, with an increasing share of these investments going to the private sector. The process driving the financial integration of developing countries is still unfolding, but some of the implications are clear: * Continuing deregulation of financial markeris in industrial coun- tries, and technological progress and financial innovations at the international level, will spur increased responsiveness of private capital to international investment opportunities. * Developing countries' markets are likely to become increasingly accessible, and policy reforms are likely to deepen in countries that are already implementing such reforms and broaden to countries that are not yet embarked on the process. As a result, the financial integration of developing countries is ex- pected to deepen and broaden over the coming decade, against a back- drop of increasing global financial integration. Indeed, given the changes that are taking place at the internationa] level-in particular, the rapid advances in technology, communications, and financial inno- vations-and the growing economic sophistication in developing coun- tries, the progressive financial integration of the lai ter in world financial markets appears inevitable. Gross private capital flows may therefore be expected to rise substantially, with capital flowing not only from indus- trial to developing countries but, increasingly, among developing coun- tries themselves and from developing to industrial countries. Aggregate net private capital flows to developing countries are likely to be sustained in the short to medium term because of the continuing decline in creditworthiness risks and other investment risks, the higher expected rates of return in developing countries, and the fact that these countries are underweighted in the portfolios of institutional investors. The rate of growth, though, and eventually, the levels, will inevitably decline. There will probably also be considerable variation among countries, depending on the pace and depth of improvements in macroeconomic performance and creditworthiness. In fact, in coun- tries where economic and policy fundamentals are quite weak, the ini- tial manifestation of growing financial integration may take the form of net outflows of private capital. With changes in the international financial environment, there are likely to be considerable year-to-year fluctuations in private capital 78 THE NEW INTERNATIONAL ElN flows to developing countries, even in aggregate. Emerging markets will probably continue to be more susceptible to shocks from the inter- national environment than industrial countries are, a prospect that raises concerns in developing countries because private capital can now respond quite rapidly to actual or perceived changes. In the international environment, three factors in particular are seen as having the potential for creating significant volatility in private cap- ital flows to developing countries: * Movements in international interest rates and other asset re- turns-especially movements in industrial country stock mar- kets. Private capital flows to emerging markets are considered to be particularly affected by changes in these factors because in- vestors regard these markets as marginal.1 * Investor herding behavior. The new investor base in developing countries-dominated by institutional investors-is widely thought to be prone to herding behavior, arising from its incen- tive structure and the relatively limited information available on developing country investments. * Contagion or spillover effects, which arise when events in one emerging market cause investors to change their investment deci- sions in other emerging markets. The likelihood of contagion is also seen to be high in the current international environment, in part because of institutional features in the current investor base. Emerging markets are, in fact, more susceptible to volatility emanat- ing from the international financial environment-particularly to global interest rate and stock market movements. However, large move- ments in international interest rates appear unlikely, given that indus- trial countries are now operating in a low-inflation environment and hence are not likely to require sharp corrections in monetary stance. There is a higher likelihood of relatively large movements in industrial country stock market returns. A sizable decline in industrial country stock markets-especially in the U.S. stock market-could result in a general retrenchment, since investors would adjust to the negative wealth effect. Such a retrenchment is likely to be disproportionate with regard to investments in emerging markets, as emerging markets are more marginal in investors' portfolios. A small correction in the U.S. stock market could, however, result in a positive stimulus to emerging markets, as investors respond to the relatively higher returns in these 79 ITAI~TAL FLOWS TO DEVELOPING COUNTRIES markets. Finally, although emerging markets are more prone to poten- tial contagion effects, the likelihood of sustained volatility arising from pure contagion is also declining, because investors are becoming more familiar with emerging markets. At the individual country level, though, there is likely to be consid- erable variation. In particular, factors in the international environment could interact with domestic policies and conditicns to give rise to sig- nificant differences in the volatility of private capital. Thus, the inter- national environment could magnify and exacerbate shocks in the domestic economy. Or domestic investors can react to foreign in- vestors' initial reactions, again leading to a magnification of the shock. This chapter discusses these issues. The next section analyzes the fac- tors driving private capital flows to developing couritries and argues that these flows are increasingly being driven by forces that are creating per- manent structural changes in international capital markets. The follow- ing section looks at the nature of these structura- forces. The chapter then analyzes the effects of these forces on the growth of emerging mar- ket investments, focusing, in particular, on portfo lio flows, the form of investment associated with the new investor base. Based on an under- standing of the structural forces at play, the next section assesses whether private capital flows can be expected to be sustained. The chapter then analyzes the implications of the changing international environment for the volatility of private capital flows to developing countries, and it con- cludes with a summary of the main findings of the chapter. The Structural Character of New Private Capital Flows A- S PRIVATE CAPITAL FLOWS TO DEVELOPING COUNTRIES began to surge in the early 1990s, coinciding with declining global interest rates, it was generally assumed that these flows were being driven primarily by cyclical conditions in industrial coun- tries. This assumption was supported by early econometric analyses (box 2.1). The persistence of these flows in spite of global interest rate increases in 1994 and the Mexican peso crisis in 1995, however, suggests that they are being driven by more than international cyclical factors. In- deed more recent analyses (box 2.2) show several striking trends: 80 THE NEW INTERNATIONAL E PVV Box 2.1 Are Private CaPital Flows a CYckial and TemPOrary PhenomentOn? The Early Answers THERE HAS BEEN CONSIDERABLE DEBATE ON with several U.S. financial variables, induding inter- whether the surge in private capital flows to develop- est rates. This suggested that the main factor driving ing countries since the early 1990s is essentiaUly a private flows to Latin America was the cyclical temporay phenomenon, driven in large part by cycli- downturn in industrial countries and the associated cal factors in the international economy, or the result decline in global interest rates. of longer-term structural changes, which would sug- Chuhan, Claessens, and Marningi (1993), on gest that private capital flows will be sustained the other hand, included Asian countries in their More precisely, the debate has been about the analysis. They found that improvements in coun- relative importance of "push" factors (factors in the tries' economic fundamentals-the country credit global economy) and 'pull" factors (factors in rating, secondary bond prices, the price-earnings emerging markets) in explaining the surge in private ratio in domestic stock markets, and the black mar- capital flows. But since studies identified the push ket premium-were as important as cyclical factors factor as global interest rates, and the pull factors as in attracting portfolio flows to Latin America. Do- the improvements in countries' economic funda- mestic factors, moreover, were three to four times mentals, the arguments have also effectively been more important in explaining capital inflows to about the relative importance of cyclical factors (at Asia. the international level) versus structural factors (at However, since the Chuhan, Claessens, and the country level). Mamingi study considered country creditworthiness In a seminal article, Calvo, Leiderman, and Rein- as being solely determiined by improvements in the hart (1993) looked to see whether private capital domestic economy (whereas, in reality, global inter- flows to Latin America were driven primarily by est rates also affect country creditworthiness), the cyclical factors in the international economy or by study may have overstated the proportion that could improveements in countries' economic fundamentals be attributed to improvements in domestic funda- during the period 1988-91. Taking international mentals, as argued by Fernandez-Arias (1994). By reserves and the real exchange rate as proxies for pri- decomposing the improvements in creditworthiness vate capital flows, they analyzed the degree of co- into those arising from the decline in global interest movement in these variables using principal rates and those arising from improvements in the component analysis. (International reserves and the domestic environment, he found that global interest real exchange rate were used as proxies because of a rates accounted for around 86 percent of the in- lack of monthly data on aggregate private capital crease in portfolio flows for the 'average' emerging flows.) They found that there was a significant co- market during the period 1989-93. movement among countries' foreign reserves and On balance then, the prevailing view in the early among their real exchange rates, and that the degree 1990s was that cyclical factors in the international of co-movement increased in 1990-91 compared environment were the driving force of private flows with 1988-89. They also found that the first princi- to emerging markets. More recent work, however, pal component of both reserves and the real ex- suggests that there are some structural forces at work change rate exhibited a large bivariate correlation (see box 2.2). 81 1A1TAL FLOWS TO DEVELOPING COUNTRIES Box 2.2 Are the New Piivate Capital flows Cyclial or Structal? Recent Emplical Evidence TO ASSESS THE IMPORTANCE OF INTERNA- TF=a+ 0.0115*IlNV%1 -0.069*CONS,ir tional interest rates on private capital flows to devel- (243)*** (1 .62)** oping countries, a panel regression (which uses boih 0.031*DR - 01 VRER cross-country and tine-senes data) of total private -4.39 - (1.44)* it long-term capital flows/GNP was rmn on total invest- ( )* (1.44)* mentr/P, private consumption/GNP (if private in- + 0.005695DBR4D vestors consider private savings to be a sign of (1.01) confiden-ce in a country's prospects, the expected sign on this coefficient is negative), the stockof total -37) (I. 17) external, debt minus international reserves/GNP, volatility of the real effective exchange rate, a - 0.00 0021USDt + 0.356*F.i dummy for the successful completion of a Bray (2.02)** (4.94*** - deal, real export growth, the 12-month U.S. treasury bond rate, and a dummy for U.S. interest rates dir- -Adusted R2 - 46 ing 1990-93 (Hernandez and Rudolph 1995). EXcept for peD, the downturn in U.S. interest rates during 1990-3 was a significant factor in x- F a + 0.097A1INV,, - 0.087* CONS a plaining flows to developing countries, although do- (1.68)** (L71)Y mestic economic factors were also important -O3DRES, - 0.000036VRER To explore the influence of global interest rates (4.30)*** (2.24)* R on portfolio flows in particular, we analyzed the ex- tent to which portfolio flows firom the United + 0.00355*DBRAD Stares to 12 emerging markets in Latin America and (0. 52) East Asia moved together and the degree to which f 0.0001 7*EXPG,t + 0.000o36*IUSt this co-movement was related to U.S. interest rate ,034 S rmovements. Co-movement in flyws was measured by the first principal component of the flows. The - 0.000746*IUSD, + 0A4351 analysis was done for countries in each region sepa- (1.24) (6.0)*** rately and then in aggregate, as is shown in box Adljusted I?=0.4S table 2.2. (See Calvo, Leiderman, and Reinhart 1993 for the analysis using reserves and real ex- Significantat the 10 percent level; '*significantratthe5 change rates as proxies for capiial flows to Latin percent feveJ; *'*signicant at the I perceat level. America during 1988-93.) The results show that countries with strong eco- a The results show that there was a high degree nomic fundamentals have received the largest pro- of co-movement in flows duaring 1990-93 for portion of private flows relative to the size of their both regions and that this co-movement was econonuies. related to movements in IJ.S. interest rates. If foreign direct investment flows (mDI) are ex- This supports the hypothesis that U.S. inter- cIu&d rom the regression, the following results are est rates played an important role in driving obtained: pottfolio flows during 1990-93. 82 THE NEW INTERNATIONAL ENYVt Box Table 2.2. C-movenent of U.S. Portflio Flows to Emergng Mauk1ts, 1990-95 1990--93 1993-95 Correlation Corretion with PC with PC of interest of interest Principal Correlation rate and Principal Correlation rate and component of with PC of stock market component of with PC of stork market Region flows (PC) interest rate rerns flows (PC) interest rate returns Tatin Amerka T.813 -0.60 -0.60 0.548 0-33 0.31 Eat Asia 0.675 -0.59- -0.58 0.408 0-25 0.25 Total 0.755 -0.61 -0.60 0.455 0.28 0Q26 NateCo-movementismeasured by thefirst prticipal component Soerce:World Bank staff estimates. • Since 1993, however, there has been a decline a statistical sense, is nor expected to reverse), from in the co-movement of ptfolio flows to both which the temporary or cyclical coymponenrs are regions, suggesting that country-specific fac- measured. tors are becoming more important. • Te decline in the co-movement of flows after a The results show that the cycical component 1993 is especially marked for E-ast Asia. for bond flows is higher than for equity flows in both regions. On average, during 1990.95, Portfolio -flows may also be susceptible to domes- 40 percent of portfolio bond flows to Latin tic cyclical or temporary factors. In order to assess - America and 16 percent of bond flows to Asia the relative importance of cyclic-A factors (whether were temporary or cyclical. For equity flows of international or domestic origin} and structural the proportions were 13 percent for Latin factors in driving private flows, we decomposed America and 5 percent for Asia. portfolio bond and equity flows from the United a Portfolio flows to Latin America show a much States to Latin America and Asia into their trend/ higher degree ofcyclicality. cycle components. Since the structural factors that a Despite the relatively high degree of cyclical- may be underlying private capital flows-such as ity, there is a clear upward structural trend in global financial innovarions or productivity im- portfolio flows to both regions. The structural provemnents in recipient countries--do not occur in trend in flows to LatinAmerica begins around a predictable manner, they would be ill captured by 1992-93. Although flows to Asia show an up- a deterministic trend. We therefore used the Bev- ward trend that began earlier, the rise in eq- eridge-Nelson methodology, which entails fitting a uity flows is more dramatic from 1992 stochasic trend (the component of the flow that, in onward. (Box continues on thefollowing page.) 83 ,,CAPITAL FLOWS TO DEVELOPING COUNTRIES Box Figjre 2.2 Actual and Permanent Components of U.S. Portfolio Flows to Latin America and Asia, 198495 Bond flows to Latin America Equity flows to Latin America Bllions of dollars Billions of dollars 40 _ - 35 Actual Trend i 30 ------- - 25- 25- . 2 25 -?--X---r~ - l 4f 0 - t- +- >4t- 'Te ist 5L - 4 t3-^ ,-- =§i Acua 5 _-> -<--.. 0 __ __ __ M_ S__ _______ 0--~~~~~~~~~~~~~~ 1984 1987 1991 1995 1984 1987 1991. 1995 Bond flows to Asia Equity flows to Asia Billions of dollars Billions of dollars 35 - i t f Trend 3 35^ - - - '0 A0 0 25 ' ' . w: ,, 5 , -'' -,: 40- 19834 1987 1991 1995 ±984 1987 1991. 1995 Source World Bank staff estimaes. 84 THE NEW INTERNATIONAL ENv * Countries with the strongest economic fundamentals, such as a high investment-to-GDP ratio, low inflation, and low real ex- change rate variability-factors that affect the long-term rates of return to investors-have received the largest flows as a propor- tion of domestic GDP. At the other end, countries with very weak fundamentals have not attracted private flows at all. * Global interest rates are not significant in explaining FDI flows, which have been the largest component of private flows to devel- oping countries. These flows are more sensitive to countries' macroeconomic fundamentals. * International interest rates were an important (that is, statistically significant) factor in driving other private capital flows to devel- oping countries during 1990-93.2 * More recently, the relative role of international interest rates has declined and country-specific factors have become more impor- tant. There is, moreover, a sizable difference among regions. Country-specific factors have become particularly important in East Asia. W '~hile there has been a relative decline in the sensitivity of port- folio flows to interest rates, portfolio flows remain quite suscepti- ble to cyclical or temporary factors-including movements in global interest rates and changes in domestic interest rates. • Although they are still quite cyclical, portfolio flows to both Asia and Latin America show a clear upward trend since 1992-93. Thus, even in portfolio flows, other factors are at work. The Structural Forces Driving Private Capital Flows to Developing Countries T HE STRUCTURAL TREND NOW EVIDENT IN PRIVATE CAPITAL flows is being driven by two primary forces: higher long-term (as opposed to short-term or cyclical) expected rates of return in developing countries and the opportunity for risk diversification. Higher Expected Rates of Return Standard economic theory predicts that if the level of capital stock is relatively low, then, other things being equal, the marginal product of 85 T AftAI FLOWS TO DEVELOPING COUNTRIES As figure 2.2 shows, policy reforms capital will be high. If not constrained by the ovailability of skilled in developing countries have labor, infrastructure, and other factors that are coinplements to capital strengthened economic performance and creditworthiness. in the production process, therefore, the rate of ieturn to investment will be relatively high in countries with low levels of capital stock.3 For the foreign investor, country creditworthiiness, or a country's ability to make resources available for external payments, is also a very important determinant of the overall rate of return to investment.4 With the onset of the international debt crisis, the early 1980s saw a dramatic decline in the macroeconomic performance and creditworthi- ness of developing countries. This was due, in part, to a deteriorating external environment, which included a sharp inciease in international interest rates and recession in industrial countries. In the mid-1980s however, the macroeconomic performance and creditworthiness of many developing countries started to improve again and this trend accelerated in the early 1990s. Developing coun- tries that have been the major recipients of private capital have seen a decline in inflation, higher growth of output and exports, and higher and more productive investment (figure 2.2). The more stable domes- tic macroeconomic environment has, in turn, irmproved prospective rates of return to investment in general, while the growth in earning ca- pacity (as manifested in the growth of output and exports), and reduc- tion in the stock of external liabilities in many of the heavily indebted middle-income countries (following the implementation of the Brady Plan), has reduced country risks for the foreign investor.5 Many of these countries have also seen a significant growtht in the skilled labor force and improvements in supporting infrastructure over the past decade. Underpinning the improvements in economic performance of de- veloping countries has been the systematic adoption of macroeconomic stabilization programs and structural reforms by a growing number of countries and a more favorable international environment. A key ele- ment of the stabilization programs has been sustained fiscal adjust- ment, with fiscal deficits declining substantially i-rom the high levels reached after the debt crisis (figure 2.2). Trade liberalization, invest- ment deregulation, and financial sector liberalization have promoted more private sector activity and outward-oriented economies. The creditworthiness of these countries has been strengthened by the low international interest rates that have prevailed since the mid-i 980s and was further boosted during the cyclical downturn in the early 1990s.6 86 THE NEW INTERNATIONAL ENVT.1 Figure 2.2 Effects of Policy Reforms in Developing Countries Receiving Large Private Capital Flows (percent) A. Economic Growth B Fiscal Deficits 10 Average fiscal balance/GDP (percent) 0 D Output growth - Brady countries Other major recipients 8 * Export growth -1 of private capital flows 8~~~~~C Dett Export growth -1Cuty rdtorhns EInvestment growth -2 6 -3 4 -4 2 -5 0 6 1986-90 1991-95 1980 1983 1986 1989 1992 1995 Note This represents the average performance of major recipients of Source IMF, World Economic Ourlook data base capital flowas Source World Bank staff estimates C. Debt-to-Export Ratios D. Country Creditworthiness Average debt/exports (percent) Index 400 44 350 42 300 40 250 38 200 36 150 34 100 32 50 Other major recipients of private capital flows 0 30 - 1980 1983 1986 1989 1992 1995 1985 1988 1991 1994 Note This represents the average of major capital inflow recipients Source World Bank data. Source Data from Institurional Investor Credit Ratings 87 IACITAL FLOWS TO DEVELOPING COUNTRIES These reforms have meant that unlike the growth of the late 1970s, which was largely inward-oriented and led by public investment, the growth of developing countries in the 1990s is more broadly based and is led by exports. What is also noteworthy is that [he improvements in economic performance and creditworthiness are being shared by a growing number of developing countries, although the process is still in its early stages in many countries (figure 2.3). Although the risks of investing in emerging inarkets remain rela- tively high, the policy reforms are resulting in the progressive decline in A growing nuimber of developing these risks and in improvements in expected rates of return. Investors countries are gaining the benefits of have thus begun to respond to the relatively highei expected rates of re- improved economic performance. turn in developing countries.7 Figure 2.3 Output and Export Growth, Selected Regions, 1986-90 and 1991-95 (percent) Output Growth Exports Growth 12 25 - Middle East and East North 10 Asia Africa 20- Middle East 4_ 1Eas andl9 9 - --- --_10-i 1_9r ---- Asi -Tota-l North South Latin Africa 15 _ Asia Americ I I ~~~~~~ ~~~~~~~~~Sou~th America 2 1986-90 1986-90 0 ___________________1991-95 1991-95 986-90 O _ ~~~~~~~~~~~~~~~~~~~~~~~1S 91-95 1986-90 1986-90 ° - 1991-95 9_ -2 1991-1)5 , ~~~~~~~~~~~~~~1991-95 2- . 1991-95 1986-90 1986-90 -4 1986-90 -5 1986-90 Standard deviation Median Standard deviation Note As the figure shows, not only has the median growtth rate of output and exports increased, but this growth has been shared more equitably across countries-thar is, the dispersion around the mean has declined in several regions (particularly in export growth) in the past five years, rela- tive to the late 1980s Source World Bank staff estimates 88 THE NEW INTERNATIONAL ENVFI Opportunities for Risk Diversification The second force behind the structural trend in private capital flows is in- vestors' desire for portfolio risk diversification. Investors can benefit from holding emerging market equities because returns in emerging markets tend to exhibit low correlations with industrial country returns-that is, they tend not to move in tandem with those of industrial countries. In general, by holding an asset whose returns are not correlated with the re- turns of another asset, investors can raise the overall return on their port- folio without a commensurate increase in risk (variance).8 The opportunity for portfolio diversification offered by emerging markets is a relatively recent phenomenon, associated with the 1 990s, that has developed as capital markets in these countries have deepened and broadened.9 As discussed in chapter 6, by the end of 1994, the combined market capitalization of the 18 major developing countries that form the IFC Emerging Markets Index (IFCI) was, at $1.2 trillion, 13 times higher than it had been in 1985.10 Although still much lower than that of industrial countries, the average market capitalization of these countries rose dramatically from 7 to 42 percent of GDP during the period, while turnover ratios increased approximately twofold be- tween 1985 and 1994.11 As a result of this growth, financial markets in developing countries are now beginning to provide foreign investors with significant oppor- tunities to diversify. As a hypothetical illustration, assume that in- vestors' holdings of international assets are allocated according to the country shares implied by the MSCI-EAFE Index.12 Given the current correlations of returns between the IFC Investible Index and the MSCI- EAFE Index, investors could both increase the expected rates of return and reduce the risks in their overall portfolio by increasing their alloca- tions to emerging markets, according to the proportions implied by the IFCI, to 41 percent of this portfolio of international holdings (figure 2.4). This compares with current allocations of around 12 to 14 per- cent of investors' international portfolios to emerging markets. More- over, given that returns among emerging markets exhibit low (and often negative) correlations with each other (figure 2.5), greater diver- sification within emerging markets than that implied by the overall IFCI could reduce portfolio risks even further.13 Of course, this calculation is only hypothetical: such a large reallocation would be an upper bound of possible gains and would not even be feasible under present market 89 TA~1TAL FLOWS TO DEVELOPING COUNTRIES condltions. In addition, correlations between developing and industrial countries are likely to rise over time as the former become more finan- cially integrated with the global economy. 4 Non etheless, it illustrates the potential for risk diversification that developing countries offer. The Changing Enabling Environment in Industrial Countries The strong response of private capital flows to the two primary forces described above has, in large part, stemmed frorn changes in the en- abling environment in both industrial and developing countries during the 1980s and 1990s. Industrial countries have seen changes in two broad areas. First, in the real sector, increasing competltion and rising Figure 2.4 Potential Benefits from Portfolio Risk Dhiersification (MSCI-EAFE and emerging markets) Annualized expected returns in percent 16 100 percent in IFCI 14 12 10 59 percent in MSCI-EAFE * 41 percent In IFCI 6 0 percent in IFCI Emerging markets offer international investors significant potential for portfolio diversification. 4 15 16 17 18 19 20 Annualized standard deviation of returns (risk) in percent Source World Bank staff estimates using the [Fc Investable IndeN and the Morgan Stanley Capital International Index (Msci-EAFE) 90 THE NEW INTERNATIONAL ENvuN, Figure 2.5 Correlations of Returns among Emerging Markets Mexico Argentina Bral'I Chile Malaysia Thailand India Indonesia Pakistan Philippines Colombla Jordan Turkey Sn Lanka MexICO Low: -0.005 to 0.25 Argentina - 1 Medium: 0.26 to 0.5 High: 0.5 to 1.0 Brazil 1 Chile 1 Malaysia _ - - 1 Thailand _ _ 1 India _ -_ --- 1 Indonesia - - 1 Pakistan - _ - 1 Philippines _ - - - - - 1 Colombia _ _ _ _ _ _ _ _ _ 1 Jordan Turkey _ - _ __ _ __ _ _ Sri Lanka Note Correlations computed using lFc Emerging Market Invstable Index returns during January 1990-August 1996 Source World Bank staff estimates costs at home, combined with falling transport and communications The low correlation of returns among costs, have heightened firms' responsiveness to opportunities to in- emerging markets suggests that crease efficiency and reduce costs by locating investments abroad. This investors can benefit from diversifi- is leading to the progressive globalization of production and has cation among emerging markets. spurred the growth of "efficiency-seeking" EDI. Second, in financial markets, a self-reinforcing process of competition, deregulation, tech- nological advances, and financial innovations has increased the respon- siveness of investors to international investment opportunities. This process is rapidly leading to the linking of domestic markets into one global market. These two forces are discussed below. 91 zT_APITAL FLOWS TO DEVELOPING COUNTRIES The globalization of production. The 1980s and 1990s have witnessed the progressive globalization of the production process. Competitive pressures from unilateral and successive rounds of multilateral trade liberalization, and stagnant demand combined with rising costs at home, have encouraged firms to seek new markets and increase effi- ciency. Initially, this involved locating the full range of production activities in a low-cost country or moving them to one. More recently, the drive for increased efficiency has involved breaking up the produc- tion process into discrete segments, each located in the best place in terms of cost and productivity considerations. This process is spurring the growth of "efficiency-driven" foreign investment flows, which encompass a wide range of corporate functions and take place in a broad number of industries with different levels of factor or skill intensity. The globalization of production has been made possible by recent technological changes and reductions in transport and communica- tions costs. Toyota, for example, has rationalized its production on an ASEAN-wide basis, with affiliates in each country specializing in the pro- duction of different parts that are subsequently exported to the country where final assembly takes place. Similarly, Generai Motors plans to es- tablish a materials and components purchasing office in Poland for all its European affiliates. This process is, moreover, self-perpetuating: heightened competi- tion is driving firms to invest abroad, basing their plans on efficiency considerations, and growing FDI is making firms more competitive, thereby intensifying the competition. Consequenmly, firms are becoming increasingly responsive to new opportunities that can strengthen their com- petitiveness; in particular, they are looking to invesc in markets that offer macroeconomic stability, supportive regulatory frameworks, well-de- veloped infrastructure (transport and telecommunications), low costs in relation to productivity of the labor force, and more open trade regimes. As a result, a significant proportion of current global FDI flows can be characterized as efficiency-seeking. The importance of this type of FDI is evident from the fact that the sales of foreign affiliates to the par- ent transnational corporation and to other affiliates of the same parent company, as a share of worldwide sales of affiliates, are high and have increased somewhat over the past decade. Moreover this increase has been more pronounced with regard to the sales of affiliates in one host 92 THE NEW INTERNATIONAL ENV-A country to affiliates in other host countries-corresponding with the specialization in production process implied by the more recent type of efficiency-seeking FDI. The growing responsiveness of financial markets. The financial markets in industrial countries have also changed significantly during the 1980s and 1990s. Driven by the self-reinforcing process of competition and financial innovation, along with deregulation and technological change, they have become increasingly global in nature. Foreign exchange markets were the first to internationalize in the late 1970s, followed by bond markets in the 1980s and equity markets in the 1990s. The internationalization of foreign exchange markets began in the late 1970s, after the collapse of the Bretton Woods system of fixed ex- change rates in 1973. The internationalization of bond markets, how- ever, gathered momentum only during the 1980s, when low inflation and positive real interest rates and yield curves made long-term bonds appealing to investors. At the same time, banks' funding costs rose for a variety of reasons. The general deflation in the 1980s and the inter- national debt crisis placed pressures on the performance of banks' assets and resulted in a slip in their credit ratings and a rise in their funding costs. Together with their high intermediation costs associated with re- serve requirements, capital, and overheads, this resulted in relatively high lending rates (Honeygold 1987). Consequently, prime borrowers such as governments and large corporations found it cheaper to raise funds directly from investors through the securities markets. The inter- nationalization of securities markets began with the strengthening of the offshore Eurobond market. Because the Eurobond market was ex- empt from many of the regulations of domestic markets-especially with regard to taxation-prime issuers could usually raise funds at lower costs than in domestic markets, while investors often received higher rates of return than they did in their own regulated domestic markets. Largely in response to market pressures, governments began to deregulate domestic financial markets by the mid-1980s. This con- tributed to greater convergence of issuing costs between offshore and onshore markets, thereby encouraging corporations and governments to seek capital in the major domestic securities markets while leading to the progressive internationalization of the latter. Japan, for example, re- laxed regulations on the Samurai bond market in 1983, issued the first 93 JT2ANTAL PLOWS TO DEVELOPING COUNTRIES Shogun bond in 1985, and relaxed restrictions orn the holdings of do- mestic and Euro Yen commercial paper by nonresidents in 1988. The United States eliminated its 30 percent withholding tax on foreigners' interest income in 1984, the same year that Germnany stopped taxing foreign investors' income from bonds. Germany also allowed foreigners to buy federal bonds in the primary market in 1988 and, a year later, eased restrictions on deutsche mark bonds. (Annex 2.1 lists the key do- mestic and international financial deregulations.) Equity markets started to globalize much more recently, essentially in the early 1990s. They have been slower to globalize for several rea- sons. First, equities are much less liquid than bonds, in part because the valuation of the equity of a company is very specific to the circum- stances of that company, making shares intrinsically more difficult to trade. Second, the information and infrastructure needed for investors to undertake global investments in equities (for example, comparable accounting standards and global settlement and custodial services) are only now being developed (McKinsey Global Institute 1994). What has driven the process of internationalization of equity markets is the more active stance of institutional investors, as discussed in more detail below. Neither multinational commercial banks (an important force in the internationalization of the bond markets) nor the issuers them- selves (that is, multinational corporations) have played a significant role in the process. In fact, the relative illiquidity and volatility of equities has meant that multinational commercial banks, which are highly leveraged insti- tutions, have generally been averse to holding equities (foreign or do- mestic) in any significant amounts. Multinational corporate issuers have been relatively less interested in listing abroacd, because the price of equities is generally driven by investors in their own countries, who hold the bulk of the company equities and know most about them (McKinsey Global Institute 1994). As discussed below, however, insti- tutional investors, who during the 1980s concentrated on domestic eq- uities and held the bulk of their international investments in bonds and currencies, are now rapidly increasing their holdings of foreign equities. At the same time, governments have been opening domestic stock mar- kets to foreign investors and issuers (annex 2.1). This is providing the impetus for the internationalization of domestic stock markets. Financial innovations during the 1980s and 1990s have also played a key role in the internationalization of financial markets. Specifically, 94 THE NEW INTERNATIONAL--N the last two decades have seen the growth of foreign exchange and fixed income and, more recently, equity-related derivatives. These innova- tions can lower funding costs, enhance yields, or unbundle some of the characteristics of securities, such as their price risk, credit risk, country risk, and liquidity, to tailor portfolios to the needs of different investors to hedge price, interest rate, and exchange rate risks. As a result, these innovations have made it more attractive for borrowers to raise capital in foreign markets and for investors to make cross-border investments. For example, the use of interest rate swaps in conjunction with cur- rency swaps has resulted in an increasingly global bond market. And, more recently, financial innovations have been promoting the interna- tionalization of equity markets. Foreign investors are now using equity swaps, in which a domestic agent passes on the gains or losses from holding domestic equities to the foreign investor for a fee. This allows the foreign investor to avoid having to pay for high local execution costs or falling victim to insider trading practices (McKinsey Global In- stitute 1994). (Annex 2.1 lists the adoption of key financial innova- tions in securities markets.) Technological advances have reinforced the effects of deregulation and financial innovations in internationalizing markets by increasing ef- ficiency in gathering and disseminating information and in processing transactions. In particular, low-cost telecommunications have been in- strumental in linking financial markets and have made possible 24-hour trading, which, in turn, has brought greater breadth and depth to trad- ing. It can be argued that improved communication also encourages fi- nancial institutions to continue to develop new instruments to meet the needs of customers in previously isolated markets (Honeygold 1987). The evolution in electronic technology has greatly enhanced the ef- ficiency of stock markets, just as the Euroclear and Cedel standard clearing mechanisms have provided low-cost dealing and delivery in the Eurobond market. For example, NASDAQ (National Association of Securities Dealers Automated Quotation), which originated as an over- the-counter market for smaller firms unable to meet the stringent re- quirements and high listing costs of the major exchanges, now provides for private firms to make block sales by linking their customers to- gether through computer terminals, at high speed and with low trans- action costs. Moreover, for small-scale operations, it has been possible, since 1984, to complete instantaneous transactions of orders of up to 500 shares via the Small Order Execution System. In Japan the CORES 95 CAPITAL FLOWS TO DEVELOPING COUNTRIES (Computerized Order Routing and Execution System) handles all but the largest stocks (which are still traded on the trading floor). In the United Kingdom, SEAQ International (Stock EKchange Automated Quotations System), introduced in 1985, is linked to NASDAQ in the United States and provides mutual on-screen access to the top 300 quotations in each market. Market makers can thereby display current prices around the world, improving the efficiency of trades. The growing importance of institutional investors. The other aspect of change in the enabling environment in industrial countries has been the growth of institutional investors. These investors, both able and willing to invest abroad, have increased the magnitude of the response of private capital flows to the fundamental forces driving these flows-that is, cyclical and long-term relative rates of return and new opportunities for portfolio diversification. The growing importance of institutional investors has been the re- sult of the same forces-competitive pressures, ceregulation, techno- logical advances, and financial innovations-that have affected the markets on the issuing side. One type of innovation-securitization (broadly defined as the process of matching savers and creditors through financial markets as opposed to closed market credits via banks and other financial institutions)-has played a particularly important role (Gardener 1991). Specifically, securitization has meant the cre- ation of instruments (including the conversion of loans into securities, or secondary securitization) that can be issued and traded directly on market. Because securitized assets are more cost-effective than bank loans, they have facilitated the growth of institutional investors that trade these assets at the expense of commercial banks, whose primary business remains in making loans. In particular: * Mutual funds, which had already grown substantially in the 1950s and 1960s, attracted more savings and thereby became more important in the financial markets during the 1970s by of- fering money market funds. Investors found these funds an at- tractive alternative to the regulated deposit rates of commercial banks when market interest rates rose. In the 1980s, following the success of money market funds, mutual funds began to invest in bonds, both domestic and international, andJ gradually in domes- tic equities. This both reinforced and was reinforced by the secu- ritization that was taking place. 96 THE NEW INTERNATIONAL ENV, U Pension funds have also become increasingly important because of broader pension coverage and-particularly since the 1970s- the rising value of contributions. The growth in pension assets was especially stimulated by changes in pension and tax laws. The United States, for example, allows companies to deduct their con- tributions to employee pension plans from their taxes. Moreover, employee contributions are not taxed, and interest on pension as- sets is not taxed until retirement. Therefore, both employers and employees have an incentive to save through a pension plan, as opposed to other forms of savings, which are taxable (Sellon 1992). Given the long-term nature of their commitments (that is, to pay retirement benefits), pension funds tend to invest in long- term instruments, including corporate equities and long-term bonds. As in the case of mutual funds, therefore-indeed, ar- guably more so-the growing importance of pension funds and the process of securitization have been mutually reinforcing. There are fundamental advantages offered by institutional investors that explain their appeal to individual investors and suggest that their role will continue to expand. Pension and mutual funds provide indi- vidual investors with a low-cost method of diversifying their portfolios: pooling funds with many other investors to purchase a number of dif- ferent assets. At the same time, technological advances have greatly re- duced the costs of dealing with a large number of investors. This trend away from banks and self-directed investment to institu- tional investors is very evident in the three countries that are major sources of funds-Japan, the United Kingdom, and the United States (figure 2.6). In the latter two, where this trend has been most pro- nounced, the share of household savings channeled to mutual and pen- sion funds doubled between 1975 and 1994. However, continental Europe, with the exception of the Netherlands and Switzerland, has seen much more modest shifts to institutional investors. These varying trends largely reflect country-specific factors such as the pace of dereg- ulation and tax policies. The variation across countries notwithstanding, there is a clear trend toward the institutionalization of savings in industrial countries. Alto- gether, pension funds, insurance companies, and mutual funds in seven major industrial countries had assets close to $17 trillion in 1994, com- pared with $5.3 trillion in 1985.15 Institutional investors now domi- 97 ikP-TAL FLOWS TO DEVELOPING COUNTRIES Figure 2.6 Institutionalization of Savings (assets as a percentage of GDP) Canada _ Pension funds 221 Insurance companies 155 1985 Mutual funds 93 France - 61 - 1994 450 497 Germany _121 590 273 Japan- - ----- 1,118 2,036 437 Netherlands --_ _ ~ ~~~~~ 131 62 United Kingdom 743 ' *__ ~~848 _ ~~214 A growing proportion of household States 3,760 savings is being channeled through 2,6ro institutional investors. _ g_2,161 0 20 40 60 80 100 Note Figures refer to the size of assets in billions of dollars For the United States, assets refer to U S open-end mutual funds only Source Data from InterSec, IMF, International Financial Statistics data base, the Investment Company Institute, the Investment Funds Institute of Canada, the [nvestment Trusts Association of Japan, the U K Association of Unit Trusts and Investment Funds; and the Nederlandsche Bank nate the financial landscape, especially the capital markets. In the United States, for instance, institutional investors are now estimated to account for more than 49 percent of U.S. equities, compared with 16.5 percent three decades ago. One reason for the success of pension and mutual funds is that indi- vidual investors are offered the benefits of professional management. Through the expertise of specialized investment advisers or fund man- agers, indivldual investors can realize higher returns from a more diver- sified and international portfolio than they couMd get by themselves, 98 THE NEW INTERNATIONAL EN*, without having detailed knowledge of the countries and individual companies issuing the securities. The trend toward international diver- sification by institutional investors has been particularly pronounced in the United States, but it is also happening in other major industrial countries to varying degrees. A distinction should be made between the international behavior of pension funds and of mutual funds, though. Mutual funds, driven by profits and subject to relatively few regulations, increased their interna- tional exposure much earlier and have always had a higher proportion of international assets. They have not, consequently, internationalized much more during the 1990s, except in the United States, where the international share of the mutual fund portfolio rose from 3.8 percent in 1990 to 8.9 percent in 1994. Pension funds, on the other hand, have always been heavily regu- lated because of their fiduciary responsibility to deliver promised bene- fits. Consequently, they have tended to be more cautious in their investment strategies. Moreover, given their orientation toward long- term investments, they have tended to focus on long-term instruments. The international diversification of pension funds did not begin until the 1980s, when the long-term securities markets became internation- alized and as governments began to deregulate the investment alloca- tions of pension funds. The trend toward greater international diversification by pension funds has been common to most industrial countries, with the excep- tion of Germany and a few other European economies (such as Nor- way, Spain, and Sweden). The degree of international diversification among pension funds varies significantly, however, ranging from 24 percent of total assets in the United Kingdom to around 17 percent for Canada and the Netherlands, 9 percent for the United States, and a low of 5 to 6 percent for France and Germany (figure 2.7 ). Moreover, with the exception of Japan and the United States, the international expo- sure of pension funds remains significantly lower than that of mutual funds. Institutional investors of the third type, insurance companies, generally have had even lower international exposure. One reason for their prefer- ence for domestic assets may be their need to match assets and liabilities- including currency composition-in the short term (Davis 1991). However, with the exception of those in Japan and the United States, even insurance companies have seen small increases in international assets. 99 APITAL FLOWS TO DEVELOPING COUNTRIES Figure 2.7 llnternational Diversification of Institutional Investors, Selected Countries, 1990 and 1994 (percentage of total assets) Canada Japan Pension | 1990 Pension | 1990 funds M 1994 funds M 1994 Insurance Insurance companies I companies Mutual Mutual funds funds 0 10 20 30 40 50 60 0 10 20 30 40 s0 60 France Netherlands Pension M 1990 Pension | 1990 funds M 1994 funds M 1994 Insurance L Insurance companies companies Mutual P Mutual funds e funds 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Germany United Kingdom Pension l 1990 Pension 1990 funds M | 1994 funds 1994 Insurance IE Insurance companies companies Mutual Mutual funds funds 0 10 20 30 40 50 60 0 10 20 30 40 50 60 United States Pension i 1990 Pension and mutual funds _ 1994 funds and insurance companies are Insurance companies _ increasing their c international invest- ments, but tEle propor- Mutual ments, ~~~~~~funds tions vary by country. 0 10 20 30 40 50 60 Notes For Fiance the data on mutual funds are for 1993 Data on insurance companies for Canada and the United States are for life insurance only Source World Bank staff estimates using data from InterSec, the WM Company, Nederlandsche Bank, the Americar Council of Life Insurance, Bank of Canada Review, the Investment Company Institute, the U K. Association of Unit Trusts and Investment Funds, the Investment Funds Institute of Caniada, the Investment Trusts Association of Japan 100 THE NEW INTERNATIONAL ENVY1R Overall, the combination of the growth in the asset base of institu- tional investors and the growing internationalization of these assets has meant a rising volume of international investments by institutional in- vestors. For example, total assets of pension funds at the global level are estimated to have increased from $4.3 trillion in 1989 to $7 trillion in 1994. At the same time, the share of international investment in their portfolios rose from around 7 percent in 1989 to 11 percent in 1994. Together, this has resulted in an increase in total international invest- ments by pension funds from $302 billion in 1989 to $790 billion in 1994, with the growth in asset base contributing to around 40 percent of the increase in international investments, and greater international diversification contributing around 60 percent of the increase. The Changing Enabling Environment in Developing Countries Changes in the enabling environment of industrial countries have meant that economic agents in these countries-both firms and port- folio investors-have become more responsive to opportunities to earn higher rates of return or diversify risks through international invest- ments. And, as discussed in the previous section, developing countries have begun to offer investment opportunities as their creditworthiness and rates of return have improved. Concomitant changes in the en- abling environment of developing countries have enabled these forces to be translated into actual investments. The most important enabling factors are simply whether private cap- ital is permitted to flow into a country and whether the restrictions on the repatriation of profits (income) and capital are prohibitive. Along with the sharp decline in expropriation risks, the 1980s and 1990s have witnessed a progressive dismantling of barriers to capital account mobil- iry in developing countries.16 During 1991-93, 11 developing coun- tries undertook full or extensive liberalization of their exchange restrictions, 23 liberalized controls on FDI flows, 15 eased controls on portfolio inflows, and 5 eased restrictions on portfolio outflows. By the end of 1995, 35 developing countries had fully open capital accounts. Trade liberalization associated with the Uruguay Round has received much attention, somewhat eclipsing significant progress made in par- allel on investment treaties, relevant for FDI. In fact, half of all invest- ment codes and bilateral treaties have been drawn up in the 1990s, as have several important multilateral agreements.i7 101 F41TAL FLOWS TO DEVELOPING COUNTRIES Whereas previously most national investment codes and bilateral investment treaties imposed few restrictions on the recipient countries with respect to market entry, recent laws and agreements have em- phasized the free flow of investment. Many of the recent laws and agreements also contain provisions for the settlement of disputes, usu- ally providing for several different mechanisms for their resolution- ranging from direct negotiations between the disputing parties to arbitration proceedings in which investors and host states may partic- ipate on an equal footing (World Bank 1997). The liberalization of restrictions on portfolio capital has also trans- lated into significant changes in the stock markets of developing countries. As recently as the beginning of 1991, only 26 percent of emerging stock markets could be categorized as having free entry for foreign investors, while 11 percent were closed to foreign investors. By the end of 1994, 58 percent of all stock markets had free entry for foreign investors, while only 2 percent remained closed (figure 2.8). In addition to the easing of regulations pertaining to the move- ment of private capital, structural changes in cLeveloping countries have meant a significant expansion of areas for potential foreign in- vestor involvement. In particular, as part of their structural reform programs, many developing countries have deregulated their invest- ment regimes and reduced the role of the public sector in directly pro- ductive sectors, with a view to allowing greater participation of private investors. One manifestation of this has been the gradual reduction of the share of state-owned enterprise investment in total domestic in- vestment. For the countries that have been the primary recipients of private capital flows, the average (unweighted) share of state-owned enterprises in gross domestic investment has fallen from around 25 percent in the late 1970s to 17 percent in 1991. At the same time, countries have deregulated to allow greater foreign participation in these sectors. Foreign investor involvement in developing countries has been fur- ther boosted by the privatization of state-ownec. enterprises. Of the $112 billion of privatization proceeds that developing countries re- ceived during 1988-94, almost 42 percent was from foreign investors. The early privatizations were largely in the form of FDI. But there has also been a steady increase in the participation of portfolio investors. Indeed, in 1994, portfolio investors accounted for over 50 percent of the foreign participation. 102 THE NEW INTERNATIONAL EN Figure 2.8 Entry Restrictions for Foreign Investors in Emerging Stock Markets, 1991 and 1994 (percentage ofstock markets) 60 D 1994 50 1991 40 30 20 10 Stock markets in emerging economies are allowing more access 0 to foreign investors. Closed Restricted Relatively free Free entry entry Note Free entry-no significant restriction to purchasing stocks Relatively free entry-some registration procedures required to ensure repatriation rights Restricted-foreigners restricted to certain classes of stocks or only approved foreign investors may buy stocks Closed-closed or access severely restricted Source 1FC, Emerging Stock Markets Factbook 1996 Finally, trade liberalizations by developing countries have provided an impetus to FDI flows in particular. As discussed earlier in this chap- ter, a large proportion of FDI flows is being driven by considerations of production efficiency. As a result, trade regimes in developing countries have become more important. Following unilateral and successive rounds of multilateral trade liberalizations since the mid-1980s, there has been a progressive dismantling of trade barriers in developing coun- tries. In fact, by 1994, 42 developing countries could be categorized as having an open regime (Sachs and Warner 1995).18 103 -CAPITAL FLOWS TO DEVELOPING COUNTRIES The Outcome of Structural Changes: Growing Investment in Emerging Markets C HANGES IN THE ENABLING ENVIRONMENTS OF BOTH INDUS- trial and developing countries have made private capital more responsive to underlying forces that are spurring investments in developing countries. Of all types of capital flows, FDI has responded most vigorously. The driving factor for FDI has been the sustained improvement in domestic economic fundamentals. For portfolio flows, institutional investors have been the driving force behind the surge, especially in portfolio equity. Although the initial impetus was the cyclical decline in global interest rates, as improve- ments in economic fundamentals and creditworthiness began to take hold and the new investor base has become more familiar with emerg- ing market investments, the impetus has become the long-term rates of return and opportunities for portfolio risk diversification. As a result, institutional investors are the new, and increasingly important, segment of the investor base in emerging markets. Accordingly, this section focuses primarily on the new investor base and the growth of their investments in emerging markets. Foreign Direct Investment Developing countries have seen an almost fourfold, increase in FDI flows in just five years-from $25.0 billion in 1990 to more than $95 billion in 1995. Indeed, the growth of FDI flows to developing countries has been much faster than that to industrial countries. Developing coun- tries' share of global FDI flows has risen from 12 percent in 1990 to around 38 percent in 1995. The nature of the FDI flows to developing countries has changed sig- nificantly. In the 1970s and early 1980s, resource extraction and im- port substitution were the primary motives for FDI to developing countries. In contrast, a high proportion of current FDI flows to devel- oping countries can be characterized as efficiency- seeking investments, associated with the globalization of production. Initially directed to- ward basic manufacturing, these flows are now increasingly going into high-value-added and skill-intensive manufacturing sectors. Develop- ing countries are also increasingly seeing FDI In services sectors, includ- 104 THE NEW INTERNATIONAL ENVPVt ing the provision of infrastructure services. Opportunities for invest- ments in services and infrastructure have expanded significantly as a re- sult of the stronger economic growth and investment deregulation in developing countries. The changing nature of FDI flows to developing countries is reflected in the regional destination of FDI flows and the relative selectivity in terms of country destination. East Asia has accounted for about 50 per- cent of FDI flows to developing countries in the 1990s. As mentioned earlier, efficiency-driven flows are sensitive not only to the strength of countries' macroeconomic fundamentals and domestic market consid- erations, but also to the supporting infrastructure (both regulatory and physical) and the quality and productivity of the labor force in relation to the cost. East Asian markets have therefore been primary reciplents of such flows. East Asia, with its strong economic growth, is also pro- viding growing opportunities in services and infrastructure. Latin America has been the second-largest recipient of FDI flows (accounting for 28 percent of FDI flows). In the early 1990s, a sizable proportion of these flows were the result of one-off privatizations, but countries in Latin America are also receiving efficiency-driven FDI. Portfolio Flows Despite a dramatic increase of portfolio flows from institutional in- vestors into emerging markets, information on emerging market place- ments by such investors remains fragmentary. For mutual funds, the available information suggests the following sequence. Initially, in the mid-1980s, investment in emerging markets was in the form of closed- end funds, including country funds, which pioneered the flow of pri- vate investment to emerging stock markets. Closed-end funds are well suited to emerging market conditions, since they automatically regulate redemption risk, which can be especially large in less liquid markets. As emerging markets have become more established, more and more open-end emerging market funds have been set up. In the third stage, emerging markets have begun to figure in the allocations of interna- tional investment funds, and finally of global funds. In all three stages, however, mutual fund investments in emerging markets have remained highly skewed toward portfolio equities, with debt-oriented mutual funds accounting for less than 10 percent of mutual fund assets in emerging markets. 19 105 tLAI?TAL FLOWS TO DEVELOPING COUNTRIES The growth in mutual funds over the last decade has in fact been dramatic. In 1986, there were 19 emerging market country funds and 9 regional or global emerging market funds. By 1995, this had ex- panded to 505 country funds and 773 regional and global emerging market funds. Initially these were mainly closecd-end funds, but by 1995 around 50 percent of all funds were open-end. The combined as- sets of all closed- and open-end emerging market funds increased from $1.9 billion in 1986 to $10.3 billion in 1989 to $132 billion in mid- 1996. In addition to the dedicated emerging market funds, interna- tional funds have also increased their allocations to emerging markets. Surveys suggest that international funds based in the United States have increased their allocations to emerging markets from a bare 2 per- cent of their portfolios in 1989 to 12 percent in 1995 and that U.S. global funds now hold around 3 to 4 percent of their portfolios in emerging markets. International and global funds together now ac- count for an estimated 30 to 40 percent of the emerging market assets held by U.S. mutual funds. These trends have meant that emerging markets are accounting for a rising proportion of international investment by mutual funds- more than 30 percent of new international investments by U.S. mutual funds went to emerging markets during 1990-94. Since international investment itself has been rising, the share of emeiging market assets in total mutual fund assets has risen quite sharply. In absolute terms, U.S.- based open-end mutual funds alone had around $)36 billion in emerg- ing markets by the end of 1995 (figure 2.9). Yet despite this impressive increase, the share of emerging markets in the portfolios of mutual funds remains small. Emerging markets still account for only about 2 percent of total mutual fund assets in the Unicled States. Although emerging market exposure of U.K. mutual funds is higher-in the range of 3 to 4 percent-mutual funds in Japan and the rest of Europe still have negligible exposure to emerging markets (figure 2.10).2O A survey undertaken for this report confirms that pension fund in- vestment in emerging markets is a relatively recent phenomenon (box 2.3). Almost 60 percent of the pension funds surveyed initiated expo- sure in emerging markets after January 1994. Interestingly, the propor- tion is exactly the opposite for the larger pension funds (those with assets greater than $1 billion)-that is, 60 percent had initiated expo- sure prior to January 1994. Larger pension funds tend to have higher exposure to emerging markets than do smaller funds, and corporate 106 THE NEW INTERNATIONAL ENVU; Figure 2.9 International and Emerging Market Assets of U.S. Open-End Mutual Funds, 1990-95 Billions of dollars Percent 400 10 * International markets 350 * Emerging markets 8 300 250 6 200 Percentage of international assets in total assets 4 150 100 2 so U.S. mutual funds have increased their investments in international and o -0 emerging markets, although their investments are still quite low. 1990 1993 1995 Source Data from Micropal, Inc, and World Bank staff estimates. and endowment funds typically have much higher exposure than do government and union funds. Although reliable data are not available, the results of two surveys suggest that U.S. pension funds currently have about 2 percent of their total assets invested in emerging markets (figure 2.1 1).21 Thus, even though pension funds in the United States began to invest in emerging markets more recently than did mutual funds, their allocations are only slightly lower. Only a fraction of these funds, however, actually have a policy of allocating investment to emerging markets. Those that do generally treat all emerging markets as a single asset class and leave country allocation to the manager of the mutual fund or to an outside manager with a specific mandate. More than half of the pension funds surveyed invest through mutual funds.22 In fact, whereas retail investors, especially high net worth 107 XA?ITAL FLOWS TO DEVELOPING COUNTRIES Figure 2.10 Emerging Market Investments of Industrial Country Mutual Funds, 1993-95 (dedicated emerging marketfunds) Billions of dollars 120 0 * Rest of the world 100 -*---- ------- ----Japan * Canada 80 - - - United Kingdom D Offshore 60--- --- El United States 404- 40 t-- ---- - ---- - - ---------- Mutual funds in most 20 ~~~~~~~~~~~~~~~~~~~~~~~~other industrial countries infvest even lower arnrounts in emerging markets than do U.S. __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _m utual funds. 1993 1994 1995 Note Offshore denotes investment channeled through a mutual fund registered in an offshore financial center Source World Bank staff estimates using data from Micropal, Inc, the U K. Association of Unit Trusts and Investment Funds, the Investment Funds Institute of Canada, the Investment Trusts Association of Japan Box 2.3 Pension Funds in Emereing Markets Pension hind interest in emnergngmarkets. Over 60 per- their portfolios in emerging market assets. This aver- cent of the pension funds invested in emerging age, however, masks wide dispariti es among types of markets started investing at the beginining of 1994. pension funds. The average for endowment funds is Large corporate pension funds have had the longest around 5 percent and that of larger corporations exposure to emerging markets, with around 8 per- around 3 percent, with a few very large corporations cent having invested in emerging markets since holding up to 7 percent of their portfolios in emerg- 1986-87. ing markets. On the other hand, the average for gov- Expo*wre in emerging markets. On average, U.S. ernment finds is currently around 1 percent. pension funds hold around 1.5 to 2.0 percent of Judging from a sample that incluc[es data from both 108 THE NEW INTERNATIONAL ENVtRI corporations and endowments, it seems that the percent). Less than 25 percent undertake quarterly larger pension funds have increased their exposure in reviews, and around 15 percent undertake monthly emerging markets from an average of around 1 per- reviews. Changes in policy allocations, however, are cent of their allocations in 1992 to just under 2 per- a lot less frequent. Over 7( percent of pension funds cent in 1995. undertake changes to their policy allocations every Inestment motiv. Around 40 percent of pension two or three years, and 13 percent every three to five funds are interested in investing in emerging mar- years. Only 3 percent change their policy allocations kers for the higher expected returns that these mar- annually. kets offer, while just under 40 percent consider both Seection and evaluation of managems. Pension finds portfolio diversification and higher expected returns that employ outside managers generally choose tobe important factors. Around 13 percent-which them on the basis of a combination of track record indudes the larger corporations-are investing tn and total asse under management. For emerging emerging markets primarily with a virew todiversify- markets, -track record' refers as much to the length ing portfolio risks. . a - Investment vehicles. Over half the pension funds of a manager's experience as to the manager s perfdo- invest in emerging markets by buying into existing mance- Around 40 percent of pension fiunds stare mutual funds. The remainder give specific mandates tiat tpe managers they selected employed a prcmar- to managers. Less than 5 percent of pension funds ily "top down" approach, with another 28 percent manage emerging market investments themselves. employing managers who have a primarily "bottom Buying into existing funds is considered better for up" approach to investing. Less than 3 percent of liquidity reasons and is also less expensive than spe- pension funds use an index alone for country alloca- cific mandates, given the current size of the alloca- tions. The bulk of the pension funds evaluate their tions in emerging markets. The majority of the managers OT the basis of performance relative to an pension funds that invest through mutual funds do index of choice as well as relative to their peers. so by buying into global funds (over 50 percent) or Around 35 percent evaluate their managers relative regional funds (23 percent). Only 5 percent buy into to indices alone. country funds. Outlook for pension fund investments in emerging mar- Types of invesbnent Almost all investments by kets. More than 43 percent of pension funds plan to pension funds are in the form of equity. Debt ex- increase their exposure to emerging markets over the posure is still minimal-with only the largest pen- next two or three years-of which the majority (60 sion funds likely to hold emerging market debt percent) plan to increase their exposure by about 20 instruments. percent over their current levels, while the remain- Policy allocaton and inestment stratea. Only a ing 40 percent expect to increase their investments small proportion of pension funds treat emerging by 50 percent. Around 57 percent of pension funds markets as a separate asset class in their policy allo- expect to maintain their current levels of exposure. caions. Most invest in emerging markets as part of The rest have recently increased their allocatons to their international allocations. Those pension funds emeri marerso that do consider emerging markets as a separate asset class in their policy allocations leave the country se- Source: Background survey undertaken by Kleiman In- lections to the fund managers. Anual review of pol- ternarional Consultants, Frank Russell Company, and icy allocations is the most common practice (over 30 World Bank staff 109 TAL TAL FLOWS TO DEVELOPING COUNTRIES Figure 2.11 International and Emerging Market Assets of U.S. Pension Funds, 1990-94 Billions of dollars Percent 400 10 I International markets 350 * Emerging markets 8 300 Percentage of international assets in total assets 250 6 200 4 150 100 2 U.S. pension funds now invest around 50 2 percent of their total assets in emerging markets. 0 0 1990 1992 1994 Source Data from InterSec and World Bank staff etimares individuals, were the primary force in the early expansion of mutual funds into emerging markets, pension funds have rapidly become the dominant source offunds for mutual funds to inmest in emerging mar- kets, especially in the United States. The remainder of pension funds (around 41 percent) give the management of their assets to outside managers, who determine country allocations. Less than 5 percent of pension funds manage their emerging market investment in-house. In absolute terms, U.S. pension funds are estimated to hold between $50 billion and $70 billion in emerging market exposure. Indeed, U.S. pension fund investments in emerging markets, including investments made on their behalf by mutual funds, have been an important factor in sustaining investment flows to emerging markets during 1994-95. As with mutual funds, most pension fund assets in emerging markets are in the form of portfolio equities. 110 THE NEW INTERNATIONAL ENYIt The Prospects for Private Capital Flows T HE STRUCTURAL FACTORS UNDERLYING CAPITAL FLOWS TO developing countries are still unfolding. The changes taking place at the international level, which are increasing the responsiveness of capital flows to cross-border investment opportuni- ties and leading to progressive global financial integration, can be expected to continue. In particular, all the forces discussed above- deregulation and breakdown of barriers, technological change, and financial innovation-continue to evolve and, as in the past, will keep reinforcing one another, creating pressures for further integration. Internal and external financial deregulation is far from complete in either industrial or developing countries. Although it is difficult to speculate on the nature of future innovation and technological change, competitive pressures and increasing integration have been stimulating investments in technology that are likely to continue to reduce transaction costs and make distant markets more accessible to small as well as large investors. Such innovations will make policy- induced barriers less effective, spurring even more deregulation and competition. As discussed, many developing countries are still in the early stages of policy reform. The policy reforms that are being embarked upon- which focus on macroeconomic stability and the promotion of more deregulated, outward-oriented, and market-based economies-are likely to increase the productivity of investments in these countries. On the expectation that developing countries will continue to strengthen and deepen their policy reforms, and that the external environment will remain broadly favorable, developing countries are likely to grow at al- most double the rate of industrial countries (figure 2.12), providing significant opportunities for productive investments. Developing countries should provide growing opportunities for port- folio investors in particular as their share of world market capitalization, which is currently around 10 percent, rises further. As financial markets in emerging economies are not fully developed, further financial deep- ening-including that of capital markets-can be expected. And be- cause developing countries are expected to grow significantly faster than industrial countries over the next decade or so, their share of world mar- ket capitalization will rise. Broadening of the base-as more developing 111 P-l ITAL FLOWS TO DEVELOPING COUNTRIES Figure 2.12 Developing and Industrial Countries' Growth, 1980-2005 Percent 6 5 Industrial countries Developing countr'es are likely to \ Developing countries grow at nearly double the rate of industrial countries. 0 1980 1984 1988 1992 1996 2000 2004 Note This excludes the former Soviet Union and Eastern Europe Source World Bank staff estimates countries undertake reforms and liberalize-will contribute to the rising share of developing countries in world market capitalization. Given these trends and the fact that the share of emerging markets in industrial country portfolios remains very small, there is still consid- erable room for an expansion of investments by institutional investors in emerging markets.23 As a result, the financial integration of develcplng countries is ex- pected to broaden and deepen over the coming decade. As part of this process, gross private flows are likely to rise significantly, with capital flowing not only from industrial to developing countries, but, increas- ingly, among developing countries themselves, and from developing to industrial countries. In the short to medium term, developing coun Lries, in aggregate, are likely to receive net private capital flows because investment risks are declining, expected rates of return are improving, and these countries are underweighted in the portfolios of institutional investors. However, the rate of growth, and eventually the levels, will inevitably decline. Moreover, there is likely to be considerable variation among countries, 112 THE NEW INTERNATIONAL ENVV~R depending on the pace and depth of improvements in macroeconomic performance and creditworthiness. Aging in Industrial Countries: A New Force An important new factor will provide further impetus to these under- lying trends-the demographic shift under way in industrial countries. Industrial countries now have a pronounced bulge in their demo- graphic structure, reflecting the aging of the baby boom generation and declining birthrates (figure 2.13). This will lead to a steady rise in the proportion of elderly to active population in all industrial countries, al- though the pace of this increase will be most pronounced in Japan (fig- The fact that developing countries ure 2.14). As figure 2.13 also shows, this is in sharp contrast to the have younger populations than indus- trial countries will have a positive situation in developing countries, whose clearly pyramidal structure re- effect on the flow of capital to flects a much younger population.24 emerging markets. Figure 2.13 Demographic Structures of Developing and Industrial Countries, 1995 Age groups: developing countries Age groups: industrial countries 75 or mor 70-7 65 65 - :-' : Male 6 Female F:L S044: i AL: A = = A~~~~~~16dA 300 200 ±00 0 100 200 300 40 30 20 10 0 10 20 30 40~ Population (millions) Population_(millions) Note Developing countries comprise low-income, lower-middle-income, and upper-middle-income countrieis Industrial countries are All highi- income countries Source World Bank data 113 CITAL FLOWS TO DEVELOPING COUNTRIES Figure 2.14 Elderly Dependency Ratios, Selected Industrial Countries, 1990-2030 (ratio ofpopulation aged 65 and over to populan-on aged 20-64) Percent 60 -+-- Europe 50 -v--- Japan . - US.Xv 40 30 20- The portion of the population formed by the elderly is growing in industrial countries. 0 1990 2000 2010 2020 2030 Source World Bank (1994) and World Bank data There are three broad implications of this difference in demographic patterns. First, the aging of populations in industrial countries could lead to an increase in savings in the short to medium term. Second, aging and the associated slowing of labor force growth is likely to exert downward pressure on the rate of return to capi Lai relative to that of labor in industrial countries. Given the demographic structure in de- veloping countries, the reverse can be expected there. Thus, differences in demographics are likely to reinforce the differe atials in the expected rates of return to capital between industrial and dleveloping countries. Both of these factors should stimulate the flow of capital to emerging markets. Third, the aging of populations in industrial countries is lead- ing to pressures for pension reform. These reforms are likely to result in greater responsiveness on the part of pension funds to investment op- portunities in developing countries. 114 THE NEW INTERNATIONAL ENVit Aging, savings, and rates of return. The aging of populations in indus- trial countries is likely to affect the rates of return to capital in these countries through two channels. First, aging could lead to some increase in savings in industrial countries over the short to medium term. Second, aging can be expected to affect the rate of return to capital through changes in the labor market. Effects through savings. The impact of aging on savings remains con- tentious, but there is reasonable basis to suggest that the demographic patterns in industrial countries are likely to lead to some increase in ag- gregate savings over the next 10 to 15 years, before leading to a decline.25 There are many channels through which aging can potentially affect savings. Since households are the primary savers in an economy, the ef- fects of aging on private or household savings are particularly important. * First, economic theory postulates that the savings rate varies by the age profile of economic agents and households. In particular, the standard model (that of Modigliani) postulates a hump-shaped life-cycle saving profile of households-that is, households save until retirement and then dissave. Empirically, however, the latter dissaving in the elderly is not observed: indeed the rate of savings of the elderly is positive in almost all industrial countries. * Second, in addition to the age effect, there is a "cohort" effect. That is, different age groups differ by birth cohort, and this may cause savings behavior to vary among cohorts, based on the dif- ferences in their earnings histories and investment opportunities. A comparison of the savings rates of different cohorts reveals that later cohorts (younger generations) have recently had a higher saving rate than did earlier cohorts. In particular, the baby boom generation in the United States has had a higher savings rate at every age compared with the cohort born one generation earlier. * Third, since savings is a product of the savings rate and income, savings are also affected by age- and cohort-specific incomes. In general, the age category at which the savings rate tends to peak, between 45 and 64, is also when incomes peak. This means that the age composition effect is reinforced by the age distribution ef- fect. Because incomes as well as savings are highest in the 45-to- 64 category, a shift of the population weight into this category will raise the annual flow of savings more than proportionately to the savings rate. 115 r,APITAL FLOWS TO DEVELOPING COUNTRIES Because of these factors, the projected population structure in in- dustrial countries can be expected to lead to an increase in private sav- ings over the next 10 to 15 years in the OECD as a whole. First, the aging of populations will put more households into the high savings age cat- egory and fewer households into the low savings age category. Focusing on Germany, Japan, and the United States-which together account for two-thirds of OECD wealth-the following trends are expected (Borsch-Supan 1996). In the United States, over the next 10 to 15 years more households will enter the 45-to-64 age group, for which the sav- ings rate is the highest. In Japan, the savings rate is found to be mo- notonically increasing with age, so an increase in the savings rate is also expected there. In Germany, however, the effect of aging on the savings rate is ambiguous: aging results in fewer households below age 37 (when savings rates are low), but it also results in a larger number of households in the saving trough after retirement. Second, the positive age effect on the savings rate is reinforced by the age-income distribu- tion effect, particularly in the United States, wheie a growing number of households will fall into the 45-to-64 age group, the point at which income also peaks.26 Third, as mentioned, there is a strong cohort ef- fect, especially in the United States, where the baby boomers-who have had high savings rates throughout-are entering the peak savings and income age.27 The combination of age and cohort effects is expected to lead to a mildly decreasing aggregate savings rate in Germany and a mildly in- creasing aggregate savings rate in Japan. But there should be a relatively strong increase in the U.S. aggregate savings rate, from 4.7 percent in 1990 to 5.4 percent in 2000, and 5.8 per cent in 2010. Overall, the age composition effect and cohort effect are expected to lead to an increase in the volume of aggregate savings for the OECD as a whole over the next 10 to 15 years: from $1.015 billion in 1990 to $1.4 billion in 2010.28 While these factors could lead to an increase in the level of private savings over the next 10 to 15 years, what happens to aggregate savings in industrial countries depends on the implications of aging on public savings and on the interaction between private and public savings.29 The aging of populations will have significant implications for pub- lic expenditures. In particular, meeting pension fund obligations under the current pay-as-you-go systems would have significant repercussions on the fiscal deficit. Aging is also likely to increase expenditures on health. Empirical evidence from Japan and the United States shows 116 THE NEW INTERNATIONAL El that social expenditures on health increase very sharply with age. And while aging also means a decline in the proportion of children and hence public expenditures on schooling, such offsetting effects are not likely to be large. Since, however, there is a strong commitment to con- taining-and indeed reducing-fiscal deficits in industrial countries, it is increasingly being recognized that pension reforms cannot be delayed for much longer. Such reforms, including a move to a partially funded system, should result in an increase in public savings. Unless private sector behavior completely offsets this increase in public savings, aggre- gate savings in industrial countries should rise somewhat in the short to medium term. Effects through the labor market. Demographic patterns, and the at- tendant slowdown in the growth rate of the labor force, may also affect rates of return to capital in industrial countries. More specifically, be- cause the slower labor force growth implies fewer workers per unit of capital, other things being equal, the returns to capital relative to labor will tend to fall. (Of course, technological innovations that raise the rate of return to capital could mitigate some of the adverse effects of a slowdown in the labor force.) As the reverse is occurring in developing countries, the differences in demographics can be expected to con- tribute to the differential in the relative rates of return to capital be- tween industrial and developing countries. Both the developments in industrial country savings as well as the changes in the labor market should provide further stimulus to invest- ments in emerging markets. Aging and pressures for pension reform in industrial countries. The growing recognition of the burden that pension obligations under the current system will place on fiscal positions in industrial countries and the need to address the problem during the next 10 to 15 years is begin- ning to spur pressures to reform existing pension systems. In particu- lar, there are pressures to move away from PAYG systems to more funded systems; to privatize pension schemes, both by increasing reliance on private employer and individual schemes-the so-called second and third pillars of old age security (World Bank 1994)-and through greater private sector management of public pension schemes; and to deregulate the investment allocations of pension funds to enable them to earn higher returns on their investments. All three fac- tors are likely to result in greater response to investment opportunities in emerging markets (box 2.4). 117 +-TTAL FLOWS TO DEVELOPING COUNTRIES The United Kingdom and the United States have been at the fore- front of the shifts discussed above, but Japan and most of continental Europe are also beginning to establish market-oriented private schemes. Japan began to implement a series of reforms in recent years aimed at improving the incentives for private and individual programs and for greater international diversification.30 Pension reform has also been high on the agenda of the European Union (EU), but it has been a contentious issue. The Pension Fund Directive, which was intended to liberalize cross-border management and investment of pension funds, was withdrawn in 1994. Shortly thereafter, however, the Euro- Box 2.4 Pension Refonms and Their Implications for InVestments in Emering Markets UNDER THE CURRENT PENSION STRUCTURE IN the present value of the benefits schleduled to be paid industrial countries-which is dominated by public, between now and the year 2150 will exceed the pres- defined beneftt schemes that are either partially ent value of expected contributions by two to three funded (apan, Sweden, United States) or financed times the current value of GDP for most OECD coun- on a pay-as you-go (PAYG) basis-the rising depen- tries. If this were to be financed by a payroll tax, it dency wfill become a very significant fiscal burden. would mean a rax increase of about 15 to 20 per- Under a PAYG system, the defined benefits to pen- centage points (World Bank 1994). Alternatively, if sioners are not actuarially tied to contributions and there were a move towards partial fimdinig now, are usutally financed from a payrol tax. If this is the with an immediate one-time increase in the contri- case, the benefits received by current retirees deter- bution rate, the payroll tax increase would be lower, mine the taxes paid by current workers. Speciftcally, at around 9 to 12 percentage points in most coun- the payroll tax rate depends on the benefit rate (av- tries (World Bank I994}. Recognition of this prob- erage benefit/average wage) times the dependency lem is spurring pressures for pension reform. These ratio (beneficiaries/covered workers). Meeting the reforms are likely to result in a greater response on defined benefit obligations in the face of a rising de- the part of pension funds to investment opportuni- pendency ratio will require either further increases ties in ermerging markets for several reasons. in the fiscal deficits or higher tax rates on the active population. (The magnitude of the overhang from a First, a move to more fully funded systems unfunded pension plans is very large, net pension fi- will entail a one-time hike in the current con- | abilities of public pension programs amount to some tribution to amass a temporary surplus to be 60 percent of GDP for major industrial countries.) It used to pay The rising pension benefits. (Note is estimated that if the current system remains un- that wien the interest rate that capital earns is changed and is financed strictly on a pay-as-youg-go hgher than the rate of wage growth-which basis, as opposed to moving toward partial funding, is likely if pension funds invest greater 118 THE NEW INTERNATIONAL EYNV pean Commission issued a communication which seeks to ensure that there are no national restrictions on fund managers and to carefully de- fine investment rules and levels of prudential control so that there are no undue restrictions on these grounds (Harrison 1995). Many Euro- pean governments have started to ease investment restrictions, but this will be a slow process, in part because of strong political opposition in France, Italy, and Spain. Even if the pace of reform is uncertain, how- ever, the direction is clear. The magnitude of the potential increase in private pension funds and its impact on developing countries can be assessed from figure 2.15. amounts in emerging markets-fully funded pension schemes that have accounted for thc pension schemes have a cost advantage over growth of pension fund investments inremerg- PAYG systems. That is, the contribution rate ing markets. will be lower under the former.) Increased a Finaly, pension funds in most countries cur- pension payments are likely to increase total ready face some restrictions on international long-term savings available for investment, investments (box table 2.4). These regulations unless households cut back on other forms of are currently not binding in countries like personal savings or increase borrowing. Canada and Switzerland (and of course the a Second, the privatization of pension schemes Netherlands, the United Kngdom, and the will result in more diversified institutional in- United States, where there are no major restric- vestments, induding investments in interna- tions), since pension fuinds in these countties tional markets. Whereas publicly managed are allocating much smaller proportons to in- pension schemes are usually obliged to invest ternational investments than they are currently in government or government-related securi- allowed by law. However, as the trend toward ties, privately managed pension schemes-in international diversification grows, these re- which workers or employers choose their fund- strictions will become binding. Moreover, in managers-do not face this restriction. They places such as Germany and the Scandinavian have the incentive to allocate capital to invest- countries, current regulations are very tight and ments that offer the best risk-yield combina- already appear binding, since allocations to in- tion, regardless of whether the securities are temational investments are close to the maxi- public or private. As a result, they are able to mum allowed by law. As these regulations are beneftt from international management exper- liberalized, we can expect to see greater intema- tise and international -iversification in invest- tional investments in these countries. ments. Indeed, It is the privately managed (Box continues on tbefollowingpage 119 TAL FLOWS TO DEVELOPING COUNTRIES Box Table 24 Cunnt Restftino en Peasio Fudse Foreg In stneut& Coieny Resitons Belgium Maximum of 65 percent in equities in the oQCD orny. Canada Ceiling on foreign investment raised from 10 to 20 percent during 199U-95. Deemark Maximum of 40 percent in 'high risk assets, a category that inclndes foreign equities. France Maximum proportion of asset& invested in shares (or mutual funds invesied in shares) is 65 percent. Foreign investlent permitted only if legal tide of ownership remains in Fran(e. Currency matching-at least 80 perent of assets must bie invested in the same currency as liabtiiIes. Germany Maximum of percent in non-EU bonds and maximum of 6 percent in non-EU equitieS. Eighty percent of assets must be invsted in the same currenay as liabilities. Italy Investment policies of the new private funds determined by funds' boards of directors and tend to be restrieted to insurance policies,propet government bonis, and bank deposits. Japan Maximum of 30 percent in fureign-currency-entominated- assets (part of 5:3:3:2 rule) Previ- ously this rule applied to pordifos managed by indivtdual asset managers, rather than to the pension fund as a whole (old requirements still apply to tax qualified pension plan assets, or TQPP). Also employe' pension, fund plans considered to have sufficient rrsanagement compe- tence and experience nay be exempted from the 5:3:3:2 rule on overal assets. Further deregula- tion planued. Netherlands No restrictions.--Prudent maen rule applies. Deregulation in January 1996 of ABP (the civil servantse fund and the largest pensionsfund in Europe). Portug Maxium of 40 percent-in foreig equities, listed in the European Union, Nsew York, or Tokyo only. Spain No restrictions.- Sweden Maximum of 5 to 10 percen in foreign ctrrcies. Switzerland Maximum freign eqaities,1g percent- AMaximmWTliseg currency honda 2 percent. Maxmum Swiss fanc bonds (by foreigners), 30 percent Total bonds (excluding Swin borrowers), 30 percent United Kingdom No restrictions. 'Prudent man rule' applies-trustees must invest assets in a prudent manner and in the most atpropriate way for the menibership. United States No restrictions (on private pensian funds). 'Prudent man" rule applies. Note. The 5;3:3~2 rue in Japan is the asset allocation ration that apphes to pension find investments. Under the new rules noted above, the pension fund as a whole must meet the isllownig requirements: 58 percent minimum in principal guaranteed issets 30 percent maximum in domestic stocks 30 percent maximum in foreiga-crreny-desominated bonds 20 percent maximum in real estate. Souree; DeRyck (1996), Harrison (1995¶ WorldBaukstafinterviews. As figure 2.1.5 shows, private pension funds in industrial countries can significantly expand their holdings of international equities. 120 THE NEW INTERNATIONAL ENVIRION Figure 2.15 Pension Fund Assets and International Diversification, Selected Countries Per capita pension assets (dollars) 30,000 25,000 20,000 15,000 10,000 5,000 Switzerland Netherlands United United Denmark Japan Italy Canada Ireland Genmany France Belgium Italy States Kingdom Private sector coverage (percent) 100 80 60 40 20 0 Switzerland FPance Netherlands Denmark United Sates United Kmgdom Germany Ireland Belgium Japan Italy International diversification (percent) 40 30 20 10 0 wine.d Belgium United Netherlands Canada Switzerland United States Denmark Japan Frnce Germany Sweden Kingdom Holdings of international equities (percent) 30 25 20 I5 10 5 0 Ireland United Krgdcm Bgi Netherlands United States Japan Switzerlad Germany Denmark Frnc Sweden Source World Bank staff estimates using data from InterSec, PDFM Penston Fund Indicators (1995). 121 .P AtP1fAL FLOWS TO DEVELOPING COUNTRIES Pension fund assets are extremely uneven across industrial countries and generally quite modest on a per capita basis. NWith aging, per capita and total pension assets will need to rise sharply. This increase can come in two ways: increases in contributions and increases in coverage. The coverage levels of private schemes, meaning the proportion of pri- vate sector employees covered by corporate pension funds, are still quite low in most industrial countries and therefore have significant potential for expansion. In fact, with pension reform, private pension fund assets are likely to increase dramatically, especially in Japan and Europe. The European Federation for RetiremenL Provision (De Ryck 1996) estimates that the combined EU pension fund asset base could expand ninefold over the next 25 years. The growth of pension assets in conjunction wit h the trend toward in- ternational diversification should result in substantial new international investments. An increasing proportion of such new international hold- Box 2.5 Defined Contribution Pension Plans and lntenational Dtvesification IN THE PAST 15 YEARS OR SO, DEFINED CONTRI- and adrministrative costs, since they need not meet bution corporate pension plans have grown an actuarial standard ensuring that a defined bene- rapidly, especially in the United States. There the fit will be provided in. the future; they are more share of defined contribution plans in total private flexible, since contributions can be linked to the pensioni assets is estimated to have risen from 16 current performance of the employee; and the em- percent: in 1980 to 23 percent in 195. Indeed, in ployee bears the subsequent finaneal risk of the 1995, 42 milion people in the United States be- pension assets. longed to defined contribution plans with assets of There is a question as to whether such pension $1.3 trillion, as compared with 17.5 million people plans-where employees generally play an active and- assets of $162 billion in 1980. Within the de- role in selecting the investment medium for their fined contribution plans, the 401 (kIs-crporate pension accounts-are likely to be more conserva- pension funds in which employees contribute a tive in their investment decisions, reducing the percen'tage of their salary each year, allowing them likelihood of significant furthet international in- to redtuce their taxable income-in particular have vestments by pension funds. In particular, it is gen- seen tremendous growth. Defined contribution erally presumed that these plans will tend to hold plans offer the employer several advantages over less risky assets such as domestic fixed income other plans: they are subject to fewer regulatory instruments. 122 THE NEW INTERNATIONAL ENV1R.. ings is likely to be in the form of equities, including emerging market eq- uities. International diversification by pension funds, and holdings of in- ternational equities in particular, is currently impeded by a number of factors. These include binding restrictions on holdings of foreign assets or international equities (as in France, Germany, Japan, and Spain); cul- tural biases against equity investments (as in other parts of Europe); the more conservative approach of the new pension plans (box 2.5); and re- strictions on the use of international fund managers and advisers, who tend to push aggressively for international diversification (as in much of continental Europe and Japan). All of these factors are in the process of change, albeit gradually. Pension funds, therefore, are likely to be a major force in further international diversification and, in particular, in the de- mand for portfolio equities from developing countries. Today such pen- sion funds hold about $70 billion of emerging market assets. This could rise very considerably over the next decade. Evidence suggests that although such plans are the specialized knowledge to undertake mnore more conservative in their investment allocations, risky investments. Thus, for example, in 1989, with very low levels of international diversification, around 32 percent of the 401(k) assets in- this situation is changing quite rapidly: vested in mutual funds were in guaranteed funds, and a further 1I percent in balanced a First, a growing proportion of defined contri- funds, with only 9 percent in diversified equity bution programs such as the 401 (k) plans, for funds. By 19g5, the percertage in guaranteed example, is being channeled through mutual funds had declined to 22 percent, while that in fimnds-which, as noted, are in a better posi- diversified equities had risen to 21 percent. Al- tion, relative to the individual, to invest in- though equity holdings by 401(k) funds are ternationalLy. Thus, in 1986, less than 8 still lower than those of traditional corporate percent of 401 (k) assets were invested in mu- pension funds, they are not all that low. It is tual funds. By 1994, this share had risen to estimated that 401(k) plans now have around 30 percent. 35 percent of their assets in equities, compared a Second, a gradual broadening in the type of in- with the average proportion oftraditional,de- vestments by these plans is under way, with fined benefit pension plans, which hold the recognition that fund managers may have around 60 percent of their assets in equities. I 123 CAPITAL FLOWS TO DEVELOPING COUNTRIES Volatility Arising from the Internaticinal Environment P RIVATE CAPITAL FLOWS AND THE PROCESS OF FINANCIAL integration hold significant potential benefits (as discussed in chapter 3). Yet the volatility of these flows can have serious repercussions on the domestic economy. These concerns have been heightened in the current international environment in which new types of investors-investing in portfolio flows, vThich are a new form of investment for emerging markets-can and do respond more quickly to changing conditions. This section deals with factors in the international economy that can contribute to the volatility of flows and asset prices in emerging markets. Two key questions arise with respect to volatility emanating from the international environment. First, what are the main factors that af- fect the volatility of flows and asset prices in emerging markets? Sec- ond, how significant are these factors in terms of [heir impact? Three factors in the international environment have been identified as having an important bearing on the volatility of flows and asset prices in emerging markets. The first is movement in international in- terest rates and stock market returns in industrial countries. The sec- ond is potential foreign investor herding, meaning that investors follow each other in investment decisions, irrespective of whether the particu- lar investment decision is warranted by changes in economic funda- mentals. Such behavior leads to excess volatility of flows and asset prices-that is, volatility unrelated to, or in excess of, changes in eco- nomic fundamentals. The third factor is contagion, which occurs when events in one emerging market change investors' behavior in other emerging markets, regardless of whether the economic fundamentals of the latter have been affected or not. The Role of International Interest Rates Private capital flows to emerging markets are seer as being particularly prone to movements in international interest rates (and other asset re- turns). In part this is because changes in international interest rates can have sizable effects on the macroeconomic performance (through their impact on trade) and creditworthiness (through their impact on the 124 THE NEW INTERNATIONAL g debt burden) of developing countries. But developing countries are also seen as being more susceptible to changes in international interest rates because foreign investors still think of emerging markets as marginal investments.31 The presumption is that because these investors con- sider investments in emerging markets as a means of adding higher re- turns to their portfolios only when their mainstream investments-in industrial countries-are underperforming, investments in emerging markets will be very sensitive to changes in industrial countries' inter- est rates. The evidence suggests, however, that this presumption is true only for portfolio flows and not for FDI, which accounts for more than 50 percent of all flows to developing countries. In fact, based on the estimated relationship between private capital flows to developing countries, U.S. interest rates, and countries' macroeconomic funda- mentals that is reported in box 2.2, we find that movements in U.S. in- terest rates account for only around 1 percent of the observed variation in total private flows-as opposed to 32 percent accounted for by changes in countries' macroeconomic fundamentals.32 In part, the relatively low sensitivity of aggregate flows to changes in international interest rates is due to the fact that FDI flows-especially those that occur as part of the globalization of production-are driven by firms' considerations of long-term profitability. Such flows are there- fore much more responsive to changes in the investment environment of the host country than to temporary fluctuations in interest rates. Cyclical changes in interest rates may affect the start of an FDI project. It may also change the form of the financing package, but temporary movements in interest rates are unlikely to affect the magnitude of the flows associated with the project. Portfolio flows, on the other hand, are quite sensitive to changes in interest rates. Because a portfolio investor buys bonds or shares in a company to get a rate of return, portfolio flows are highly sensitive to differentials in rates of return among countries. Moreover, should the rates of return rise elsewhere, portfolio investors, unlike FDI investors, can divest themselves of their stocks of equities or bonds relatively eas- ily at the market price for those securities. For FDI investments, on the other hand, the value of the project is not public information, so an in- vestor who wants to sell may not get the fair price because of problems of asymmetric information. Hence FDI investments are much more costly to reverse than are portfolio investments.33 125 ½ KCAUTAL FLOWS TO DEVELOPING COUNTRIES Portfolio flows are also more volatile than FDI because portfolio in- vestors consider emerging markets to be marginal investments. When international interest rates rose in the first quarter of 1994 (for the first time since 1990), portfolio flows to emerging markets declined sizably. Despite the further increases in interest rates during the year, however, portfolio flows to emerging markets recovered by the third quarter. Portfolio flows did decline sharply during the first quarter of 1995 when there was another spike in interest rates, but this period also co- incided with the Mexican crisis. In fact, the recovery in portfolio flows in the second quarter of 1995 was sharper than the recovery following the interest rate increase in the first quarter of 15994, even though in- terest rates were around 2 percentage points higher in 1995. Interna- tional interest rates rose again in early 1996, with seemingly little effect on portfolio flows to emerging markets. The effect of global interest rates on portfolio flows appears therefore to have declined somewhat (figure 2.16).34 This is most likely because institutional investors are becoming more familiar with emerging markets and increasingly con- sider them to be mainstream investments. Despite some decline in sensitivity, however, movements in interna- tional interest rates-if they are sizable-will undoubtedly continue to affect the volatility of portfolio flows to emerging rnarkets. At the coun- try level, if portfolio flows account for a sizable proportion of total flows, movements in global interest rates could translate into sizable volatility in flows and in domestic macroeconomic variables. Investor herding. A second factor that has recently received a great deal of attention as a cause of volatility in portFolio flows and asset prices is herding on the part of foreign investors. Investor herding is generally attributed to problems of asymmetric information. It has been suggested that the current structure of thie investor base-in which the assets of primary investors (retail investors and pension funds) are managed externally by professional fund managers-is par- ticularly susceptible to herding behavior (box 2.6). The essence of this argument is that fund managers will follow the investment decisions of other fund managers in order to show clients that they know what they are doing. If they follow other fund managers' decisions and the investment turns out to be unprofitable, they are more likely to be thought of as unlucky than as unskilled, since other fund managers will have made the same mistake. Given that such a high percentage of household savings is now channeled through institutional investors 126 THE NEW INTERNATIONAL EI Figure 2.16 International Interest Rates and Portfolio Flows to Emerging Markets, 1993-96 Billions of dollars Percent 30 7 Interest rate increase, 6 25 February 1994 5 20 December 1X994- 4 i5 3 10 2 5 The effect of global interest rates on portfolio flows has declined 0 0 somewhat. 1993:4 94:1 94:2 94:3 94:4 95:1 95:2 95:3 95:4 96:1 96:2 1 IN Total portfolio flows U.S. dollar LIBOR I Note LIBOR is the London interbank offered rate for six-month deposits. "Portfolio flows" refers to international issues of bonds and equities Source Euromoney Bondware and World Bank staff estimates that employ professional fund managers (it is estimated that 14 per- cent of household assets and 60 percent of pension fund assets are managed externally), the potential for such behavior clearly exists. Much also depends on how the mandates of fund managers are set: some mandates could reinforce these incentives to herd. Mandates of external fund managers of pension funds in the United Kingdom, for example, often stipulate that the fund manager perform at least as well as the median fund. Consequently, underperformance relative to the median is penalized, while overperformance is not rewarded propor- tionately. This could arguably provide more of an incentive to herd than if fund managers are evaluated against an index as well as against peers, as is often the case in the United States, or asked to achieve a minimum yield, as in Japan (Griffith-Jones 1996b). 127 tPITAL FLOWS TO DEVELOPING COUNTRIES Box 2.6 Is the Current Investor Base Prone to Herding? HERDING ON THE PART OF INVESTORS IS GENER- Clearly, to the extent that household and pension ally explained by a variety of asymmetric informa- fund assets are managed externally by professional tion problems. For example, it has been argued that fund managers (that is, to the extent that a principal- asymmetric information between find managers agent setup exists)-and to the extent that the per- and the primary investors (retail investors or pension formance of a fund manager is evaluated relative to fund rnanagers) can give rise to "principal-agent' that of other find managers-incentives to herd out problenis and result in herding behavior by futnd of reputational concerns can eximt, It is estimated managers (see Scharfstein and Stein 1990). The rea- that 14 percent of the assets of individuals and about soning is as follows. Fund managers can be either 60 percent of pension fund assets in the major highty skiled or of low abilities. Highly skilled man- industrial countries are managed t'hrough fund man- agers will tend to receive informative signals about agers, including bank managers, insurance man- the value of an investment, while those of low abil- agers, and mutual fund managers (InterSec). In ity will receive purely random signals. Since the abil- general, though, the performance of a fund manager ity of fund managers (the agents) cannot be tends to be evaluated both in relation to the perfor- determined with certainty, they need to signal the mance Of other fund managers and in relation to an quality of their abilities to the primary investor (the appropriate index. (Thus, if the fund manager spe- principal). And as the investment decisions of highly cializes in emerging markets, his performance may skilled managers will tend to be correlated (because be evaluated relative to the performance of the IFC they are all observing the same piece of 'truth), Emerging Market Index, for example.) The fact that whereas the decisions of low-ability managers will managers are also evaluated relatire to an index may not (their signals will be random), it pays for an in- somewhatdampen the incentives :o herd out of con- dividual fund manager to make the same investment cerns for reputation (although, of course, if foreign decision as other fund managers in order to signal investors account for a significant proportion of a that he is a manager of high skill. Even if the invest- country's market capitalization, their herding ac- ment ttIrns out to be unprofitable, an unprofitable tions could influence the index as well). In addition, decision is not as bad for the reputation of a fund since fund managers' salaries depend on the returns manager when others make the same mistake-in a they achieve, they will clearly also attach weight to world in which there are systematically unpre- the profitability of their investment decisions. Their dictable components of investment value, they can salaries are usually a percentage of the assets under all share in the blame. their management, which, in turn, is a function of Thus, even if a fund manager's own information their performance over the preceding two years or suggests that an investment has negative expected so. Third, there may be a 'superstar" effect (see value, he may choose to pursue it if other fund man- Rosen 1981), in which the top-ranked fund man- agers have done so. Conversely, he may choose not agers earn disproportionately higher wages. For to undertake an investment even if he believes it to managers ranked in or near the top 100, the incen- have a positive expected value if other fund man- tives to herd will also be much lower. agers have chosen not to make the investment. 128 THE NEW INTERNATIONAL ENVIRl, Herding behavior may also be stronger in emerging markets because the investors are less familiar with these markets and will thus be reluc- tant to rely solely on their own assessment of the fundamental value of an investment. Under these circumstances, investors will at least partly adjust their investment decisions according to those of other investors, behavior that could lead to herding or trend chasing. A factor that will have some dampening effect on such behavior, however, is the high transaction costs associated with equity investments, which constitute the bulk of institutional investors' investments in emerging markets. Because of high transaction costs, turnover ratios of equity investments are generally less than one-third that of bonds. It has been argued that pension funds are less prone to short-term trend chasing and volatile behavior than are mutual funds (Griffith-Jones 1996b). In particular, the argument has been that retail investors have shorter investment horizons and are hence more prone to responding to short-term trends than are pension funds, and that since retail investors are the primary investors in mutual funds, mutual funds are more likely to respond to short-term trends. An important factor in this regard is the frequency with which fund managers' performance is reviewed. Pension funds tend to review their external managers' performance and policy al- locations every two or three years, although monitoring generally takes place annually (box 2.3). The performance of mutual fund managers who serve retail investors as well, on the other hand, does tend to be eval- uated by the market over a much shorter horizon.35 However, over 50 percent of pension fund investments are undertaken through the pur- chase of shares in mutual funds, so the distinction between mutual funds and pension funds is, in practice, more blurred. While the theory of investor herding is plausible, it is difficult to demonstrate empirically. Studies have tested for investor herding and positive feedback trading (buying stocks whose prices are rising and selling those whose prices are falling) in the United States (Lakonishok, Shleifer, and Vishny 1992). In particular, they have looked at the extent of correlation across fund managers' buying and selling of individual stocks. There is little evidence from these studies to support the notion that there is investor herding in the United States.36 For emerging markets, it is difficult to test for investor herding by looking directly at the buying and selling patterns of fund managers for individual stocks-given the paucity of data. An indirect way of assess- ing whether foreign investors herd in emerging markets, though, is to 129 DAflTAL FLOWS TO DEVELOPING COUNTRIES look at the behavior of stock market prices and returns. If investors herd, changes in stock market prices and returns in one period will tend to be accentuated in the next period, so prices will tend to exhibit peri- ods of upward or downward swings (that is, prices will tend to be some- what predictable), which eventually reverse, leading to excess volatility. Since there are other reasons why stock prices and returns may exhibit such patterns, however, price swings could suggest investor herding only if they are accentuated when foreign investors become important in an emerging market.37 Moreover, such herding could be on the part of foreign and domestic investors-who could be overreacting to the decision and actions of foreign investors, thereby contributing to excess volatility. Figure 2.17 illustrates the behavior of asset returns in a sample of emerging markets as institutional investors have become important. A ratio greater than one indicates positive autocorielation in returns- that is, it shows that returns exhibit successive periods of increases (or decreases). As figure 2.17 shows, the degree to which these markets were prone to such periods of upward or downward swings increased in the period in which foreign investors first became important in these markets-which could suggest foreign investor herding or do- mestic investors' overreaction to the actions of foreign investors. What is noteworthy, however, is that all the markets in the sample (with the exception of Mexico) saw a decline in excess volatility as the duration of foreign investor presence increased.38 This linding is consistent with the hypothesis, discussed below, that the poi ential for foreign in- vestor herding at the country level diminishes as these investors be- come more familiar with markets and as these markets become more financially integrated. As discussed in chapter 6, foreign investor pres- ence in emerging markets can also promote, among other things, bet- ter information disclosure, which in turn can reduce problems of asymmetric information and the potential for investor herding- whether by foreign or domestic investors. Again, this is consistent with the finding that excess volatility declines as the length of foreign investor presence increases. The possibility that emerging markets may be subject to investor herding was also analyzed by testing whether Brady bond prices have spillover effects on local stock market returns over and above what is justified by changes in a country's economic fundamentals. If so, this would be consistent with less informed investors (foreign or domestic) 130 THE NEW INTERNATIONAL ENVI Figure 2.17 Excess Volatility in Emerging Markets, Selected Countries and Years (measured by the variance ratio test) 2 *pre-FP 1.5 * LI post-FP2 I~~~~~~~~~~~~~~~~~~ 0 5 Countries are more susceptible to investor herding in the early stages of integration, but such herding tends to decline the longer foreign _ investors are active in a market. Brazil Chile India Korea Malaysia Thailand Mexico Composite Statistically significant at the 10 percent level Note FP denotes foreign presence FP1 indicates the first period of increased foreign pres- ence, which differs across countries This threshold date was taken to be the period in which total assets of open- and closed-end dedicated emerging market funds reached $500 million or more. FP2 is the following period, up to 1996 For example, mutual fund assets in Thailand passed $500 million in 1988. Thus the period prior to foreign participation (pre-FP) is taken to be from early 1986 to the end of 1987, the initial period of foreign participation (FPI) is from early 1988 to the end of 1990, and the later pernod of foreign participation (FP2) runs from early 1991 to the end of 1996 Cross-country comparisons are not meaningful because other factors can affect the variance ratio (see note 37) Source World Bank staff estimates using the IFC Investable Return Index. overreacting to Brady bond price movements in making their invest- ment decisions in local stock markets-in the belief that those invest- ing in Brady bonds (primarily foreign traders, commercial and investment banks, and hedge funds) have more information about changes in the economic conditions of the country and in the value of the investments than they do. As box 2.7 shows, there is in fact some evidence to suggest that such spillover effects exist. 131 -tAPITAL FLOWS TO DEVELOPING COUNTRIES Box 2.7 Do Investr Ovenreact to Changes in Brady Bond Pfices? ONE POSSIBLE FACTOR THAT HAS BEEN WIDELY may indeed overreact, or herd, in response to move- cited as a cause for market ineffielency and excess ments in the price sf Brady bonds. volatility in emerging markets is herding by in- vestors, particularly international investors, whose Coefficient Measuring Effect ofBrady Bond access to information may be limited. Prices on Excess Volatility When such information does become available, Counry Coefficient investor s may overreact, so stock price movements in Argentina 0.59* emerging markets are greater than would be justified Brazil 0.37* by the subsequent impact of the news on dividends. Chile 0.30 One such example is the discount on Brady bonds. It Jordan 0.03* has been suggested that investors in emerging mar- Mexico 0.59* kets ofien watch changes in Brady bond prices to Nigeria 0.47* infer information about changes in countries' eco- Philippines 0.47* nomic fimndamentals and that they tend to overreact Venezuela 0.20* to these changes in prices. Significant at the 5 percent level. Thi.s proposition was tested by examining 1. Using the methodology develop.-d by Campbell and whetheir there was excess volatility in stock markets Shiller (1987), we examnined the relationship between the and whether this excess volatility was correlated with actual dividend/price ratio p. and the theoretical divi- movements in Brady bond discounts (prices).' Effi- dend/price ratio p equal to the present discounted value cient nmarket hypothesis suggests that stock prices of the growth of dividends Dd According to this method- should be equal to the discounted present value of fu- ology, all information relevant to forecasting Dd is included in pt Therefore divergences between the actual ture dividends. The extent to which the volatility of price/dividend ratio and the theorer ical price/dividend actual stock prices exceeds the volatility of the theo- ratio, which is correlated with movements in Brady bond retical stock prices is the measure of excess volatility. prices, is evidence of overreaction in equity prices to In seven out of the eight emerging markets we ex- movements in Brady bond prices. A significant positive amined, stock market prices were found to exhibit coefficient indicates that stock markeE prices p rise above the theoretical pricesp*when there is good news in Brady excess volatility that was correlated with movements bond prices. in Brady bond prices. This suggests that investors Source. World Bank staff estimates. The key question, though, is whether such herding behavior results in significant volatility in asset prices and returns in emerging markets. Studies that have compared the behavior of asset returns in emerging markets with those in industrial countries have Found that emerging markets do exhibit greater volatility. These studies have found evidence of return reversals at long horizons-that is, periods of overperfor- mance are followed by periods of underperformance-evidence that, again, could be consistent with investor herding. ]-lowever, these rever- 132 THE NEW INTERNATIONAL ENVR 4 sals are not much different from those in industrial countries (Richards 1996). Furthermore, data on the short-term volatility of returns do not suggest that there has been a generalized increase in the overall volatil- ity of returns following liberalization of emerging markets (Richards 1996, Bekaert and Harvey 1995). Several factors also suggest that emerging markets should become less susceptible to volatility related to foreign investor participation as these markets become more integrated: * The conditions under which emerging market investments are undertaken are changing rapidly. In particular, institutional in- vestors are becoming more familiar with emerging markets-as evidenced by the significant increase in both the volume and quality of broker research into emerging markets over the past four to five years. The potential for herding should diminish as better macroeconomic, industry, and company information be- comes available to foreign investors-a result of financial inte- gration and capital market development. * The form of investor participation changes and broadens as countries become more integrated (box 2.8). In the initial, or "pre-emerging," stage of integration, the likelihood of herding is low because investors are specialized and dedicated. Countries in the newly emerging stage are more vulnerable to herding because investments are in a less dedicated form and investors are invest- ing in an environment of possible asymmetric and incomplete information. Finally, as markets become more mature and inte- grated, the investor base tends to become more heterogeneous and better informed, dampening the potential for excess volatil- ity at the country level. * As markets become more integrated, another factor comes into play: inefficiencies at the country level will have been exploited, so that excess returns are made not at the country level but at the level of individual stocks. Accordingly, fund managers' invest- ment strategies will tend to move away from top-down alloca- tion and active management at the country level to bottom-up allocation and active management at the stock level. That is, their strategies will combine more passive, indexed-based invest- ment at the country level with the active selection of stocks. The potential for herding at the country level should therefore di- 133 0T A___R~TAL FLOWS TO DEVELOPING COUNTRIES Box 2.8 The Form in Which Foreign Investors Participate in Emerging Markets Wll Vary with the Extent of Financial Integration IN THE VERY EARLY STAGES OF FINANCIAL INTE- on an active discretionary allocation strategy at the gration, foreign participation tends to be in the form country level. With less dedicated flows and the pos- of boutique investment funds. These funds are usu- sible problem of asymmetric and incomplete infor- ally not very large, and their investment objectives are mation among fund managers (particularly since generally quite narrowly defined; they are looking for there are less specialized managers participating), selective opportunities in specialized markets. The in- there is a greater potential of herding behavior. vestments of these funds tend to be based on detailed Finally, as markets become more mature, more country, sector, and company knowledge. At this integrated, and better known to investors, there ap- stage, therefore, the likelihood of herding is low. pear to be two opposing forces at work. First, the in- As countries reach the newly emerging stage, in- vestor base broadens further, with global fund vestor participation broadens. For most emerging managers also investing in these countries. These countries, the process is one of first having country fund managers are, other things being equal, less funds--which are relatively dedicated money-and likely to have detailed knowledge at the country and then entering regional funds and global emerging stock levels than are specialized emerging market market funds (box figure 2.8). In the latter cases, in- managers. On the other hand, as mentioned above, vestments are made across a region or across several at this stage the level and quality of information emergintg market regions, and the funds are corre- available to investors in general is much greater at spondingly less dedicated at the country level. At the country, sector, and stock levels. This dampens this stage, moreover, investments tend to be based the tendency toward excess volatility. minish, even if it remains at the level of individual stocks. In- vestors at this stage of integration are also likely to be more tacti- cal in the timing of their investments (selecting national markets according to current levels of key benchmark. asset returns relative to their long-run equilibrium value). They act, from the country's perspective, in a countercyclical manner. In general, therefore, the potential for volatility arising from foreign investor participation can be expected to follow an inverted U-shaped curve, as countries move from the pre-emerging stage of financial inte- gration to the mature emerging stage. It is likely to be highest when countries are in the newly emerging stage of financial integration. The factors underlying this process are summarized in table 2.1 below. What is more, the experience of countries that have begun the process of integration earlier suggests that the peak of hercLing behavior and the 134 THE NEW INTERNATIONAL ENVIRP! Box Figure 2.8 Changes in the Form of Investor Participation as Countries Become More Integrated Number of l(orea country funds 96 Philippines Thailand x~ India Indonesia 2 Malaysia Pakistan X 22 Mexico 'i1 Argentina X6 8T8ZIl,'~~~~~~~~~~~~~~~ ~~~55 Brazhiel' ~ Peru) __ _ _ _ _ 3 * ~~~~~~~Turkey~ _ _ _ _ _ _ _ _ _ _ 14ungary~ X 2 Poland *-X 2 Morocco K 3 Asian regional32 Latin American teglonal * 15 East European ieglonal * 38 0 0 ~~~~~~~~~~~~~~~~~~~~~~~29 African reigional and'snl coity 1986 1986 1987 1988 1989 1990 199± 1992 1993 j1994 1995 1998a ~~~ ~~International x - Country * * Regional funds funds funds a. Second quarter. Saurce. Micropal; World Bank staff estimates 135 <2APITAL PLOWS TO DEVELOPING COUNTRIES Table 2.1 Volatility in Various Stages of Financial Integration Investor Market St cture and Conditions Pre-emerging N-ely emerging Mature emerging Factors that Countries begin to have regional In addition to managers of coun- can increa,e and emerging market funds that try, regional, and emerging mar- volatility are less dedicated at the country ket funds (and managers with level. Also, most investors in these speciali:ed mandates), there these emerging markets, while will also be global fund managers, perhaps using an index-as a who, unlike the specialized man- benchmark, are likely to have a agers, may be operating with rela- discretionary allocation strategy, tively little information at the This will typically involve looking country level. at macroeconomic conditions in countries (a top-down approach), but also include considerations of market liquidity and what other investors are doing. This strategy- may result in herding behavior. Factors that Boutique investors under- In the-early stages of the newly As markets become better known can dampen rake careful stock anialysis emerging markets, there are dedi- to investors and morecinforma- volatility for a select group of coun- cared (country} funds. tiOto is available, the speciahst or tries, which have such low - knowledgeable managers are in a market capitalization that -Although in later stages there will better positonr to undertake tacti- they are not in any inter- be less dedicated mtoney and cal asset allocations (which national index. higher risks of herding, in prac- involve selecting national markets tice active managers undertaking according to current levels of key discretionary allocations typically benchmark asset returns relative tend to take only small bets off to their long-run equilibrium). the index. Therefore, even if This is likely to have some coun- there is a risk of herding, the tercyclicality and dampen absolute magnitudes of flows are volatility. not likely to be large. As markets become large enough to enter the global index and inefficiencies at the country level have been exploited, investors are more likely to use the index to allocate over countries (that is, to undertake passive investing) and use stock picking as a rneans of adding to the value within coun- try allocations. This is likely to reduce the risk of herding at the country level 136 THE NEW INTERNATIONAL ENVIR RN associated increase in excess volatility appear to last for a relatively short period, as discussed above (figure 2.17). Cross-Country Contagion The Mexican crisis highlighted the issue of contagion-spillover effects from one country to another-which can be another source of volatil- ity for emerging markets. Conceptually, two types of contagion can be distinguished. The first is fundamentals contagion, in which a shock in one country can affect investments in other countries because the countries share similar fundamentals or are exposed to common exter- nal shocks. Or it is possible that the shocks in one country are trans- mitted through trade or financial channels and thereby affect the economic fundamentals of other countries.39 The second type of contagion occurs when shocks in one country af- fect investments in other countries, even if the economic fundamentals of the latter have not changed. This could be termed "pure" contagion. From the country's perspective, such changes in investments would constitute excess volatility. Investors' initial reactions in the wake of the Mexican crisis, which erupted in the last quarter of 1994, were an ex- ample of pure contagion. Because investors were not sufficiently dis- criminating among emerging markets, portfolio flows to almost all emerging markets declined very sharply in the first quarter of 1995. In- deed, international equity and bond issues by emerging markets de- clined by around 86 percent in the first quarter of 1995 over the previous quarter. Moreover, all regions were hit significantly-al- though Latin America was the most affected. In fact, the only countries that issued international bonds of any significant magnitude in the first quarter of 1995 were Korea and Portugal. By the second quarter of 1995, however, many emerging markets had returned to the international financial markets, and there was a much greater differentiation in the terms of borrowing. This initial contagion was also reflected in equity prices in emerging markets. Al- most all emerging markets saw declines in domestic equity prices dur- ing the first quarter of 1995. By the end of May, however, countries whose macroeconomic fundamentals were relatively strong saw a re- covery in prices. In fact, there is evidence to suggest that the magnitude and duration of the decline in stock market prices was clearly related to some key macroeconomic fundamentals-average inflation rate and 137 2CALTAL FLOWS TO DEVELOPING COUNTRIES growth of output and exports-before the onset of the crisis (figure Although stock prices in emerging 2.18).40 In sum, there is evidence of contagion alising from investors markets declined after the Mexican behavior. However, evidence also suggests that International markets crisis, the reaction was relativelyh mild and transitory for countries with have discriminated among countries relatively quickly after the initial good macroeconomic fundamentals. reaction. Figure 2.18 Impact of the Mexican Crisis on Stock Prices in Emerging Markets, 1994-95 Perc5ent einotutgot . ecn ±5 , [ Median inflation 8.5 percent N li zd i Median export growth 9.3 percent Median output grrowt . percent l l E i g gI ~~~~~~~~~~~~~Median inflation 16.0 percent 10 l l i E llS_ glw BC ll ,4| ~~~~~~~Median export grrowth 4.6 percent -15-- E -20 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0 4 1 Impact index Export growth Note The imrpact is measured both in terms of the magnitude and the duration of the decline in stock market prices from December 16, 1994, to the end of M/ay 1995 As the impact is measured as an index, it is a negative number, as shown on the botrom half of the graph The top half shows the average export growth in percent per year in each country, as well as the median output and export growth and rates of inflation of the two groups of countries-the group in which the Mexican crisis had a relatively small impact (an impact index of -1 ()0to - 5.0) and the group in which the impact was relatively large (an impact index of - 0 to - 20.0) As the graph shows, the group of countries in which the impact was relatively small had stronger economic fundamentals Source World Bank staff estimates 138 THE NEW INTERNATIONAL ENV Even if investors are becoming more discriminating, however, two factors, related to the investment strategy (rather than behavior) of in- stitutional investors and to the structure of the current investor base, have the potential to contribute to contagion. * Institutional investors appear to invest in emerging markets using a two-step allocation process in which they first allocate a pro- portion of their portfolio to emerging markets as an asset class, and then allocate among individual emerging markets within this. Asset returns in one emerging market are therefore evaluated on the basis of returns to both the world portfolio and a broad emerging market portfolio. Under these circumstances, shocks that lead to any changes in asset returns in one emerging market will lead to changes in investment allocations to other emerging markets (Buckberg 1996). However, the magnitude of contagion arising from this process is likely to be relatively small, since it in- volves portfolio rebalancing within the overall allocation to emerging markets as an asset class, which itself is still quite small.41 * A high likelihood of contagion also derives from the fact that a sizable proportion of mutual funds that invest in emerging mar- kets are open-end funds. Unlike closed-end funds, open-end funds are required to redeem claims on demand. Since managers are particularly averse to large losses, redemptions-or the fear of redemptions-will tend to affect the portfolio decisions of open- end mutual funds. (In fact redemption risk is an important rea- son why open-end mutual funds tend to have much higher turnover ratios, averaging around 100 percent, compared with those of closed-end funds, at 50 percent or less.) Such funds may be forced, by capital losses in one country, to sell holdings in other countries, either in order to keep their country shares in the correct proportions or to meet investor redemptions.42 Evidence shows, for example, that following the Mexican crisis, U.S. mu- tual funds sold shares in all emerging regions-including Asia. In the longer run, though, fund managers were able to substitute be- tween regions, reallocating a bigger share of their portfolios to the assets of the regions where macroeconomic fundamentals were perceived to be strong. Thus, contagion arising from the purely institutional nature of mutual funds appears to be short-lived.43 139 ME E CAPITAL FLOWS TO DEVELOPING COUNTRIES In sum, a shock in one emerging market may well lead to some volatility in flows and asset prices in other emerging markets. However, experience suggests that such contagion will be short-lived-lasting only until investors have reevaluated the prospects in individual emerg- ing markets. Conclusion P RIVATE CAPITAL FLOWS TO DEVELOPING COUNTRIES HAVE continued to grow despite the increase in IJ.S. interest rates in 1994 and the Mexican crisis in 1994-95-albeit at a slower rate. Aggregate private flows have shown resilience because a significant pro- portion has been in the form of FDI , which is less susceptible to cycli- cal and other short-term shocks than other types of flows. While port- folio flows initially declined by substantial amcunts in response to these shocks, their recovery has also been impressive. Despite the large drop in the first half of 1995, portfolio flows recorded only a small decline for the year as a whole: it is estimated that in 1996 they rose to levels comparable to the peak reached in 1993. This chapter has argued that the sustained increase in private capital flows reflects the fact that these flows have now reached a new phase, one driven by increased financial integration, although low interest rates in industrial countries provided the initial impetus. The two fun- damental forces driving the growing investor interest in and integration of developing countries are: * higher long-term expected rates of return, as a result of policy re- forms and improved creditworthiness * the opportunities that developing countries provide for risk di- versification because of low correlation between returns in devel- oping and industrial countries. The chapter has also argued that the magnitude and speed of re- sponse of private capital flows to investment opportunities is much higher than it was in the late 1970s and early 1980s, because of signif- icant changes that have taken place in the enabling environment of both industrial and developing countries. As a result of these changes, emerging markets lhave seen growing in- vestment by institutional investors. Whereas in 1986 there were only 140 THE NEW INTERNATIONAL E 19 emerging market country funds and 9 regional and global funds, in 1995 there were 505 country funds and 773 regional and global emerg- ing market funds. And although pension funds have started to invest in emerging markets only recently, in the United States the average share of emerging markets in pension fund portfolios is comparable to that in mutual funds' portfolios. In the case of some of the larger U.S. pen- sion funds, the share of emerging markets in total assets is considerably higher. The factors that are driving financial integration of developing countries are still unfolding: * At the international level, the factors that are increasing the re- sponsiveness of private capital to cross-border investment oppor- tunities-deregulation, competition, financial innovations, technological advances, and reductions in communications costs-are still far from complete, and are likely to continue to re- inforce one another. _ This process will receive added impetus from the aging of popu- lations in industrial countries. The most important element with regard to the demographics in industrial countries is the pressure for pension reform and the deregulation of pension fund invest- ments it is likely to exert. * In developing countries, market accessibility can be expected to increase, and policy reforms are likely to deepen-in countries that are already implementing such reforms-and broaden, as more countries embark on the process of integration. As a result, the financial integration of developing countries is ex- pected to deepen and broaden over the coming decade, against a back- drop of increasing financial integration at the global level. Indeed, given the changes that are taking place at the international level-in particu- lar, the rapid advances in technology, communications, and financial in- novations-and the growing economic sophistication in developing countries, the progressive financial integration of developing countries appears to be inevitable. Gross private capital flows may therefore be ex- pected to rise substantially, with capital flowing not only from industrial to developing countries but, increasingly, among developing countries themselves and from developing to industrial countries. As noted above, net private capital flows to developing countries can expected to be sustained, even though the rate of growth, and eventu- 141 tJTB.CAPITAL FLOWS TO DEVELOPING COUNTRIES ally the levels, will inevitably decline. Moreover, there is likely to be considerable variation among countries, depending on the pace and depth of improvements in macroeconomic performance and credit- worthiness. Indeed, in countries where economic and policy funda- mentals are very weak, the initial manifestation of growing financial integration may take the form of net outflows of private capital. Private flows to developing countries are also likely to be subject to fluctuations from year to year, even in aggregaie. International in- vestors tend to be very responsive to changes in the international envi- ronment, including interest rates, and to events in other countries, since they see each country as one of many optiolis. Although the de- gree to which international interest rate movements affect portfolio flows to emerging markets is declining, emerging rnarkets are still quite marginal in foreign investors' portfolios. These rmarkets therefore re- main quite susceptible to cyclical conditions in industrial countries. However, since large increases in international interest rates are un- likely-industrial countries are now operating in a low inflation envi- ronment and hence are unlikely to require sLharp corrections in monetary stance-the volatility in aggregate flows to developing coun- tries arising from international interest rate movements is not likely to be very large. Foreign investors are still relatively unfamiliar with emerging mar- kets, so there remains a potential for volatility arising from investor herding. However, there is a learning process involved, and the peak of the volatility arising from investor herding in a market does not appear to be sustained for very long. Foreign investors are also beginning to differentiate among emerging markets. This fact suggests that while a shock in an individual emerging market is still likely to result in some volatility in other emerging markets as investors reevaluate the prospects in the latter, such pure contagion is likely to be relatively short-lived. The main risks of volatility and large reversals lie at the individual country level. As argued above and elaborated in subsequent chapters, while some of the major recipients have relatively well established pol- icy and performance track records, most developing countries lack strong macroeconomic, banking sector, and institutional underpin- nings, and hence remain vulnerable to potential Instability and rever- sals of flows. While international investors are becoming more discerning, market discipline tends to be much more stringent when 142 THE NEW INTERNATIONAL ENVII investor confidence is lost-triggering large outflows-than during the buildup to a potential problem. Financial integration can therefore magnify shocks or the costs of policy mistakes, leading to greater instability. 143 :CAPITAL FLOWS TO DEVELOPING COUNTRIES Annex 2.1 Table 2.2 K(ey Deregulations, Financial Innovations, and Technological Advances Yearl Dm estiefinancial Externalfinancial Teehnolog- country liberalization liberalization Financial innovations ical advances 1970-80 Public ownership of mem- Warrants. NASDAQ United bers of the stock market. Options introduced on computer States Deregulation of securities CBOE. quotation. firms' commissions (May Money market funds DOT Day). certificate. computer Beginning of phasing out Treasury bond futures quotation. of deposit price controls. on CBOT. ITS. International Banking Act Futures on IMm. for national treatment of U.S. and foreign banks. United Relaxation of foreign exchange Options market. TALISMAN Kingdom controls. centralized dearing (1979-93). Japan Securities firms offer Overseas stock listing. Negotiable certificates, medium-term govern- New Foreign Exchange and For- New bond funds. ment bond finds. eign Trade Control Law. Foreign exchange banks established. First Samurai bond issued and listed. Interbank foreign exchange trad- ing begins in Tokyo. Foreign stock listing on TSE. Ban on issuance of Japanese cor- porate bonds overseas lifted. First issue of Euro yen bonds by a nonresident. First issue in Japan of unsecured yen bonds by a foreign private company. Gensaki bonds offered to nonresidents. Domestic DC market begins, open to nonresidents. Law on foreign securities firms 144 THE NEW INTERNATIONAL ENV1R-R6 Yearl Domesticfinancial Externalfinancial Technolog- country liberalizatian liberalization Financial innovations ical advances 1981-90 Banks allowed to affiliate 30 percent withholding tax on wySE composite index AMX and United with securities firms for interest income paid to future. Toronto States underwn.ting. foreigners repealed. NYSE composite index line. Security Pacific is the first NYSE, AMEX, NASD allow foreign option. ITS and bank to set up a limited issuers if they comply with NYSE composite index NASDAQ securities firm subsidiary. home country laws. future option. link. Federal Reserve approves Primary Dealer Act requires reci- S&P index option. Coordinated limited securities procity before foreign financial Treasury bond futures circuit underwriting institutions can become dealers on NYFE. breakers CFTC approves GLOBEX. in the U.S. securities market. Currency options installed. Federal Reserve approves Rule 144a exempts from registra- introduced. cBaoT begins limited debt-equity tion privately placed debt and NYSE trading of instru- evening underwriting by com- equity offered to qualified insti- ments linked to trading. mercial banks. tutional buyers. foreign equities. United Financial Services Act. Foreign membership of stock LIFFE futures market. ITS, SEAQ Kingdom Deregulation of markets. Treasury bond futures. Inter- commissions. Negotiable bankers' nationaL acceptances. SEAQ auto- Commercial paper. mated Gilt warrants. quotation. Japan Japanese bank allowed to Commercial paper of foreign Warrants. cORES, lend long-term Euro yen issue. Treasury bond futures. automated to borrower of their Foreign mernbership of stock Money marker funds quotation, choice market. certificates. second New Bank Law and Secu- Samurai bond regulations Negotiable bankers' section. rities Exchange Law. relaxed. acceptances. Centralized Securities firms allowed to Foreign exchange trading no Stock futures 50. depository sell foreign exchange longer tied to commercial trade, TOPIX stock index and CDs and CPs in the hedging and swaps allowed, future. dearing. domestic market. Yen-foreign exchatnge conversion Nikkei stock index JASDAQ Banks allowed to deal in limits for foreign banks abol- future. introduced. government bonds. ishied. Introduction of gov- TIFFE opens. Interest rate deregulation Withholding tax on Euro yen ernment bond future. begins. bonds issued by Japanese resi- Euro yen floating rate Taxes on bond transac- dents removed. notes, zero-coupon tions reduced. Medium- and long-term Euro bonds, DCs, Commissions for large yen loans liberaized, warrants. transactions lowered. First Shogun bond issued. Foreign banks open trust First Euro yen straight bond subsidiaries. issued (Table conrinues on tbe followingpage.) 145 CITAL FLOWS TO DEVELOPING COUNTRIES Table 2.2 (continued) Yearl Domestic financial Externalfinancial Technolog- country liberalization liberalization . Financial innovations ical advances Japan Japan Offshore Banking Market Domestic and Eure (cont.) opened. yen CP markets Japanese firms make markets on introduced. SEAQI. Stock index futures Restrictions on Japanese purchases traded on Osaka of foreign securities removed. exchange Insurance company and pension fund trust accounts allowed to increase foreign exchange assets. Japanese financial institutions allowed to trade in overseas fiLtures markets. FourJapanese securities firms become primary dealers in U S. government securities markets. Restrictions on domestic and Euro yen CP issues by nonresi- dents relaxed. All financial institutions allowed to trade as brokers in overseas financial futures. 1991-96 Restrictions on U.S. banks After-hours United are eased to allow them trading on States to comnpete with securi- NASDAQ ties houses. Interna- tional. United Alternative Kingdom Investment Market for small firms (AIM); Trade- point. Japan Commercial banks allowed Ministry of Finance authorizes to establish securities sub- trading of non-Japanese finan- sidiaries. -cial futures through GLOBEX. Securities houses allowed Domestic bond marktr further to establish trust bank dereguated, easing the criteria subsidiaries. for issuance on the Samurai Securities houses allowed market. Issuers no longer have to offer money narket to have investment-grade rating. funds. 146 THE NEW INTERNATIONAL RNV-tEi Year! Domestic financial Externalfinancial Technolog- country liberalization liberalization Financial innovations ical advances Japan Deregulation of Valuation miethod for foreign (cont.) commissions. bonds held by institutional Securities and Exchange investors is changed. Previously Council proposes that had to be valued at cost; now banks and other financial can be valued at market value institutions be allowed to or cost-the same treatment own securities that is accorded to Japanese subsidiaries. government bonds. Diet eliminates separation Law requiring companies to mark of commercial and to market their foreign currency investment banking. assets whenever the yen moved Deregulation of time by over 1 5 percent is deposits; other savings eliminated. deregulation to follow. Regulations on foreign exchange Across-the-board deregula- positions of foreign exchange tion of financial markets banks is eased, to promote planned by 2001: easing investments in foreign- restrictions on portfolio currency-denominated bonds. allocations of pension funds, removal of restric- tions on foreign exchange transactions, and further liberalization of domestic financial markets, including liber- alization of fixed com- missions of securities transactions. Note:AMEX, Amernican Stock Exchange: CBOE, Chicago Board Options Exchange; cBoT Chicago Board of Trade; cFTc, Commodity Futures Trading Commission, coirsS, Computer-Assisted Order, Routing, and Execution System; DC, deep discount bonds; DOT, Desig- nated Order Turnaround System; 5Mm, Chicago International Money Market; iTs, Intermarket Trading System; JASDAQ, Japanese Associa- tion of Securities Dealers Automated Quotation; LIFFE, London International Financial Futures Exchanges; NASDAQ, National Association of Securiues Dealers Automated Quotation; NYFE, New York Futures Exchange; NYSF, New York Stock Exchange; SFAQ, Stock Exchange Automated Quotation; sIPC, Securities Investors Protection Corporation; TIFFE, Tokyo Financial Futures Exchanges; TOPIX, Tokyo Price index; TSE, Tokyo Stock Exchange. Sources Sobei (1994); McKinsey Global Institute (1994); Goldstein and Mussa (1 993). 147 CAPITAL FLOWS TO DEVELOPING COUNTRIES Notes 1. Developing countries are also more susceptible than the resources available for external payments, it is directly industrial countries to real sector shocks resulting from, sensitive to the investor's discount rate-that is, to inter- for example, terms-of-trade changes or a slowdown in in- national interest rates (Fernandez-Arlas 1994). A decline dustrial country activity. Their greater susceptibility anses in international interest rates increases the present value of from the fact that their economies tend to be less diversi- resources available for external payments and therefore fied and because their stock of outstanding debt is gener- country creditworthiness. Second, international interest ally higher. However, the susceptibility of developing rates can affect a country's debt burden if debt is held at countries to such shocks is declining because these variable interest rates. Hence a decline in international in- economies are becoming more diversified terest rates can reduce a country's external debt burden. 2. The analysis suggests that the decline in global in- 7. Such risks include not only sovereign risk, which is terest rates it 1990 provided a strong impetus to private still quite high in developing countnes, given their rela- flows (especially portfolio flows) to developing countries; tively short policy track record, but also higher investment allocation was determined largely by economic funda- risks, such as legal and custodial risks, settlement and op- mentals. This is corroborated by the second piece of erational risks, information and regulatory risks, and non- analysis which shows that there was large co-movement of market risks. U.S. portfolio flows to developing countries and that this co-movement was correlated with U.S. interest rates. In 8. According to the Capital Asset Pricing Model other words, U.S. interest rates appear to have had an im- (CAPM), the optimal portfolio allocation in a fully efficient portant bearing on portfolio capital movements to devel- and integrated capital market is the world market portfo- oping countries during 1990-93. lio. Therefore, an investor with a portfolio that is under- invested in foreign assets (as a percentage of world market 3. Since savings tend to be a function of per capita in- capitalization) can, by increasing his holdings of foreign come, savings are lower in developing countries (under fi- assets, either lower risk without sacrificing expected re- nancial autarky) than in industrial countries, and turn or increase the expected rate of return for a given therefore so is the stock of capital. level of risk The benefits of such diversification arise from undertaking investments that are relatively uncorrelated 4. A high default risk increases the likelihood of non- with each other. Because domestic securities have a com- compliance with contractual payments-either explicitly mon exposure to country-specific shocks, domestic secu- or implicitly through, for example, discriminatory taxes rties are likely to exhibit stronger correlation with each against the foreign investor. Declining creditworthiness other, so that international diversification may reduce therefore increases the variability and reduces the ex- risks more quickly than domestic diversification (see pected rates of return to the foreign investor (Fernandez- Reisen forthcoming). Arias 1994). [n contrast, domestic investors are concerned only about returns to the individual project. 9. That private capital flows in the 1990s are being dri- ven by portfolio diversification mo yives as well as expecta- 5. Macroeconomic stability has been found to be a tions of higher returns is reflected in the fact that very important determinant of investment. Macroeco- developing countries have now become recipients of port- nomic volatility is estimated to have reduced developing folio equity flows-the form of Iv estment for which risk country growth between 1960 and 1990 by as much as diversification is most important--and the fact that gross 0.9 percentag,e points per year (Schmidt-Hebbel 1995). flows are sizably larger than net flows. 6. In principle, international interest rates can affect 10. The growth and development of stock markets is, country crecitworthiness through two main channels. in general, very dependent on the availability and disclo- First, since creditworthiness reflects the present value of sure of high-quality information. This may explain why 148 THE NEW INTERNATIONAL ENVIkO stock markets tend to become more important in the fi- 16. Expropriations by governments, which were sub- nancial structure of economies as these economies de- stantial in the 1960s and 1970s, have virtually ceased velop: companies become larger and countries develop since the mid- 19 80s. the requisite accounting and disclosure standards. Mishkin (1996) for example, discusses how securities 17. These include the General Agreement on Trade in markets are more prone to problems of asymmetric infor- Services (GATS), which sets standards for market entry and mation and adverse selection than are commercial for the uniform treatment of firms-whether domestic or banks-so the development of the securities markets is foreign; the Agreement on Trade Related Investment likely to evolve after that of the banking sector. Measures (TRIMS), which imposes equal treatment of all Demirgu,-Kunt and Levine (1 996b) provide evidence on firms (excluding services); and the Agreement on Trade how financial structures differ across countries and Related Aspects of Intellectual Property Rights (TRIPS), change as countries develop. which is beginning to address the protection of intellec- tual property rights-a concern that is particularly impor- 11. As argued in chapter 6, foreign investors them- tant for many high technology transnational companies selves have made a significant contribution to this devel- (World Bank 1997). opment-and the process has been self-reinforcing. As these investors have entered the markets, they have con- 18 An open trade regime is one that fulfills the follow- tributed to their deepening, which in turn has increased ing criteria: nontariffbarriers covering less than 40 percent their attractiveness as markets. of trade, average tariff rates of less than 40 percent, a black market exchange rate that is depreciated by less than 20 12. The MSCI-EAFE is the Morgan Stanley Capital In- percent relative to the official rate, the lack of a state mo- ternational Index, including Europe, Australia, and the nopoly on exports, and a nonsocialist economic system. Far East. The IFCI is the IFC Emerging Markets Index, which takes into account restrictions on foreign investors. 19. The main holders of emerging market debt (Brady bonds, Eurobonds, global bonds, and Yankee bonds) are 13. In fact, portfolio diversification benefits can be commercial and investment banks, hedge funds, and a reaped as long as the correlation among returns is less than few of the larger insurance companies and pension funds. one. 20. Other than those in the United Kingdom and the 14. Even based on the current rates of return and cor- United States, most emerging market funds or interna- relations among these returns, the above example would tional funds with emerging market exposure are, for tax be an upper limit, since any large shift in the portfolio mLx reasons, domiciled in offshore centers. of institutional investors would likely raise prices-in the short to medium term-and could also result in an in- 21. One survey was undertaken for this study; the other crease of their volatility, given the relative thinness of de- is based on emerging market allocations of the Frank Rus- veloping country stock markets. sell universe. This figure is much higher than had been previously estimated-that is, IMF (1995) put the figure at 15. Canada, France, Germany, Japan, the Nether- 0.5 percent and Chuhan (1994) at 0.2 percent. lands, the United Kingdom, and the United States. There is some double counting entailed, since pension funds 22. Even the 401(k) plans (the most popular type of typically make a significant proportion of their invest- defined contribution corporate pension plan) have seen a ments through intermediaries, including other institu- growing trend of investing through mutual funds, rising tional investors. For instance, in Europe and the United from around 8 percent in 1986 to 30 percent in 1994. States, pension funds account for about half of mutual fund assets. In Japan, pension funds make most of their 23. The existence of "home bias"-or a preference for investments through insurance companies and trust home assets-may mean that investors will continue to banks. hold smaller shares of emerging market assets than that 149 APITAL FLOWS TO DEVELOPING COUNTRIES implied by the latter's share of world market capitaliza- possibility of investing in regions with higher rates of re- tion Many reasons have been put forward for the exis- turn to capital, such as emerging markets-feedback ef- tence of home bias, such as the fact that optimal portfolio fects are less likely. calculations do not take into account factors like the im- portance of nontraded goods, the role of human capital, 30. The two main changes aie liberalization of the transaction costs, and the ability to achieve quasi-interna- rules that apply to Japanese corporate pension funds on tional diversification by holding shares in domestic multi- the use of foreign fund managers, and relaxation of the re- nationals. These factors may explain part of home bias, striction on investment, including foreign investment, on but by no means all of it. To the extent that home bias ex- all fund managers. ists because investors are not sufficiently familiar with emerging m arkets, it can be expected to diminish over 3 1. If interest rates affected the volatility of flows to time. At a minimum, the share of emerging market assets emerging markets purely through their effects on devel- in investors' holdings of international assets should rise. oping country creditworthiness, much of differences in the volatility of flows among emerging markets would be 24. The demographic shift is more advanced, however, explained by the differential impact of international inter- in China andl some Latin American countries. est rates on countries' debt burcdens. However, a very small proportion of the differencz in volatility of flows 25. Accoi ding to the findings of some studies, such as among emerging markets appears to explained by differ- Gokhale, Kotlikoff, and Sebelhaus (1996), aggregate sav- ences in the (floating interest rate) debt burden. ings in industrial countries have already peaked and can be expected so show continued declines. Also, it should be 32. A vanance decompositior shows that domestic noted that aging is only one factor to affect savings rates macroeconomic factors as a block account for 32 percent and the aggregate level of savings. of the variation in total private flows (including FDI) in the short run and 66 percent in the long run, while U.S. 26. The age effect on the savings rate is slightly nega- interest rates account for 1 percent in the short run and 2 tive in Germany, because of the early retirement age. percent in the long run. 27. In 1990, the vanguard ofthe U.S. baby boom gen- 33. The distinction between FDI and portfolio flows eration was 44 years old, while the youngest baby should not however, be exaggerated. It has been argued, boomers were 27. In the year 2010, therefore, the baby for example, that an FDI investor who wants to get out of boomers will be between ages 45 and 64. a country can borrow on the domestic market against his investment and then take his money out. And sustained 28. These effects were channeled through age-and co- shocks, such as deterioration of a country's longer-term hort-specific household rates, savings rates and disposable economic fundamentals, can lead to a reduction in rein- household incomes They also allow for population vested earnings and a faster repatriation of earnings as a growth. means of divesting. 29. Aging could also affect the demand for private in- 34. As noted in box table 2.2 , rhe strong negative cor- vestment, and hence net pnvate savings. However, there relation between (the first principal component of) pri- is little empi rical basis on which to assess the direction and vate capital flows and (the first principal component of) magnitude of these effects. One channel is through the global interest rates, which was clearly evident during labor marker and the effects on the relative rate of return 1990-93, is not seen in the more recent period. to capital. As discussed below, the slowdown in labor force growth is likely to place downward pressure on rates 35. There is evidence to suggest that mutual fund of return to capital in industrial countries. In a closed managers alter the riskiness of their portfolios at the end economy, this could have feedback effects, reducing the of the year, presumably because thelr performance on an savings rate. However, in an open economy-with the annual basis is critical (Chevalier and Ellison 1995). 150 THE NEW INTERNATIONAL ENYR 36. There is slight evidence of investor herding in 1994, to the period of the maximum fall in prices and small cap stocks only. the duration of the decline up to the end of May 1995. A regression of the impact index on export growth and a 37. The variance ratio test is used to assess the presence of dummy (which took the value of one if either inflation investor herding (see Aitken 1996). A variance ratLo greater was below 10 percent or if output growth was over 6.5 than one indicates positive autocorrelation in asset returns. percent when inflation was above 10 percent, and zero This means that price increases (or decreases) in one period otherwise) yielded the following result: are followed by price increases (or price decreases) in the fol- lowing period. (A variance ratio of less than one indicates ind= 14.7 - 4.96*DUM- 0.61*EXPGR negative autocorrelation-pnce increases followed by price (-1.79Y) (-2.03) decreases, or price decreases followed by price increases.) It Adjusted R' = 0.48. should be noted, however, that deviations from a variance ratio of one do not necessarily indicate stock market ineffi- * Statistically significant at the 10 percent level. ciency. Other reasons indude market microstructure-related factors such as infrequent trading or time-varying risk premi- 41. Buckberg (1996) finds that emerging market re- ums. Evidence of price predictability therefore should not turns are determined by both world market returns necessarily be taken as evidence against stock market effi- (which include industrial countries) and by other emerg- ciency. Significant increases in price predictabllity that coin- ing market returns. In fact, evidence suggests that emerg- cide with increased foreign investor participation, though, ing markets are more sensitive to returns in other may suggest foreign investor herding (or domestic investors emerging markets than to returns on the world portfolio overreacting to foreign investors' actions). for 12 of the 13 countries. Note that contagion arising from portfolio rebalancing implies a positive shock to 38. Note that the Mexican crisis is included in the other emerging markets when one emerging market is hit sample period. negatively. 39. Goldfajn and Valdes (1996) discusses how shocks 42. An open-end mutual fund, unlike a closed-end can be propagated to other countries through the channel fund, must sell its underlying shares to meet investor re- of financial intermediation. demptions. 40. The impact index was constructed by multiplying 43. See Frankel and Schmukler (1996). the percentage of decline in prices from December 16, 151 TER 3 The Benefits of Financial Integration UCH OF THIS REPORT IS DEVOTED TO UNDER- M A standing and avoiding the pitfalls on the road to financial integration. In this chapter, we focus on the goal itself, the long-term benefits obtain- able from a well-managed integration, concen- trating deliberately on the best-case scenano. Fundamentally, the benefits of integration fall into two main cate- gories. On the production side, integration permits greater international specialization and facilitates the allocation of scarce resources to their most productive uses independent of location, thereby accelerating growth. On the consumption side, integration allows individuals (both in the newly integrating and, to a lesser extent, in the already integrated economies) to insure themselves against adverse developments in their home economy both through international portfolio diversification and by tapping global capital markets to smooth temporary declines in in- come. We look at both benefits in turn. While, from a purist stand- point, the gain to financial integration is quite narrowly defined, we also touch upon many of the accompanying benefits of integration (such as the development of good accounting standards or improvements in set- tlement systems), which could, strictly speaking, be attained without in- tegration but in practice are part of the overall benefits package. Integration and Growth T WO ENGINES DRIVE ECONOMIC GROWTH CHANGES IN THE quantity of inputs-in particular, labor and capital, broadly defined-and changes in their quality. 153 C TAL FLOWS TO DEVELOPING COUNTRIES Investment In a financially closed economy, investment must be funded by domestic savings. The two are matched by movements in the real interest rate- the cost of borrowing and the return to savings. WVithout international capital flows, there can be no presumption that the returns to investment are equal across countries, resulting in resource misallocation: on the margin, highly profitable investment projects in some countries may not be undertaken for lack of financing, although lower-return projects are funded elsewhere. Financial integration, whether on the national or the international level, severs the link between local savings and local invest- ment, allowing savings to "run as surely and instantly where it is most wanted, and where there is most to be made of it, as water runs to find its level" (Bagehot 1924, p. 12). Investors are able to borrow and savers are able to lend internationally, thus the global rather than the national in- terest rate becomes the relevant cost of capital and return to savings. Of course, savings and investment must still match, b at they now do so on a global rather than a national scale. Following integration, investment is, under ideal circumstances, reallocated toward the most rewarding pro- jects (on a risk-adjusted basis), regardless of their location, financed by corresponding capital flows seeking the highest risk-adjusted returns. World production increases, a gain partly accruing to savers in the less productive economies, who now receive higher returns, and partLy accru- ing to firms in the more productive economies, who now enjoy a reduced cost of capital. The new pattern of capital flows depends on the relative return to investment across countries. Extensive research over the last decade has established a by-now-familiar set of characteristics of successful economies, including (most important) sound macroeconomic policies and good microeconomic fundamentals. The point is important: it is the relative return rather than the relative development level that drives capital flows; in consequence, as we saw in the previous chapter, finan- cial integration has primarily resulted in capital flows to countries al- ready enjoying high domestic investment levels. I ntegration thus raises the payoff for establishing good fundamentals; it does not substitute for fundamentals. In consequence, the trend towarcL integration may, at least temporarily, widen the gap between the high- and the low-growth developing economies, although in the long terrn this very widening raises the benefits of improving fundamentals in the low-growth group. 154 THE BENEFITS OF FINANCIAL 1 For the growing subgroup of developing countries already offering at- tractive investment opportunities, integration permits an acceleration of investment by augmenting domestic savings with foreign savings. The po- tential gains vary from country to country, depending on the relative prof- itability of investment opportunities and on the difference between the domestic cost and the world cost of capital before integration. If capital in- flows reach 3 to 4 percent of GDP (a typical figure for current large capital importers), the growth rate, based on typical capital shares and capital output ratios, would increase by about half a percentage point (Reisen 1996). This temporary growth effect is by no means negligible yet it is dwarfed by the range of growth rates among developing economies, again illustrating that the successful tapping of foreign capital markets is only one of many ingredients in a successful growth strategy. Can integration permanently raise growth rates in the opening econ- omy, or even in the world economy? Any permanent growth effect must come through an increase in world saving rates or faster produc- tivity growth. There is little reason to believe that integration boosts world saving rates. Indeed, a decline IS more likely, since diversification of income risk and access to world capital markets to smooth out tem- porary income fluctuations reduce the need for precautionary savings. 1 Any permanent gains from integration are thus more likely to come through the quality rather than the quantity of investment. Productivity Integration may boost productivity growth by shifting the investment mix toward projects with higher expected returns, a shift brought about by an improved ability to reduce and to diversify the higher risk typi- cally entailed in higher-return projects. The risk of any particular in- vestment can conceptually be divided into four parts: a global component (for instance, a world recession), a national component (for instance, higher profit taxes or political instability), a sector component (for instance, a technological advance), and a component specific to the particular investment. The sector- and the project-specific risks can be diversified domestically, resulting in sizable benefits from the national financial integration typically preceding international integration. Global risks by their nature cannot be diversified away. This leaves na- tional risk, which cannot be diversified within a country but can be di- versified internationally. 155 ctCAPITAL FLOWS TO DEVELOPING COUNTRIES If national shocks are a quantitatively important part of the overall risk of investment projects, the diversification benefits obtained by adding an- other domestic investment project to a portfolio olf other domestic proj- ects are lower than the benefits of adding the same project to a portfolio of internationally diversified projects. In consequence, a foreign investor can pursue projects with higher project risks and returns while keeping porqfi- lio risk to an acceptable level. A move from a system of closed national economies to an integrated world economy thus permits a global switch toward projects with higher expected returns, thereby increasing the aver- age world growth rate (Obstfeld 1995). The effect is reinforced if integra- tion also permits a reduction in project risk. For instance, if opening is associated with an expansion of the market for domestic production, sus- ceptibility to domestic demand changes declines, enabling a shift to more specialized-and more productive-capital (Saint-]'aul 1992). The quantitative significance of a switch to higher-return invest- ments in the wake of integration depends on three factors: the techno- logical lag of a particular country, risk aversion, and the speed with which existing capital can be reallocated. Simulation studies suggest that the gains are of a very large magnitude. The size of gains reflects the magic of compound interest: even small increases in growth rates have first-order level effects over time. For the economies lagging farthest be- hind current best practice, the gains comprise both catch-up and the global efficiency gain, while the leading economies benefit from only the latter change. Yet even for these economies the benefits of higher output, and hence consumption growth rates, are very significant: a well-managed integration is a win-win situation. The equalization of returns across countries and the higher growth rate resulting from a globally more productive capital stock provide the long-term benefits of integration proper. In addinon, integration pro- vides more immediate benefits through knowledge spillovers, particu- larly in financial markets and via FDI. While these benefits are not strictly dependent on financial integration, integration tends at least to accelerate the knowledge transfer. Financial System Spillovers The cost of capital to the end borrower consists of four components: the real interest rate on minimum-risk liquid assets, r:he risk premium im- posed by the financial intermediary for financing particular ventures, 156 THE BENEFITS OF FINANCIAL INT ._T the cost structure of the financial system, and the profit margin charged. Integration affects all four components. We have discussed the effect of integration on the real interest rate and the risk premium above: inte- gration replaces the local with the world interest rate and reduces the risk premium to the extent that foreign lenders can better diversify the country risk. The other two components-the cost structure of the fi- nancial system and the profit margin-are also affected if integration enhances the depth of the financial system (allowing fixed costs to be spread over a broader base) or its competitiveness. Beyond the effect on the cost of capital, a deepening of the domes- tic financial system in the wake of integration also aids growth through a more efficient allocation of resources, in turn leading to further fi- nancial deepening. This positive feedback loop between financial and real development is well established. International integration is likely to augment the feedback. Foreign financial institutions will introduce or create access to new financial instruments, while participation of for- eigners in domestic securities markets enhances liquidity, in turn facil- itating long-term maturity transformation. The empirical evidence is quite firm: domestic financial deepening is strongly associated with both higher investment and faster produc- tivity growth (figure 3.1 ). 2 Indeed, there is increasing evidence that ex- ternal financial markets are more important in less developed economies than in more developed ones. For example, analyzing data on the 100 largest corporations listed in the stock market in 10 devel- oping countries during the 1980s, Singh (1994) finds that these corpo- rations relied strongly on external sources for financing, and in particular on equity markets.3 Equally unambiguously, financial integration is associated with an increase in financial system depth (and hence, indirectly, growth) both in terms of bank lending and in terms of asset market liquidity and turnover (box 3.1). Figure 3.2 shows that stock market capitalization and turnover in countries that have received the highest levels of port- folio equity inflows have increased more than in countries receiving lower levels of flows. The figure correlates, for a group of selected emerging stock markets, aggregate portfolio equity inflows with the change in market capitalization and trading value during their respec- tive inflow episodes (as defined in chapter 4), all as a ratio to GDP. The data also suggest significant spillovers to domestic activity during these episodes. This can be seen in table 3.1, which compares the ratio of 157 0 -A-311TAL FLOWS TO DEVELOPING COUNTRIES Figure 3.1 Initial Financial Depth and Future Growth Average annual growth of per capita GDP 1960-89 0.07 0.06 0.05 0.044 0.03 11.1. 0.02 ' U 0.01 5 * Ul 0.00 * Domestic finaincial deepening is strongly associated with increased -0.01 -- n-E investment aind faster productiviy growth. -0.02 - 0.0 0.1 0.2 0 3 0.4 0.5 0.6 0.7 Initial ratio of liquid liabilities/GDI' Source Levine (1991) trading value to GDP in the year immediately preceding the inflow episode with the value of trading activity in the last year of the episode, adjusted for the proportion accounted for by foreign investors and nor- malized by GDP.4 Table 3.1 also shows that the number of new listings on the stock exchanges rose during the inflow period in 10 out of the 13 countries. As shown in figure 3.2, the increase in both market capi- talization and trading volume has been associated with an increase in foreign portfolio flows. Most convincingly, figure 3.3 shows that those countries that received the largest portfolio equity flows were also those As lFigure 3.2 shows, market that saw the largest increase in the volume of domestic trading as a ratio capitalization and trading volume to GDP during the inflow period. The table, of course, establishes only increased more in countries that correlation, not causation. Disentangling the causal effect of interna- received the largest portfolio equity corrlain,n caation . Disentanging thecsalef of externa- inflows than in countries that tional integration is quite difficult, since most episodes of external fi- received smaller amounts. nancial liberalization are accompanied by domestic financial 158 THE BENEFITS OF FINANCIAL IN N- N Figure 3.2 Portfolio Equity and Capital Market Development, Selected Countries (percent) A. Change in the market capitalization/GDP ratioa Correlation coefficient = 0.71 250 Malaysia 200 150 Chile 100 Thailand P 50 Colombia Indonesia Mexico E 0 i anI nka Argentina * Korea * Peru -N:Zuela Hungary -50 0 1 2 3 4 5 6 7 8 9 Aggregate portfolio equity inflows/GDP (percent) a Change in market capitalization as a percentage of GDP between the start of the inflow period and 1994 (percent) B. Change in the trading volume/GDP ratio Correlation coefficient = 0.71 180 1 60 Malaysia 140 120 100 80 60 Korea * *Thailand 40 20 Braz Philippines Mexico Turkey Pakistan Indonesia Chile U OCo iArgentina * Peru enezuela Hungary India -20 0 1 2 3 4 5 6 7 8 9 Aggregate portfolio equity inflows/GDP (percent) Note Aggregate portfolio equity inflows/GDP IS equal to the sum of portfolio equity inflows during the inflow episode divided by the 1994 GDP Source IFC, Emergzizg Stock Markets Factbook l9S6, World Bank data 159 I'TAL FLOWS TO DEVELOPING COUNTRIES Box 3.1 The Empirical Link bebveen Integration and Deepening THE LINK BETWEEN FINANCI INTEGRATION companied b additional liberalization steps for and financial deepening can be assessed by regress- domestic financial markets, ing indicators of financial system development on indicators of itegration and a set ofother determi- FRnaniatreelopment ants, including the development stage (proxied by FinaiAl intgrn Bank ns Turnover per capita GDP), the rate ofinflation, and economic openness (proxied by the trade-to-GDP ratio). We g i - - - - - - . - S~~~~Ecess returns 0.033' 0.09** base our analysis on regressions performed both for a composite measure of financial integration * Statisticaygnificantat 10percent. (Montiel 1994) taking into account observed capi- Statistiady significant at 5 pei cent. tal flows, excess consumption volatility, interest parity violations, and the extent to which invest- Since financial deepening is stronigly and at least ment is funded by domestc savings, and for a sec- in part causally associated with growth, the positive ond measure based on the -sie of excess returns effiect offinanial integration on financial deepening (Levine and Zervos 1995a). For both measures, a provides an indirect link between integration and highet value denotes a greater degree of integra- growth. To examine whether there is also a direct tion. T he regression was performed for two rnea- lik, we estimate a standard growth regression (Barro sures oFfinancial sector depth: the ratio of banking and Sala-i-Mantin 1995) for a group of 24 develop- sector loans to the private sector to GDP, and stock ing countries for which the financial integration market turnover as a fraction of GDP. The coeffi- measures were available. Indudirtg only the two fi- dients on the financial integration measures, re- nancial integration measures reveals a positive and, ported below, were positive and, for the excess for the excess returns measure a sigificant, link be- returns measure, significant at the 5 percent level. tween integration and econoniic growt. Adding the The regressions also confirmed the familiar nega- financial development indicators reveals that, look- tive relation between inflation and financial sector ing&backward, the major benefits of integratioa have development, again illustrating the role of good come through enhanced financial development. findamenrtals as prerequisites for a country to ben- Giveni the typical lags involved in both investment efit filly from financial integration. The results are and saving responses, not too much should be read thus quite suggestive of a positive link between in- into the absence of a sigtificant direct link between tegration and financial deepening. It stands to rea- integration and g-rowth; since significant improve- son thit the linkage is at least partly causal, though ments in financial integration have been achieved the causality pattern is difficult to establish empir- orny in the last few years, a full assessment of the ically, since financial integration is frequently ac- longer-terr growth effects cannoe yet be undertaken. liberalization efforts. However, indirect evidence suggests that interna- tional financial integration deserves at least partial credit for domestic financial market deepening. The quantitative benefits of deepening are like y to be augmented by improvements in the quality of financial intermediation. Part of the im- provement will come through prerequisites for successful integration, 160 THE BENEFITS OF FINANCIAL INT:'4 Table 3.1 Change in Domestic Activity in Selected Emerging Equity Markets Domestic trading activity! GDPa (percent) New listings Pre-episode During CounSty Pre-iepz¶sedeb 1994 dphase episode Argentina 0.60 - -27 1 Brazil 3.74 12.79 70 -48 Chile 2.53 - -7 74 India 9.33 7.06 261 460 Indonesia 0.53 1.68 33 159 Korea 29.94 70.83 314 30 Malysia 7.56 89.36 44 240 Mexico 3.34 -3 3 Pakistan 0.49 1.78 78 213 Peru 0.31 2.46 117 -76 Philippines 2.31 10.88 -59 48 Poland 0.04 4.13 9 35 Thailand 9.17 36.36 48 264 - Not available. a. Based on the value of shares traded accounted for by foreign investors in 1995, as detauled in table 6.2, chapter 6 b. Pre-epssode refers to the year preceding the beginning of the inflow episode, as detailed in chapter 4. c. Number of newly listed companies during the period of equal length preceding the inflow episode for each country. Source. IFC, Emerging Stock Markets Fa-tbook 1996. such as improvements in domestic accounting and supervision stan- dards. Another part will come through learning about better practice coupled with the pressures created by greater competition. Most im- portant, however, will be the momentum for additional reforms and development unleashed even by limited first steps toward integration. For example, an initial deepening of the banking system will raise the return to introducing an efficient payment system; the development of domestic equity markets will motivate efforts to improve settlement systems; the potential for issuing securities in foreign markets will en- hance incentives to improve domestic accounting standards; and so forth. Once begun, integration is thus likely to trigger reinforcing dy- namics, creating a virtuous cycle of further institutional development that makes possible even more complete integration. Of course, the dy- namics thus unleashed also carry significant risks and need to be well managed, as discussed in detail in subsequent chapters. 161 TAL4TAL FLOWS TO DEVELOPING COUNTRIES Figure 3.3 Growth in Domestic Trading in Selected E:merging Equity Markets Change in domestic trading/GDP (percent) Correlation coefficient 0.74 100 Malaysia 80 . 60 Korea 40 Thai lnd 20 -~~~~~ - ~ ~ ~--------r--- ---------- - -- -- - - -------------- EBrazi Philippines Poland Developing countries that received 1 P1 kista Indonesia Peru the largest portfolio equit flows saw 0 the largest increase in the value of domestic trading as a ratio to GDP during the inilow peniod. -20 0 1 2 3 4 5 6 7 8 9 Aggregate portfolio equity inflows/GDP (percent) Note Aggregate portfolio equity inflows/GDP is equal to the sum of portfolio equity inflows during the inflow episode divided by the 1994 GDP Source IFC, Emerging Stock Markets Factbook 1996, World Bank c ata The evidence on the ancillary benefits of integration is necessarily anecdotal, but it is suggestive. For example, as shown in table 3.2, av- erage bank profits and average costs in the 1980s were higher for those OECD economies that prohibited entry of foreign banks (closed) than Table 3.2 Bank Operating Ratios, OECD Economies, 1980s Indicator Open Closed Gross earnings margin/volume 3.21 4.48 Pretax profits/volume 0.58 0.78 Operating costs/volume 2.27 3.25 Note. "Open" denotes ewonomies that permitted entry offoreign banks; "closed' denotes economies that prohibited entries. Source. Terell (1986). 162 THE BENEFITS OF FINANCIAL LNT for those permitting entry (open), suggesting that the increased com- petition in integrated economies led to efficiency gains. Table 3.3 shows that even the fairly efficient and integrated Euro- pean Union members stand to reap sizable benefits from adopting best- practice methods, suggesting even larger gains for less efficient and integrated economies. Another indirect effect of integration, the impact of enhanced compe- tition on interest rates, is shown in table 3.4, reporting, for France, Ger- many, and Spain, the spread between the short-term interest rate and the deposit rate before and after a period of significant financial integration. In sum, integration promises significant indirect benefits through improved efficiency of the financial system, including the liquidity ef- fects of enhanced deepening, knowledge spillovers, enhanced competi- tion, and improved accounting and supervision standards. The Table 3.3 Potential Savings from Adopting Best Practice, Selected EU Countries (percentage of current cost) Sec:er France Germany Italy Spain Banking Consumer credits 105 136 0 39 Credit cards 0 60 &9 26 Mortgages 78 57 0 118 Foreign exchange 56 31 23 196 Commercial loans 0 6 9 19 Insurance :lie 33 5 83 37 Home 39 3 81 0 Public liability 117 47 77 24 Securities Institutional equity 0 69 47 153 Institutional gilt 57 0 92 60 Overallpotential reductions Banking 25 33 18 34 Insurance 24 10 51 32 Securities 23 11 33 44 Total 24 25 29 34 None- An entry of zero implies a cost at or below the average of the four lowest-cost providers. Best practice is measured as the average of the four lowest-cost providers. Source: Dermine (1993). 163 , #P PITAL FLOWS TO DEVELOPING COUNTRIES Table 3.4 Spread between Short-Term and Deposit R'ates: France, Germany, and Spain, 1980-89 Prance Germany Spain 1980-85 11.7 6.5 14.5 1986-89 8.8 5.5 9.3 Swtrre Dermine (1993). improved ability to allocate scarce resources in turn raises the quality of investment, increasing growth and hence further stimulating financial sector development, giving rise to a positive feedback loop. The avail- able evidence suggests sizable potential gains even for highly developed economies, raising the prospect of significant gains for less integrated countries. FDI Spillovers International financial integration also opens the door to both inward and outward FDI, which may influence growth through three channels. First, FDI may increase the total volume of investrnent in the recipient economy. Second, even if it substitutes for rather than augments do- mestic investment, it may be more productive than the capital it replaces. Third, it may generate spillover effects that raise the produc- tivity of existing domestic capital. Although it is difficult to determine the extent to which FDI adds to the total volume of investment, the majority of studies conclude that net inward FDI tends to raise total domestic invesiment. Indeed, more tentative evidence suggests that the value of doniestic investment re- placed by FDI is often less than the additional domestic investment trig- gered by FDI (for instance, follow-up investment by domestic suppliers), so that a dollar of FDI raises the sum of domestic and foreign investment by more than a dollar (Borenzstein, de Gregorio, and Lee 1995). The second benefit of FDI-higher productivity of foreign, com- pared with domestic, investment-is fairly noncontroversial.5 Domes- tic firms benefit from better knowledge of local rnarkets; to compete, foreign firms must offer a compensating advantage, notably greater ef- ficiency. The third potential benefit of FDI consists of spillover effects through a variety of channels, including learning by doing, introduc- 164 THE BENEFITS OF FINANCIAL INT~ tion of new techniques in upstream and downstream collaborators, pressures to improve the efficiency of domestic institutions, and so on. Privately financed additions to total infrastructure investments-an emerging trend-will likewise enhance the productivity of existing en- terprises. Significant spillover effects have been established in a number of illuminating case studies, though the precise nature of benefits is specific to the particular circumstances of the recipient and source countries. Strong indirect evidence of the importance of spillover ef- fects is provided by the stylized fact of a sizable increase in exports fol- lowing inward FDI-an increase not solely attributable to reexporting by the new foreign enterprises. FDI thus affects growth in two ways: by contributing to the volume of investment and by helping to improve efficiency. The empirical evi- dence suggests that the primary growth impact comes through im- proved efficiency rather than an increased quantity of investment. In interpreting these results, however, three caveats must be kept in mind. First, the quality effects are self-eliminating: as domestic practices are adjusted, the scope for additional spillover benefits decreases. Second, the evidence strongly suggests that FDI follows rather than initiates growth. The benefits from FDI, much like the financial market benefits of integration, are thus more likely to buttress existing virtuous cycles than to originate new ones; integration magnifies the benefits of good fundamentals, it does not substitute for them. Third, the net effect of FDI on growth will differ across recipients. In highly distorted economies, FDI may primarily exploit domestic rent-extracting oppor- tunities; in economies with functioning markets, relative returns are more likely to channel FDI to bottlenecks, with substantially greater benefits. Summary International financial integration promises substantial growth benefits that can come through a number of channels. First, integration severs the link between domestic savings and investment, enabling increased investment in the more productive economies while raising the returns to savers in less productive ones. Second, integration permits an im- proved global pooling of risks, making possible a shift of the invest- ment mix toward projects with higher expected returns. Third, integration enhances the depth and efficiency of the domestic financial 165 11 B'ITAt FLOWS TO DEVELOPING COUNTRIES system, with important positive feedback effects to investment and growth. And fourth, integration allows for important spillover effects through FDI, enhancing the efficiency of the domestic capital stock. The composition and magnitude of these benefits will differ by coun- try and will depend crucially on the presence of promising fundamen- tals and sound economic policies. Diversification Benefits E PRODUCTION-SIDE BENEFITS OF INTEGRATION DISCUSSED above find a mirror image on the consumption side. Individuals , in the newly integrated economies can si:abilize their income and consumption by holding foreign assets (augmenting labor income depending on the domestic economy with capital income depending on foreign economies) and by using international capital markets to buffer consumption against temporary swings in the domestic economy. Inte- gration thus allows individuals to reduce the dependence of their con- sumption on the fortunes of the domestic econorny, which is likely to be the source of most of their labor income. While residents of already integrated economies benefit from contin- uing integration by gaining another outlet for their savings, and hence additional diversification, the benefits accrue primarily to the residents of the newly integrating developing economies, for three reasons. First, the starting degree of diversification is likely to be very limited. Second, terms of trade and other shocks tend to have a particularly large impact on poorer economies with less diversified production structures and less effective automatic stabilizers. Third, the welfare cost of reducing con- sumption in the face of adverse developments is largest for countries with low per capita incomes, and hence so are the gains from reducing consumption volatility through integration. How large the gains from risk reduction will be depends on how much the volatiliry of consumption can be reduced, and on the value individuals place on this reduction. A sizable literature on the compo- sition of portfolios (partly reviewed in the previous chapter) suggests that consumers in both mature and developing economies do not reap the full benefits possible from diversification. The same conclusion is suggested by a look at consumption volatility. If individuals effectively used capital markets to stabilize consumption, income volatility would 166 THE BENEFITS OF FINANCIAL INTr- significantly exceed consumption volatility. Furthermore, the growth of consumption would be more highly correlated across countries than the growth of income, since consumers in all countries would hold daims on-and hence base their consumption choices on-global rather than national production. Neither prediction is supported by the data: for most low- and middle-income countries, the standard devia- tion of consumption growth exceeds the standard deviation of GDP growth, while the correlation of domestic GDP growth with world GDP growth exceeds the correlation of domestic consumption growth with world consumption growth. Table 3.5 illustrates these stylized facts, reporting, for eight regions, the mean and standard deviation of consumption growth, as well as their correlation structure. For the four developing regions in the col- umn headings-South America, Central America, Africa, and Asia ex- cept East Asia-the standard deviation of consumption growth (the second figure in parentheses) exceeds the average consumption growth rate (the first figure in parentheses), suggesting substantial scope for stabilizing consumption. The ranking is reversed for the four more highly developed regions, yet again volatility of consumption growth is sizable and differs substantially across regions, indicating unexploited diversification benefits, which are also suggested by the low correlations of consumption growth rates across regions. Table 3.5 Cross-correlation of Consumption Growth, Selected Regions, 1960-87 Asia South Central (except America America Africa East Asia) Region (3.11-4.57) (1.68-296) (1.31-3.59) (0.91-3.02) North America (2.35-1.76) -0.24 -0.11 -0.41 0.11 Northern Europe (2.97-1.31) 0.44 0.2g -0.03 -0.29 Southern Europe (3.13-3.03) 0.39 0.11 0.32 -0.16 East Asia (3.64-2.12) 0.13 0.36 0.07 -0.29 Note: The first figure in parentheses is the average consumption growth between 1960 and 1987; the second figure is the standard deviation of consumption growth for the samne period. Source: Obstfeld (1995). 167 '1~ CAPITAL FLOWS TO DEVELOPING COUNTRIES The evidence thus quite strongly suggests that substantial reductions in the volatility of consumption could be obtainecl by enhanced use of international financial markets to smooth out temporary domestic in- come disturbances and to diversify national risks. 'Whether this poten- tial reduction is a significant benefit depends on the value individuals place on reducing the variability of consumption. While the quantifi- cation of these gains remains an active area of research, the preponder- ance of evidence suggests sizable, although not very large, gains from reducing risk. For developed markets, upper bounds on gans are on the order of a 5 percent increase in permanent consumption, with somewhat larger benefits likely for developing countries with higher consumption volatility (Tesar 1995, van Wincoop 1994). To summarize, the gains from integration on thLe consumption side come in two ways: the scope for reduced risk through improved diver- sification and the ability to employ international financial markets to offset temporary income movements, and the ability to shift toward a portfolio with higher expected returns, the mirror image of the shift to- ward investment projects with higher expected retuLrns discussed above. The empirical evidence suggests that while there is substantial scope for risk diversification, the primary gains do not come through the reduc- tion of risk per se, but rather through the ability to raise average port- folio returns (and hence consumption growth rates) as a consequence of diversification. Conclusion W -ELL-MANAGED INTERNATIONAL FINANCIAL INTEGRA- i ton promises substantial benefits. The benefits accrue on both the production and the consumption side. Integra- tion promises to boost growth rates, partly because better diversifica- tion allows a shift to riskier but more productive investments, and partly because of spillovers in the financial sector and via FDI. Integra- tion also promises to reduce the volatility of consumption by allowing a better diversification of portfolios and by permitting international borrowing and lending to offset the effect of temporary swings in national fortunes. Empirical studies suggest that ihe primary benefits will come not through the first round of diversification effects (that is, through the change in the geographic ownership pattern of existing 168 THE BENEFITS OF FINANCIAL INTEG assets), but rather through the more gradual change in new invest- ments toward higher-return projects made possible by the scope for risk reduction and international equalization of returns. It bears emphasizing again that our focus in this chapter has been on the best-case scenario. If the integrating economy suffers from distor- tions outside the financial markets, the welfare effects of integration may differ quite significantly, and at the margin might even lead to im- poverishing inflows. The following chapters are devoted to this chal- lenge of managing the transition to full financial integration. Notes 1. The overall effect of integration on savings is am- we assume that the proportion of trading activity ac- biguous. While precautionary savings will likely decline, a counted for by foreign investors during the last year of higher return may stimulate savings, depending on the fa- the inflow episode was that shown in table 6.2, which ac- miliar tradeoff between income and substitution effects. tually refers to 1995. It is unclear to what extent and in what direction this assumption distorts the results How- 2. The literature is quite voluminous. See Levine and ever, for countries for which time-series data are avail- Zervos (1995a) and De Gregorio and Guidotti (1995) for able, (for example, Thailand), the share of foreign recent work. The 1989 WorldDevelopmentReport(World investors in tradingwas lower in 1994 than in 1995. This Bank 1989) provides an in-depth assessment of the link- suggests that domestic trading was larger than shown in ages between financial sector and overalt development. the table. 3. Demirgu,-Kunt and Maksimovic (1994) and Glen 5. An exception occurs if FDI is motivated by the and Pinto (1994) present additional evidence avoidance of tariffs on imports The importance of "tar- iff jumping" as a major cause of FDI is, however, widely 4. This makes the conservative assumption that all viewed as on the wane (World Bank 1997). trading activity in that year was of domestic origin. Also, 169 -!?TER4 Challenges of Macroeconomic Management W RHILE INCREASED FINANCIAL INTEGRA- tion may provide substantial macroeco- nomic benefits for developing countries, the integration process also carries with it some difficult macroeconomic challenges. v T Policymakers in these countries have been concerned with three types of problems: * The potential for macroeconomic overheating, in the form of an excessive expansion of aggregate demand as a consequence of cap- ital inflows. * The potential vulnerability to large, abrupt reversals of capital flows because of changes in creditor perceptions. * The more general, long-term implications of financial integration for the conduct of macroeconomic policy. As integration advances, policymakers will have to manage the enhanced macroeconomic volatility that may prevail when the economy becomes more ex- posed to external shocks. In addition, policymakers will need to face these and other shocks with reduced policy autonomy. This chapter will examine how macroeconomic management can cope with all these challenges. As explained in chapter 2, two forces are driving the growing in- vestor interest in and integration of developing countries: improve- ments in long-term expected rates of return, following policy reforms and strengthening creditworthiness, and the opportunities offered by developing countries for risk diversification. These two forces are being abetted by favorable changes that have taken place in the enabling en- vironments of both industrial and developing countries-that is, in the 171 CAPITAL FLOWS TO DEVELOPING COUNTRIES economic, regulatory, and technological conditions that influence pro- duction, and govern and underpin the operation of capital markets. As a result of these factors, the initial manifestation of increased integra- tion has been one-way net flows of capital into the newly integrated countries that are quite large relative to the size of the recipient economies. Thus the danger of macroeconomic overheating has been very real. Countries have a broad menu of macroeconomic policies at their disposal to address this problem. This chapter describes the choices that have been made from this menu in a large group of countries, as well as the macroeconomic implications of such choices. A broad find- ing is that countries have generally succeeded in avoiding overheating but have done so by bringing to bear nearly the fuill array of policies at their disposal. Cross-country comparisons reveal both similarities and differences among policy approaches. All the countries examined tended to maintain officially determined exchange rates, and almost all relied heavily on sterilized intervention in the foreign exchange market. Capital controls were used by those countries that received the largest inflows of capital relative to the size of their economies in order to re- duce up front the potential for overheating. While the available evi- dence is not strong and covers only some of these countries, it does indicate that capital controls were effective in the short run in reducing the overall magnitude of capital inflows as well as influencing their composition. Cross-country differences also emerged in the area of nominal exchange rate management (with some countries resisting real appreciation more than others) and the magnitude of fiscal tightening. These differences in approach reflected the relative magnittde of the capital inflow and initial macroeconomic conditions as well as broad development strategy. There are indications that the differences in strategy produce different macroeconomic outcomes, with countries that were able to avoid substantial real appreciation and that tightened fiscal policy tending on average to have lower current account deficits, a mix of absorption more oriented toward investment, and faster eco- nomic growth. A particularly important aspect of macroeconomic management during the inflow period concerns the potential link between postin- flow macroeconomic performance and vulnerability-that is, suscepti- bility to changed perceptions of creditworthiness. The chapter uses the Mexican crisis to examine the relationship between macroeconomic 172 CHALLENGES OF MACROECONOMIC MANAQ performance and vulnerability by assessing how the incidence and severity of the so-called tequila effect emanating from Mexico were linked to macroeconomic performance in other developing countries. Our key findings are that certain initial interpretations of the crisis, which attributed it to the role of short-term capital and the size of the current account deficit in Mexico, are not well supported by the cross- country evidence. Instead, countries that were less vulnerable to the Mexican crisis tended to be those that had avoided real appreciation and imposed a tight fiscal policy during the inflow period, thereby ab- sorbing a larger share of investment and achieving more rapid growth than they would have done under different policies. Thus the policy mix employed in the effort to combat overheating seems to matter from the perspective of vulnerability as well. The two macroeconomic challenges of overheating and vulnerability are likely to be particularly important during the transition to financial integration, when international investors are adjusting their portfolios and when policy credibility is not yet well established in the capital-im- porting countries. As indicated above, however, new macroeconomic challenges will continue to arise even when the process of integration is well advanced. In particular, financially integrated developing countries will find themselves operating in a very different macroeconomic envi- ronment, one in which capital movements are highly sensitive to changes in prospective foreign and domestic rates of return. In that con- text, the macroeconomic consequences of domestic macroeconomic shocks (including changes in policies) will be altered, and the country will also be faced with new types of shocks arising in international fi- nancial markets. This new environment may thus be characterized by enhanced volatility, which poses an ongoing challenge for macroeco- nomic management. In this chapter we consider the role of four types of policy instruments in this new context: fiscal policy, exchange rate pol- icy, monetary policy, and policies regarding the free movement of capi- tal. The picture that emerges from this discussion is consistent with the lessons from experience mentioned above: in a financially integrated en- vironment, both fiscal responsibility and flexibility are very important, and although nominal exchange rate management is not precluded, the scope for allowing the real exchange rate to deviate from its equilibrium value is much reduced. Sterilized intervention, while having the advan- tage of flexibility, is likely to become less effective as integration in- creases. The same is true of restrictions on capital movements. The best 173 tAtITALTAL FLOWS TO DEVELOPING COUNTRIES case for such restrictions on macroeconomic grounds is that they can act as transitional devices to preserve monetary autonomy while other pol- icy reforms are implemented. In the next section we will review how countries have responded to the problem of overheating associated with the recent surge in capital inflows. We will then look at the vulnerability of different countries in the aftermath of the Mexican crisis to assess the role of policies as de- terminants of vulnerability. Next we will examine the long-term impli- cations of increased financial integration for macroeconomic management. The final section will summarize the key implications for macroeconomic policy. Capital Inflows and Overheating: The Country Experience O UR DISCUTSSION OF THE PROBLEM OF OVERHEATING AND the elements of a successful response is divided into four parts. First, we briefly describe the transmission mechanism through which a surge in capital inflows can trigger overheating of the domestic macroeconomy and the range of options available to policy- makers to prevent overheating. We then describe the policies that were adopted by developing countries in response to the most recent surge in capital inflows. With this background, we next examine the macroeconomic performances of these countries and evaluate their success in avoiding overheating. We then sumrnarize the main policy conclusions that can be drawn from this country experience. The Macroeconomic Consequences of Capital Inflcws To examine how individual developing countries have fared in coping with the macroeconomic consequences of capital inflows, we have compiled data on capital flows to a sample of 21 developing and tran- sition economles as shown in table 4.1. These countries together ac- counted for 95 percent of the total FDI and portfolio flows to economies of these types during 1988-95. This table tells a now-famil- iar story. Substantial inflows began in 1988-89 in several East Asian economies that had weathered the international debt crisis of the 1980s relatively well and made significant domestic policy adjustments in 174 CHALLENGES OF MACROECONOMIC MAN Table 4.1 Net Private Capital Inflows to 21 Developing Countries, 1988-95 (percentage of GDP) Cumu- lative flows/ Coeff- GDPat cient of Inflow end of Mean vari- Country episodea 1988 1989 1990 1991 1992 1993 1994 1995 episode ratio ation Argentina 1991-94 13 3.8 2.9 3.1 9.7 2.8 0.38 Brazil 1992-95 2.8 2.3 1.9 4.8 9.4 3.0 0.44 Chile 1989-95 3.3 8.6 3.1 7.4 6.3 7.7 4.0 25.8 5.8 0.39 CoLombia 1992-95 1.8 5.6 5.6 6.2 16.2 4.8 0.42 Costa Rica 1988-95 10.6 12.0 4.4 4.7 9.2 9.1 2.5 5.3 44.0 7.3 0.44 Hungary 1993-95 17.5 8.5 18.4 41.5 14.8 0.37 India 1992-95 1.2 1.7 2.7 1.7 6.4 1.8 0.35 Indonesia 1990-95 2.5 1.9 1.3 0.2 1.1 3.6 8.3 1.8 0.66 Korea 1991-95 2.6 2.5 0.6 2.4 3.5 9.3 2.3 0.45 Malaysia 1989-95 2.9 5.7 11.1 15.3 23.2 1.2 6.6 45.8 9.4 0.82 Mexico 1989-94 2.6 2.2 7.5 7.6 8.5 3.3 27.1 5.3 0.54 Morocco 1990-95 4.6 2.9 2.5 3.0 5.0 3.2 18.3 3.5 0.29 Pakistan 1992-95 2.5 4.9 3.8 3.3 13.0 3.6 0.28 Peru 1990-95 3.9 5.4 5.3 4.6 10.8 8.2 30.4 6.4 0.41 Philippines 1989-95 2.1 3.9 4.4 2.3 4.4 7.9 5.2 23.1 4.3 0.45 Poland 1992-95 4.1 6.8 3.1 12.0 22.3 6.5 0.61 Sri Lanka 1991-95 3.9 5.3 8.2 6.5 3.5 22.6 5.5 0.36 Thailand 1988-95 7.4 10.4 12.3 12.3 8.6 7.7 8.3 12.1 51.5 9.9 0.21 Tunisia 1992-95 4.1 7.1 5.1 4.1 17.6 5.1 0.28 Turkey 1992-93 1.9 4.1 5.7 3.0 0.53 Venezuela 1992-93 3.3 2.0 5.4 2.7 0.34 a. The period during which rhe country experienced a significant surge in net private capital inflows Sou ree. World Ban k data; imF, World Economic Outlook data base; iMw, Internationat Financial Statisics data base. mid-decade. Outside East Asia, large inflows started to appear after the inception of the Brady Plan, with Chile and Mexico leading the way in 1989 and other Latin American countries following.' By 1991, several more countries in both East Asia and Latin America were participating in the episode, as were Morocco and Sri Lanka. One year later, private capital was flowing to all regions represented in this sample, although flows were far from uniform across countries. During this period, an- nual inflows of 4 to 5 percent of GDP were not unusual, and several countries experienced one or more years in which net private inflows exceeded 8 percent of GDP, compared with near-zero or negative values during the debt crisis period. Cumulative flows have varied substan- tially, from extreme values of over 40 percent of 1995 GDP in several 175 C- PJTAL FLOWS TO DEVELOPING COUNTRIES countries in the sample (Hungary, Malaysia, and Thailand), which re- ceived very large flows for an extended period, to a low of about 6 per- cent of GDP in India, Turkey, and Venezuela. Overall, then, financial integration has been associated with levels of cap ital inflows that have been both large and persistent for many of these countries, resulting in cumulative net inflows that in several cases are bound to carry enor- mous challenges for domestic stabilization policy. The transmission mechanism. The key short-run macroeconomic con- cern associated with a surge in capital inflows us that of an excessive expansion of aggregate demand-that is, macroeconomic overheating. This outcome can be produced through the fo lowing transmission mechanism: * If a country maintains an officially determined exchange rate, the commitment to defend the parity causes the central bank to in- tervene in the foreign exchange market to purchase the foreign exchange generated by the capital inflow. To do so, the central bank creates high-powered domestic money. I This expansion of the monetary base creates a corresponding ex- pansion in broader measures of the money supply, lowering do- mestic interest rates and raising domestic asset prices. I This action in turn triggers an expansion of aggregate demand. If the economy possesses excess capacity, the short-run implications may be to increase domestic economic activity and cause the cur- rent account of the balance of payments to deteriorate. Eventually, however (and perhaps rather quickly if dom,zstic excess capacity is limited), excess capacity will be absorbed an(d the expansion in de- mand will trigger an acceleration in domestic inflation. * If the exchange rate peg is maintained, rising domestic prices will cause the real exchange rate to appreciate, abetting the current ac- count deterioration associated with the expansion in aggregate demand. Policies to Control Overheating To avoid potential overheating, developing counirxies can and have in- tervened at every step in this transmission process. Policy can attempt to reduce the required scale of intervention in the foreign exchange market, restrict the monetary expansion associated with a given magnitude of in- 176 CHALLENGES OF MACROECONOMIC MAh42A tervention, and offset through other means the effects on aggregate de- mand of a given magnitude of monetary expansion. These policies are not exclusive, and most countries have brought a wide variety of these instruments into play. An overview of the specific policy choices made by individual countries in a subsample of 13 countries for which de- tailed information is available is provided in table 4.2. The specific forms that some of these policies have taken are discussed below. Reducing net inflows of foreign exchange. Some policies have restricted the required scale of intervention in the foreign exchange market, either through reducing the capital account surplus of the balance of payments or through an offsetting increase in the current account deficit. The main instruments available to the authorities are the following: * The magnitude of gross capital inflows can be reduced by impos- ing a variety of direct or indirect controls on inflows. * Even if gross inflows are freely allowed, the liberalization of capi- tal outflows or the accelerated repayment ofpublic debt can be un- dertaken to attempt to reduce net inflows. * The implications of a net capital account surplus on the foreign exchange market can be counteracted by accelerating trade liber- alization to increase the current account deficit. * The most extreme option in this category would be to eliminate all foreign exchange market intervention by floating the exchange rate. The resulting appreciation of the domestic currency would both reduce net inflows through the capital account and create a current account offset. Except for floating the exchange rate, all of these instruments were used by policymakers in recipient developing countries. Capital con- trols were adopted in several countries, in particular those that received the largest amounts of private capital flows (Chile, Malaysia, Mexico, and Thailand) or those that were constrained in their ability to use other policy instruments to reduce potential overheating and vulnera- bility (Brazil, Colombia, and Indonesia). In most of these countries, capital controls took the form of new restrictions. Brazil, for instance, enacted financial transaction taxes on foreign purchases of domestic bonds in 1993, and on purchases of domestic stocks in 1994. Mexico in 1992 imposed restrictions on the foreign exchange liabilities of banks. Indonesia in 1991 took several measures to discourage swaps 177 T A L.-II-iP TAL FLOWS TO DEVELOPING COUNTRIES Table 4.2 MIain Policy Responses to Surges in Capital Inflows, 1988-95 New restrie- Liberal- Higher tios izatian Trade Nominal Sterilized reserve Tighter on Of liberal- appred- ter- require- Jiscal Country inflows outflws ization ation vention ments poliy Other EastAsta Indonesia 1991 1990-ongoing 1990-93 1990-94 Repayment of public external debt, 1994; widen- ing of exchange rate band, 1994-95 Korea 1989-94 1992-94 1989 1989, 1992-93 1990 1992-94 Malaysia 1992, 1994 1988-94 1988-94 1992-93 1989-92, '94 1988-94 Philippine; 1992, 1994-95 ongoing 1992 (small) 1990-93 1990 1990-95 Accelerated debt repayment, 1994-95 Thailand 1995 1990-94 1990 1988-95 1938-91 Accelerated debt repaymnent, 1988-90 Latin America Argentina 1991-93 Brazil 1993-95 1994-95 Chile 1991-93 1990-94 1991 slight, 1991-92 1990, 1992 1999-95 Widening large, 1994 1992-93 of exchange rate band, 1992 Colombia 1993-95 1991-94 1991 slight, 1991 1992-95 1991 Exchange large, 1994 rate band adopted, 1994 Mexico 1992 1991 1989-93 1990-93 19B9-93 Exchange rate band adopted, 1991 SoauthAsia India 1992-95 ongoing ongoing 1993-94 1993-94 ]992 Pakistan ongoing ongoing 1993-94 SriLanka 1993 1993 1992-95 1991-93 1991-93 178 CHALLENGES OF MACROECONOMIC Mh and offshore borrowing by state-owned enterprises and tightened the limits on net open foreign exchange positions of commercial banks. The most concerted and sustained efforts at constraining capital in- flows, however, were made by Chile and Colombia, and, in 1994, by Malaysia, as described in box 4.1. In other cases, countries reacted to the surge in capital inflows by not eliminating existing controls. Thai- land, for instance, has not eliminated restrictions on the net foreign ex- change and foreign liability positions of commercial banks and finance companies, and on holdings of deposits in foreign exchange by domes- tic residents. But in 1995 and 1996, Thailand, too, imposed new con- Box 4.1 Capital Controls in Chile, Colombia, and Malaysia CHILE INTRODUCED CONTROLS ON CAPITAL IN- percent. Unlike in Chile, the requirement was to re- flows in June of 1991. These took the form of mini- main in place for the duration of the loan, but it ap- mum nonremunerated reserve requirements of 20 plied only to loans with a maturity of 18 months or percent on new foreign credits (except credits less, except for trade credits. The control regime was granted to exporters) with maturities of less than one tightened in 1994. In March the maturity of loans year. The requirement was extended to all external subject to reserve requirements was extended to three credits regardless of maturity one month later. These years, and in August to five years. In addition, the re- deposits, although applied uniformly across mxaturi- quited rate was set on a graduated scale, vith higher ties, were required to be maintained only for one rates for shorter maturities. The range was from 140 year, whatever the maturity of the loan. This implied percent for loans with maturities of 30 days or less to a tax rate that varied inversely with the maturity of 42.8 percent for those with 5-year maturities. the loan. In 1992, the 20 percent reserve requirement In Malaysia, the experience with capital controls on foreign loans was extended to foreign currency was quite different. Controls were imposed at the be- bank deposits. In May the required reserve ratio was ginning of 1994, in response to a sharp acceleration increased to 30 percent, except for foreign credits of capital inflows in 1993. Measures included the im- with maturity of more than one year that were regis- position of limits on the foreign exchange liabilities tered with the central bank under Article 14 of the of banks, the extension of reserve requirements to foreign exchange regulation. The latter also became such liabilities, a ban on the sale of short-term securi- subject to the 30 percent reserve requirement in Au- ties to foreigners by residents, and the imposition of gust 1992. Finally, in July of 1995 reserve require- a regulation requiring that domestic currency de- ments were extended to all types of foreign posits of foreign institutions be non-interest-bearing. investments in Chile, including the issue of sec- This was followed in February by a halt to trade-re- ondary American depository receipts (ADRs). lated swaps and the imposition of fees on non-inter- In Colombia, capital controls were introduced in est-bearing foreign deposits. Controls were gradually September of 1993. These took the form of nonre- removed over the course of 1994, and by January munerated reserve requirements on direct external 1995 only the reserve requirement for the foreign borrowing by firms, with the reserve ratio set at 47 currency liabilities of banks remained in place. 179 II A1ITAL FLOWS TO DEVELOPING COUNTRIES trols in the form of nonremunerated reserve requirements on a variety of offshore borrowing by banks and finance companies. A decline in net official capital inflows has offset, at least partially, the surge in private flows in several countries. TIl is offset was signifi- cant in Chile, India, Indonesia, Morocco, the Philippines, Sri Lanka, Thailand, and Tunisia (table 4.3)*2 In Chile, Indonesia, and Thailand, the offset was the result of a deliberate policy decision to accelerate the repayment of public external debt in order to reduce the surplus on the capital account of the balance of payments and thus the extent of re- Table 4.3 Disposition of Private Capital Inflows during Inflow Episodes (change) As a percentage of GDP' Net Net Current Change inr Change in private official account Reserve current reserve inflows .nflows deficitb accumuztionc account accumulation Country (1) (2) (3) (4} (percent)(t (percent)e Argentina 2.30 0.02 1.79 0.53 77 23 Brazil 2.75 -0.17 0.64 1.94 25 75 Chile 1.83 -2.06 -4.86 4.62 2,070 -1,970 Colombia 4.91 -0.90 4.90 -0.89 122 -22 Costa Rica -0.26 -3.26 -4.52 0.99 128 -28 -Hungary 15.50 -1.75 9.78 3.97 71 29 India 0.63 -0.56 -1.21 1.28 -1,720 1,820 Indonesia 1.42 -0.65 0.15 0.61 20 80 Korea 4.20 0.26 4.97 -0,51 111 -11 Malaysia 5.95 -0.31 2.86 2.78 51 49 Mexico 7.17 -0.14 7.05 -0.02 100 0 Morocco 3.74 -2.25 0.11 1.39 7 93 Pakistan 2.51 -0.26 0.92 1 33 41 59 Peru 4.43 -0.27 1.38 2.79 33 67 Philippines 3-78 -0.65 0.66 2.47 21 79 Poland 5.93 0.41 3.89 2.44 61 39 Sri Lanka 4.81 -1.26 -0.19 3.74 -5 105 ThailandJ 6.83 -1.21 2.31 3.31 41 59 Tunisia 4.38 -1.47 3.73 -0.82 128 -28 Turkey 2.57 -0.58 1.38 0.61 69 31 Venezuela 9.03 -1.15 14.59 -671 185 -85 a. Columnis 1-4 show the change in the main components of the balance of payments during the respective inflow periods as compared with the tm mediately preceding period of equal length b. A mrinus sign means an improvement in the current account balance. c. A minus sign means a decline in reserve accumulation, d. Column 31(columnns I + 2). e. Column 4t(columns I + 2). Source, World Bank data; iMF, Internataonnl Fiancial Siatistzcsdata base. 180 CHALLENGES OF MACROECONOMIC MWA' serve accumulation required by the central bank. As shown in table 4.2, most countries have continued to liberalize private capital outflows during their inflow episodes, partly as an aspect of ongoing financial liberalization. In some cases, however-Korea being a notable exam- ple-the pace of outflow liberalization has been varied in response to the strength of inflows. Finally, free-floating exchange rates were not adopted by any of the countries in the sample, but Chile, Colombia, and Mexico adopted ex- change rate bands, providing added flexibility for exchange-rate move- ments. In Chile and Colombia, the exchange rate proved to be fairly flexible within the band, accommodating significant nominal appreci- ation in response to inflows, while in Mexico the band was designed to accommodate a small amount of nominal depreciation. The floor of the band was fixed at the peso price of the U.S. dollar in November 1991, when the band was adopted, and the ceiling was set according to a system of minidevaluations. By way of contrast, in most East Asian countries (Indonesia, Korea, Malaysia, and Thailand) nominal ex- change rate policy continued to be directed to the preservation of ex- ternal competitiveness (that is, to prevent excessive appreciation of the real exchange rate) in the face of capital inflows. Offsetting the impact of capital inflows on domestic monetary aggregates. There are two policies that restrict the magnitude of the monetary expansion associated with a given amount of intervention in the for- eign exchange market: * Expansion of base money associated with a given amount of in- tervention can be restricted by sterilizing the effects of interven- tion on the monetary base-that is, by contracting domestic credit to offset the expansion of the net foreign assets of the cen- tral bank, through mechanisms such as open market operations or transferring public sector deposits from commercial banks to the central bank. * Increasing reserve requirements on domestic financial institutions reduces the impact of the expansion of the monetary base on the growth of broader monetary aggregates. Sterilized intervention was the most widely and intensively used pol- icy response to the arrival of capital inflows among the countries in our sample. The extent of sterilized intervention in a country can be esti- mated by examining the composition of the change in its monetary 181 C-APITAL FLOWS TO DEVELOPING COUNTRIES base. Changes in net foreign assets and in domesric credit in opposite directions are suggestive of sterilization operations. This decomposi- tion of the annual change in the monetary base for the countries in the sample, provided in annex 4.1, suggests that most of the countries ster- ilized heavily for at least some portion of their inflow episode, and some (Indonesia, Korea, Mexico, the Philippines, Sri Lanka, and Thai- land) did so fairly continuously over the entire episode.3 The intensity of sterilization tended to vary countercyclically in many cases, with the degree of sterilization increasing when domestic economic conditions were strong and decreasing when they were weak. For example, in Latin America, Brazil sterilized heavily when inflation accelerated in 1992, Chile did so following an acceleration of inflation in 1991, and Colom- bia did so when inflation accelerated in 1992-S3. In East Asia, In- donesia sterilized heavily in response to excess demand pressures through 1992, but eased monetary policy when the economy weak- ened in early 1993. Sterilization was intensified in association with eco- nomic recovery in 1994. The evidence suggests that sterilization proved to be effective in these cases in restraining the expansion of the mone- tary base.4 Sterilization took the form of open market bond sales, central bank borrowing from commercial banks, shifting government deposits from commercial banks to the central bank, raising int,rest rates on central bank assets and liabilities, and curtailing access to rediscounts. Chile, Colombia, and Indonesia all pursued sterilization through open market operations very early and aggressively, seeking to offset all effects of cap- ital inflows on the monetary base, while Korea, Mexico, the Philip- pines, and Thailand were not so ambitious, seekir.g only to ameliorate the effects on the base. Transfers of government or public enterprise de- posits to the central bank took place in Indonesia, Malaysia, and Thai- land, while Mexico sterilized by placing its privatization proceeds in the central bank, in effect selling real assets to absorb the monetary base. India sterilized almost entirely by altering required reserve ratios. In contrast to these countries, Argentina did not rely on sterilized inter- vention, permitting capital inflows to influence its money supply. Changes in reserve requirements have taken various forms, ranging from altering required reserve ratios on all domes nlc currency deposits to raising marginal reserve requirements on only a subset of bank lia- bilities. Chile, Colombia, Malaysia, the Philippines, and Sri Lanka, in particular, relied on repeated increases in average or marginal reserve re- 182 CHALLENGES OF MACROECONOMIC MA quirements. In most of these countries, the increases in reserve require- ments show up as decreases in the broad money multiplier in annex 4.1. Some countries, such as Thailand in 1995, imposed quantitative credit ceilings on banks. Offsetting the impact of monetary expansion on aggregate demand. If the arrival of capital inflows is permitted to result in the expansion of broad monetary aggregates, the expansionary effects on aggregate demand can be neutralized through fiscal contraction. Fiscal adjustment was a key component of the stabilization and mar- ket-oriented reform programs that many countries undertook prior to receiving capital inflows. Consequently, it is difficult to interpret a tight fiscal stance, or a further tightening of that stance, as a policy response to capital inflows rather than as a continuation of an ongoing adjust- ment process. Whatever the reason, a tighter fiscal stance during the in- flow episode does help reduce aggregate demand pressures. In almost all the countries in the sample, the central government's annual fiscal balance as a ratio to GDP improved relative to its average value during the pre-inflow period. As shown in annex 4.1, Brazil, and Sri Lanka were the exceptions. In summary, it is evident that the sample countries reacted rather vigorously to the arrival of capital inflows and that they employed vir- tually the entire arsenal of macroeconomic policy to combat overheat- ing. All of these policy options are available not just to respond to the possibility of overheating, but also to offset the destabilizing effects of external financial volatility on the domestic economy when integration is well established, as will be discussed later in the chapter. The Results of Policies to Prevent Overheating How successful were these policies in preventing the overheating asso- ciated with a surge in capital inflows? Once again, macroeconomic per- formance can be understood in terms of the transmission mechanism that operates during the inflow episodes. According to this mechanism, two conditions are necessary for capital inflows to have an expansion- ary impact on aggregate demand. First, a change in the net private cap- ital inflow must represent an addition to the country's capital account surplus, rather than merely a change in the identity, whether private or public, of the country's external creditors. Second, the central bank must intervene actively in the foreign exchange market, purchasing the 183 CAPITAL FLOWS TO DEVELOPING COUNTRIES additional foreign exchange generated by the capital inflow. To check whether these conditions were present in the sample countries during the inflow periods, table 4.3 decomposes the balance of payments in each of the countries, to account for the disposition of the change in the private capital account surplus among changes in net inflows from official sources, the current account deficit, and reserve accumulation. The table shows that, as noted above, a reduction in flows from offi- cial creditors indeed provided an important offset to the increase in pri- vate inflows in several countries. In particular, Costa Rica did not register a net increase in the capital account surplus during the 1980s and 1990s. Indeed, both net private and official inflows declined during the period, and thus Costa Rica did not face an overheating problem of the type described above. Hence, Costa Rica will be excluded from the sample from now on. The table also confirms that recipient countries did not prevent overheating by allowing their currencies to float. For the majority of countries, reserve accumulation absorbed a significant por- tion of the change in the capital account surplus--more than half in nine countries.5 Thus, in reaction to the arrival of capital inflows, cen- tral banks intervened heavily to prevent the nominol appreciation of the domestic currency. In the absence of foreign exchange intervention, of course, the current account deficit would have completely offset the capital account surplus. This implies that the arrival of capital inflows, together with central bank intervention, resulted in substantial upward pressure on the money supplies of the vast majority of recipient coun- tries-the traditional channel for the generation of overheating. Symptoms of overheating. To see whether these countries actually expe- rienced macroeconomic overheating, we can examine the behavior of four variables during the periods when the countries began to receive large private capital inflows. These variables are: acceleration of eco- nomic growth, large current account deficits, accelerating inflation, and an appreciation of the real exchange rate. The countries in the sample showed no systematic pattern of accel- erating inflation, deteriorating current account, and appreciating real exchange rates. For each of these variables, approximately half the countries experienced the outcome associated with overheating and half did not. Moreover, in each country the foar variables seldom moved consistently in the direction associated with overheating (for ex- ample, many countries in the sample experienced faster growth but lower inflation, other countries experienced higher inflation but lower 184 CHALLENGES OF MACROECONOMIC MAN_ current account deficits, and so on). Moreover, none of the variables provided a stronger signal of overheating in countries that received the largest cumulative capital inflows. Finally, even when a variable moved in a direction consistent with the emergence of overheating, it was often fairly easy to identify factors other than the capital inflow that could have accounted for its performance. For example, among the sharpest accelerations of growth in the sample were those registered by Argentina, Hungary, Peru, and Poland-two transition economies and two that had undergone discontinuous changes in their macroeco- nomic regimes. Clearly in these cases both economic growth accelera- tion and capital inflows were responding to a third factor-the change in policy regime. This is confirmed by the fact that all four of these countries underwent a sharp deceleration of inflation at this time. In- deed, the experience suggests that an acceleration of inflation, perhaps the single most direct symptom of overheating (both because it is un- desirable in itself and because it contributes to an increase in the cur- rent account deficit through real exchange rate appreciation), is by no means unavoidable during a surge in capital inflows. Thirteen of the countries in the sample reduced their inflation rates during their inflow episodes, and five of the 13 saw large decelerations.6 The macroeco- nomic outcomes for these countries are summarized in table 4.4. The surge in capital inflows also did not necessarily result in an ap- preciation of the real exchange rate, since only 12 of the countries in the sample registered a real appreciation during the inflow period. Moreover, none of the five East Asian countries in the sample experi- enced a large real appreciation during their inflow periods: four had large real depreciations, while the other kept its real exchange rate ap- proximately stable. By contrast, six of the seven Latin American coun- tries in the sample registered real appreciation, and in most cases the magnitude was substantial. Argentina, Colombia, Mexico, and Peru registered by far the largest real appreciations in the sample. In addi- tion, at least for most of the countries affected, the mechanism gener- ating the real appreciation was clearly not an acceleration of inflation, as posited at the beginning of this section, because of the 12 countries that registered a real appreciation, 10 did so with reduceddomestic in- flation. This was true, for example, in Argentina, Mexico, and Peru. In contrast to the inflation experience, a stylized fact emerging from this sample is that such inflows are often associated with upward pres- sure on the current account deficit. The current account deficit fell 185 t-I TAL FLOWS TO DEVELOPING COUNTRI ES Table 4.4 Nlacroeconomic Performance during Inflow Episodes (changeft6m immediately preceding period of equal length) Average Average annual annual Average Inflow GDP growth inflation current Average CSounltry epitsodte (percent) (percent) account a J2EE? Argentina 1991-94 9.1 -801.1 1.8 91.7 Brazil 1992-95 3.1 -93.5 0.6 7.4 Chile 1989-95 5.7 -4.1 -4.9 -25.5 Colombia 1992-95 1.6 -4.8 4.9 14.7 Hungary 1993-95 7.5 -5.5 9.8 18.9 India 1992-95 -0.7 0.1 -1.2 -30.8 Indonesia 1990-95 2 2 1.3 0,2 -29.4 Korea 1991-95 -2.5 0.8 5.0 4.4 Malaysia 1989-95 4.0 1.4 2.9 -24.5 Mexico 1989-94 2.9 -74.4 7.1 20.0 Morocco 1990-95 -3.3 0.1 0.1 -6.5 Pakistan 1992-95 -2.3 1.7 0.9 -9.0 Peru 1990-95 3.3 -79.1 1.4 120.9 Philippine, 1989-95 2.2 -3.1 0.7 -10.7 Poland 1992-95 8.5 -146.7 3.9 37.9 Sri Lanka 1991-95 2.0 -2.2 -0.2 0.6 Thailand 1988-95 3.9 -1.1 2.3 - 8.9 Tunisia 1992-95 0.5 -2.2 3.7 0.6 Turkey 1992-93 1.4 5.0 1.4 1.0 Venezuela 1992-93 -5.0 -2.7 14.6 9.8 a. As a pe rcentage of GDP A minus sign indicates an improvement in the current account balance. b. Percen-age change in the real effective exchange rate. A positive number indicates an appreciation Source iM F, World Economsc Outlook data base; IMF, International Financial Statistics data base. during the inflow period in only three of the 20 countries, and in- creases in current account deficits tended to be rmuch larger than de- creases. Only Chile managed to achieve significant reduction in its current account deficit during the period when capital inflows increased. The real exchange rate and current account outcomes have been as- sociated, even if only weakly, with each other. As figure 4.1 suggests, countries with the largest real appreciations have also tended to exhibit the largest increases in current account deficits. Omitting Argentina and Peru, whose enormous real appreciations are clearly an exception, the simple cross-country correlation between real appreciation and in- creases in the current account deficit in the sample is 0.60. 186 CHALLENGES OF MACROECONOMIC M Figure 4.1 Change in the Cunrent Account Deficit and the Real Effective Exchange Rate during Inflow Episodes, Selected Countries Appreciation of the real effective exchange rate (percent) 140 Peru 120 U 100 Argentina 80 60 Poland 40 U Colombia Mexico Hungary 20 U Venezuela Bradil Turkey U Korea Thailand -20 India Chile U * Malaysia Indonesia -40 I -6 -4 -2 0 2 4 6 8 10 12 14 16 Change in current account deficit (percentage of GDP) Note Inflow episodes are compared with the immediately preceding period of equal length Source IMF, International Financial Statzstics data base In summary, the countries in the sample have avoided the overt Larger current account deficits are symptoms of macroeconomic overheating, except for pressures on the associated with real exchange rate current account. While four of the countries have exhibited large spurts appreciation during capital inflow of economic growth and one country (Brazil) underwent a burst of near hyperinflation, the factors behind these experiences were unusual, easily identifiable, and not attributable to the arrival of capital inflows. Substantial accelerations of inflation-the factor most directly indica- tive of overheating-have been almost completely absent in this group of countries, although inflationary pressures may recently have in- creased in those countries that received the most flows in East Asia. While several countries did have a large appreciation of the real ex- change rate, regional differences suggest these may have been related to country-specific nominal exchange rate policies rather than to a gener- alized phenomenon such as the arrival of capital inflows. This interpre- 187 $C-APITAL FLOWS TO DEVELOPING COUNTRIES tation is supported by the observation that real appreciation has tended to be associated with a deceleration, rather than an acceleration, of in- flation. However, consistent with concerns about overheating, increases in the current account deficit were much more common and larger than decreases, and tended to be correlated with the real exchange rate outcome. Although, unlike countries in the previous lending boom, most countries have not absorbed all the private capital flows through a rise in the current account deficit and have increased reserves, the size of the current account deficits is of concern in certain countries, espe- cially some of those that received the largest amounts of flows. The composition of absorption. To the extent that capital inflows have altered the current account-to-GDP ratio in these countries, the level of absorption relative to income will have changed. The composition of this change-that is, its allocation between consumption and investment-may in turn influence the effects of the inflow on the real exchange rate as well as on the economy's growth rate. Thus the next step in analyzing the macroeconomic consequences of the inflows is to examine changes in the composition of absorption. For 15 of the 20 countries in the sample, as shown in table 4.5, the arrival of capital inflows was associated with an increase in the ratio of investment to GDP. In some cases the increase WaLs dramatic. In Thai- land, for example, the ratio of investment to GDP increased by 13 per- centage points. Only in Peru and Poland did the ratio decline sharply. The increase in the investment ratio in the large majority of countries implies that the consumption-to-GDP ratio would have to decline for the current account to improve.7 By and large, however, this did not happen: while three-quarters of the countries had increases in the in- vestment ratio, few of them financed these incrzases through reduc- tions in the consumption ratio, thus resulting in the higher current account deficits described above. How are these changes in the composition of absorption related to symptoms of overheating described previously? First, figure 4.2 sug- gests that the behavior of economic growth in the sample countries is related to changes in the composition of absorption. Countries that in- creased the share of investment in GDP also tended to register larger in- creases in the rate of economic growth during the inflow period. Leaving aside the outliers in growth performance identified earlier (Ar- gentina, Hungary, Peru, and Poland), the simple correlation between changes in the investment ratio and in the grewth rate among the 188 CHALLENGES OF MACROECONOMIC MAN Table 4.5 Composition of Absorption during Inflow Episodes (change from immediately preceding period of equal length) As a percentage of GDP Country Iflow episode Current a"count deficita Total investment Total consumption Argentina 1991-94 1.8 O.6 4.4 Brazil 1992-95 0.6 -2.0 3.6 Chile 1989-95 -4.9 10.2 -8.5 Colombia 1992-95 4.9 0.9 4.1 Hungary 1993-95 9.8 1.6 6.4 India 1992-95 -1.2 -1.3 -1.7 Indonesia 1990-95 0.2 5.7 -5.2 Korea 1991-95 5.0 4.7 1,1 Malaysia 1989-95 2.9 4.8 -1.8 Mexico 1989-94 7.1 2.4 6.7 Morocco 1990-95 0.1 -1.1 0.8 Pakistan 1992-95 0.9 1.0 -2.0 Peru 1990-95 1.4 -4.0 3.1 Philippines 1989-95 0.7 1.7 6.1 Poland 1992-95 3.9 -11.1 11.3 Sri Lanka 1991-95 -0.2 2.2 -1.9 Thailand 1988-95 2.3 13.4 -11.2 Tunisia 1992-95 3.7 2.6 -1.4 Turkey 199_-93 1.4 1.3 -0.5 Venezuela 1992-93 14.6 6.8 6.8 a. A minus sign indicates an improvement in the current account balance. Source. iMF, World Economic Outlook data base; IMF, International Financial Statistics data base. countries in the sample is weakly positive, amounting to 0.32.' Second, as shown in figure 4.3, changes in the composition of absorption also appear to be correlated with changes in the real exchange rate. In par- ticular, an increase in the consumption-to-GDP ratio is associated with real appreciation, the simple cross-country correlation (excluding the two outliers, Argentina and Peru) being 0.74. While this positive correlation may reflect causation running from consumption to the real exchange rate (for example, reflecting a greater intensity of nontraded goods in consumption than in investment) or the reverse (reflecting higher domestic real interest rates when the real exchange rate appreciates), the underlying stance of fiscal policy ap- pears to have played a role. Several countries that significantly tight- ened fiscal policy (as indicated in annex 4.1 )-including Chile, Indonesia, Malaysia, and Thailand-also reduced their share of con- sumption in GDP during the inflow period. Moreover, all of these coun- 189 ¾$ CAITAL FLOWS TO DEVELOPING COUNTRIES Figure 4.2 Change in Growth and the Composition of Absorption during Inflow Episodes, Selected Countries Change in average annual GDP growth rate (percent) 10 * Argentina Poland O * Hungary 6 -- - -ilChiie Thailand 4 r - *Malaysia Peru Brazil 2 * * B * Indonesia Tunisia O 0 India n Pakistan * Korea Morocco O -4 * Venezuela -6 -15 -10 -5 0 5 10 15 Change in investment rate (percentage of GDP) Note Inflow episodes are compared with the immediately preceding period of equal length Source IMF, 'World Economic Outlook data base Improvements in growth performance tries also achieved a substantial real exchange rate depreciation during are associated with increases in the inflow period. Among these countries, Chile also reduced its cur- investment raltes during capital rent account deficit during the inflow period. Wlhile the current ac- inflow episodes. count deficit increased in Malaysia and Thailand these two countries also experienced very substantial rises in the share of investment in GDP, as well as in their rates of economic growth, in association with the arrival of capital inflows. Policy Conclusions Four main conclusions can be drawn from this country experience: * The majority of countries attempting to cope with the effects of integration eschewed direct intervention to restrict capital in- 190 CHALLENGES OF MACROECONOMIC MA-- Figure 4.3 Change in the Composition of Absorption and the Real Effective Exchange Rate during Inflow Episodes, Selected Countries Change in consumption (percentage of GDP) 15 Poland 10 Phiippnes Venezuela Mexico Philippines * P Hungary 5 U oomi U Colombia Argentina S Peru * * Korea 0 Malaysia India -5 * Indonesia * Chile -10 * Thailand -15 -40 -20 0 20 40 60 80 100 120 140 Appreciation of the real effective exchange rate (percent) Note Inflow episodes are compared with the immediately preceding period of equal length Source IMF, International Financial Statistics data base, IMF, World Economic Outlook data base flows. However, capital controls were adopted by seven of the Real exchange rate appreciation is countries in the sample, in particular those that received the associated with increases in largest amounts of private capital flows or those that were con- consumption rates during capital strained in their ability to use other policy instruments to contain overheating. * Countries did not make wholesale changes in their foreign ex- change regimes. Bands were introduced by Chile, Colombia, and Mexico, but only those of Chile and Colombia seemed designed to accommodate some nominal appreciation in response to capital inflows. In the context of adherence to predetermined exchange rates, outcomes with respect to the real exchange rate generally de- pend in the short run on how the nominal exchange rate is man- aged. In this respect, the experience of the recipient countries was diverse. Several countries indeed experienced real appreciation, 191 IMM APITAL FLOWS TO DEVELOPING COUNTRIES but this was typically associated with a decline in the inflation rate. Thus, the emergence of real appreciation in these countries did not reflect overheating but rather the use of a he exchange rate as a nominal anchor in the context of price stabilization. Where the nominal rate was managed with the goal o F increasing competi- tiveness, as in the several East Asian countrizs in our sample, real appreciation either failed to materialize or was minimal. * Tight monetary and fiscal policies were the key weapons in the fight against overheating. Tight monetary policy generally took the form of sterilized intervention in the foreign exchange mar- ket. This tool was effective in restricting expansion of the mone- tary base (suggesting that financial integration had not proceeded to the point of removing short-run monetary autonomy from these countries). Sterilization also proved Lo be a flexible tool, with several countries varying the intensity of sterilization in re- sponse to domestic economic conditions. Similarly, the majority of countries appear to have tightened fiscal policy in response to inflows, albeit to very different degrees. * These differences in fiscal performance appear to be closely re- lated to outcomes with regard to the composition of absorption and the real exchange rate. A particular reliance on tight fiscal policy appears to have been conducive to avoiding the symptoms of overheating while producing more limiled real appreciation, lower current account deficits, and a higher share of investment in absorption. Policy Regimes and Vulnerability IN THE CONTEXT OF FINANCIAL INTEGRAT]ON, VULNERABILITY refers to the possibility that a country may Find itself confronted with a sudden, large, and relatively long-lasriing reduction in net capital inflows. The macroeconomic dislocations produced by such an event can be extreme, as experienced by many developing countries in the context of the international debt crisis of the 1980s, and more recently by Mexico. In addition to avoiding overheating, therefore, newly integrated developing countries that are importing large quanti- ties of capital need to set the avoidance of such vulnerability as an important policy objective. The challenge to policymakers is to iden- 192 CHALLENGES OF MACROECONOMIC MAN Ct tify and implement policies that can minimize (or at least do not magnify) vulnerability to external financial shocks. One way to identify such policies is to examine the experience of capital-importing countries in the aftermath of the Mexican crisis. Macroeconomic characteristics of countries that were successful in weathering the accompanying financial storm (the tequila effect) can be credited with reducing their vulnerability. The severity of the tequila effect for a developing country is measured in terms of two criteria: the ability of the country to sustain private capital flows in 1995 at a level approaching those of earlier periods and the behavior of equity prices in the country during the first half of the year. These indicators were chosen because portfolio flows and equity prices were relatively volatile in the wake of the Mexican crisis, and thus were highly sensitive to changes in cross-country investor sentiment. These criteria are used below to separate countries into two groups-those heavily affected by the Mexican crisis and those not so affected. The sorting of countries into those heavily influenced by the tequila effect and those not so affected is shown in table 4.6. The tequila effect proved to be important for Argentina, Brazil, India, Pak- istan, Turkey, and Venezuela, in addition to Mexico itself.9 Countries that weathered the storm well, according to our criteria, include Chile, Colombia, Indonesia, Korea, Malaysia, and Thailand. Only three countries did not clearly fall into one or the other category-Chile, Colombia, and Turkey. Although Chile and Colombia were somewhat affected by the Mexican crisis, they were classified as less affected on the basis of their stock market performance, which remained some- what stable despite their having undergone reductions in portfolio flows. The stock market measure is more reliable for Latin American countries because it is less likely to be contaminated by regional con- tagion among international investors. Turkey is also an ambiguous case because stock market prices remained strong despite a sharp drop in capital flows. But because of this drop, Turkey has been placed in the "strongly affected" category. What distinguishes these two groups of countries? Table 4.6 pro- vides a list of potential discriminating characteristics. It may be useful to begin our discussion with the size of the current account deficit. While Mexico's current account deficit was the largest among the coun- tries in the crisis group for several years before the crisis, four of the six countries in the noncrisis group had current account deficits in 1993 193 Table 4.6 Vulnerability to the Mexican Crisis, Selected Countries, 1988-93 (percent) Impact of shack Change Comry, dament in Change Cs fupdreentalf portfolio in Av. As a perentage af D equity stock FDII 1993 Average Aerag flows, prices, total Av. Central Central DODI GDP Export Change Group and 1994- 12/94- private GD/ Av. gov't. CA gov't. CA XGS growth growth in Country 95 3/95 inflows GDP inflation balnce balance balance balance 1993 rate rate BEER Strongly afected Argentina -75.6 -25,7 65.9 16.3 990.9 1.7 -3.1 -2.9 -0.4 408.5 2.6 6.5 121,0 Brazil -41.4 -36.1 28.0 22.3 1,375.4 -0.1 -0.2 -0.9 0.3 312.3 0.7 9.3 16.6 India -73.1 -18.3 4.9 21.9 9.5 -7.1 -0.4 -7.4 -2.1 290,6 5.3 7.6 -41.5 Mexico -64.2 --58 7 39.5 19.6 34.8 0.4 -6.5 -2.6 -4.2 195,1 2.7 3.6 40.7 Pakistan -45.3 -21.0 58 5 18.0 8.5 -7.7 -7.7 -7.9 -4.1 271.5 4 8 7.1 -15.1 Turkey -52.8 16.7 16.7 22.8 66.6 -6.7 -3.5 -4.3 -0,5 226,9 4.2 7.2 30.1 Venezuela -31.0 -14.1 64.8 18.7 43.1 -2.5 -3.3 -2.5 1.3 211,9 3.3 7.1 -11.6 Average -54.8 -22.5 39.8 20.0 361,2 -3.1 -3.5 -4.1 -1.4 273.8 3.4 6.9 20.0 Less affected Chile -77.3 -8.7 55.4 23.2 17.9 2.2 -4.5 3.6 -1.8 167.8 7.5 11,0 9.6 Colombia -59.1 0.6 35.8 17.1 27.2 -0.3 -3.8 -0.8 0.6 172.4 3.8 9.3 0,9 Indonesia -9.4 -6.0 18.9 27.2 8.2 0.0 -1,7 -0.3 -2.2 211.7 8.2 9.7 -3.4 Korea -10.5 -5.2 16.3 33.2 7.0 0.3 O.l -0.2 0.8 48.9 7.9 7 8 2.8 Malaysia 60.5 2.8 60.5 31.3 3.5 0.2 -4.6 -2.1 -2.3 53.0 8.7 13.6 -3.4 Thailand -378.4 -4.5 30.8 37.6 4.8 2.0 -4.9 3.3 -5.4 86.9 10.3 16.8 2.3 Average -79.0 -3.5 36.3 28.2 11.4 0.7 -3.2 0.6 -1.7 123.5 7.7 11.4 1.5 Note: CA, current account; GDI, gross domestic Investment; DOD/XCs, debt outstanding disbursed/exports of goods and services, RFFR, real effective exchange rate. All averages and differences are for 1988-93 unless otierwise noted. Source: IMF, World Ecoononzc Outlook data base; World Bank data. CHALLENGES OF MACROECONOMIC M that, relative to the size of their economies, exceeded those of all coun- tries in the crisis group except Mexico itself and Pakistan. The factors that seem most important in discriminating between the two groups can loosely be associated with either macroeconomic out- comes or macroeconomic policies: * Macroeconomic outcomes. The two groups differ with respect to the precrisis rate of real output growth, the composition of ab- sorption, and the rate of inflation. Of the six noncrisis countries, all except Colombia exhibited growth rates faster than the fastest- growing crisis country. Overall, these countries averaged 7.7 per- cent annual real growth; the country with the lowest share of investment in GDP among the noncrisis countries had a higher ratio than any of the crisis countries. The noncrisis countries av- eraged a 40 percent higher share of investment in GDP than the crisis countries. As mentioned earlier in this chapter, changes in growth and in the share of investment in absorption are positively correlated in the broader country sample from which this sub- sample was drawn. Finally, the inflation outcomes of the two groups were also quite different. Countries in the crisis group had much higher average inflation during 1988-93 than those in the noncrisis group, even after excluding Argentina and Brazil from the former. * Macroeconomicpolicies. Real exchange rate outcomes were quite different in the two groups of countries. Four of the seven coun- tries in the crisis group experienced very substantial real appreci- ation prior to the crisis. In contrast, only Chile in the noncrisis group experienced real appreciation, and Chile's total real appre- ciation was much less than that of any of the appreciating coun- tries in the crisis group.10 Overall, the crisis countries registered a 20 percent real appreciation, on average, during 1988-93, com- pared with 1.5 percent for the noncrisis countries. Fiscal policy also proved to be very different for the two groups for the five years preceding the Mexican crisis. On average, the central gov- ernment budget in the noncrisis countries had a surplus of 0.6 percent of GDP during 1988-93, while the crisis countries had an average deficit of 4 percent of GDP over the same period. In 1993, the central government balance still registered an average surplus (of 0.7 percent of GDP) in the noncrisis countries while the crisis 195 A I -CAPITAL FLOWS TO DEVELOPING COUNTRIES countries remained in deficit. The third difference between the two groups was the debt-export ratio, with al L but one of the crisis countries exhibiting a ratio larger than the largest value registered among any of the noncrisis countries." How do we make sense of these results? The most important lesson for policymakers is that since creditors (both domestic and foreign) can be expected to take their capital out of a country when they think that a pol- icy change could impair the value of their investment, then vulnerability will arise when the perception is created that a devaluation, nonpayment of public sector debt, or the imposition of restrictions on capital outflows is about to occur. Such expectations are likely to arise when the real exchange rate is perceived to be out of line, the government's debt obliga- tions are large, fiscal adjustment is perceived as politically or administra- tively infeasible, or the country's growth prospects are bleak. From the perspective of creditors, therefore, a high share of investment in absorp- tion and a strong record of growth, a low stock of government obliga- tions coupled with demonstrated fiscal flexibility (in the form of small deficits and low inflation), and a real exchange rate broadly perceived to be in line with fundamentals all augur well for future debt service. The evidence suggests that these characteristics are most directly associated with a country's ability to avoid vulnerability. From a policy perspective, countries need to have an active exchange rate policy that avoids sub- stantial appreciation of the real exchange rate and responsible fiscal poli- cies. Evidence from the previous section suggests that this policy mix, in turn, is associated with stable real exchange rates, a relatively larger share of investment in GDP, and faster economic growth. Macroeconomic Management with Growing Financial Integration S HORT OF THE CRISIS SITUATION ASSOCIATED WITH VULNERA- bilbty, cross-border capital flows may exhibit substantial volatil- ity-that is, temporary shocks to either capital inflows or out- flows caused by changes in world economic conditions or by less extreme changes in portfolio managers' perceptioris of domestic cred- itworthiness. This increased volatility may well be a fact of life for 196 CHALLENGES OF MACROECONOMIC MANA4 developing countries as they become more integrated with interna- tional financial markets, even after stock adjustment inflows associ- ated with the transition have tapered off, and even if severe crises are avoided. This section examines general principles of macroeconomic management for newly integrated countries facing relatively brief and frequent fluctuations in cross-border capital flows. The question for policymakers is how to frame macroeconomic policies in the presence of such volatility, and in particular, what broad principles should guide policy in several key areas of macroeconomic management.12 Earlier in this chapter we listed the tools available to domestic poli- cymakers for coping with financial volatility. In broad terms, they con- sist of policies regarding the free movement of capital, exchange rate policy, monetary policy, and fiscal policy. This section will consider the role of each of these policies in macroeconomic management and how they may be linked. Restricting the Free Movement of Capital The most direct response to external financial volatility is to contain its impact by restricting the movement of capital-that is, imposing capi- tal controls. However, economists have long questioned both the effec- tiveness and the desirability of such restrictions. We take up each of these issues in turn. Capital controls are of questionable effectiveness. Although economists tend to have strong views about the effectiveness or ineffectiveness of capital controls, empirical evidence on the issue is ambiguous. Assess- ing the effectiveness of controls is complicated by a host of factors, including the definition of effectiveness itself (that is, what is the objective to be achieved) and the conditions that determine the effec- tiveness of different types of controls (see box 4.2). Are Capital Controls Desirable? Even if capital controls can be effective, that does not mean they should be used. Since effective controls distort private economic decisions, the benefits of imposing them should outweigh the costs. This section eval- uates four arguments for restricting capital movements as a way to con- trol volatility and achieve macroeconomic stabilization under a high degree of financial integration. 197 T 0-A1 TAL FLOWS TO DEVELOPING COUNTRIES Box 4.2 Objectives of Capital Contros and Conditions for Effectiveness Objectives. The effectiveness of capital controls can a The extent of intermational cooperation in re- only be measured relative to the objective they are porting cross-border claims. In the case of con- designed to achieve. There are three basic types of trols on outflows, for example, host countries objectives: to preserve some degree of damestic may be unwilling to declare the assets of resi- monetaly autonomy over a period of time, to restrict dents from capital-exporting countries, the magnitude of net capital flows into or out of the whether or not these residents try to circum- country. and to affect the composition of capital vent foreign exchange regulations. flows. H-iowever, because monetary autonomy can- a The size of the misalignment motivating inflows not be achieved unless the magnitude of net flows is or outflows. If effecting a capital inflow or out- restricted, these three objectives boil down to two: flow for an arbitrage operation involves a fixed restricting the size of net capital flows and affecting cost imposed by the control regime, then the their comnposition. deterrent effect of this cost may be nullified if Coonditions for effectiveness of diferent types of eon- the arbitrage gains promise to exceed the trots. The effectiveness of controls may depend on a cost,2 variety cf factors that not only differ across countries a The desngn of the controls themselves, in particu- but also change over time as the process of financial lar their comprehenswveness. Controls that are integration proceeds.1 These factors indude: not comprehensive-for example, those that apply to only certain types of flows-can be * The state of technology This affects tmnsaction evaded by changing the com position of flows. costs in conducting arbitrage among different fi- While such controls would be ineffective in nancial centers. The lower the costs, the more preserving monetary autonomy (or limiting difficult itwiXl be to implementeffective controls. the total magnitude of flows), they may, how- * "Incredible" reform. The first argument is that controls may be ef- fective in coping with distortions that arise vvhen the government implements an incredible stabilization or trade liberalization (one that observers do not believe it can carry out). Lack of confidence in the government's ability to sustain its announced exchange rate or tariff policies may make external borrowing appear temporar- ily cheap, with potentially destabilizing effects on the domestic economy.13 The best solution to this problem is for the govern- ment to take steps to achieve credibility-most commonly, by getting its fiscal accounts into order-since fiscal problems are often the cause of policy reversal. If this is not feasible, controls on capital movements may be a second-best policy. * Insulation from shocks arising in the internati6 nalfinancial market. The second argument for capital controls on both inflows and out- flows is that they can insulate the domestic economy from identifi- 198 CHALLENGES OF MACROECONOMIC MAN, ever, still be able to alter the composition of a The efficiency of the controlling bureaucracy. flows. * The structure of the domestic financial system. In practice, restrictions on capital movements This may determine the effectiveness of par- have taken many forms, ranging from the outright tial controls that leave room for evasion by prohibition of a wide array of capital account trans- channeling flows through alternative domes- actions (such as those that have given rise to parallel tic intermediaries. foreign exchange rnarkets in both industrial and de- * The size of trade flows. This would determine veloping countries) to prudential restrictions on the the scope for under- and overinvoicing, as acquisition of foreign assets by domestic pension well as for altering leads and for lags on trade funds or of foreign liabilities by domestic banks. credit. Different restrictions designed to achieve different * The types of cross-border flaws targeted by the objectives can be expected to vary in effectiveness, contro--that is, controls on capital outflows both in achieving their stated purposes and in the may differ in effectiveness from controls on wider sense of financially insulating the domestic inflows, all other things (that is, potential ar- economy from external shocks. bitrage margins) being equal. One argument 1. Some analysts have argued that controls of any type is that the residence of the capital-importing can be effective under any arbitrary set of circumstances, or -exporting agent may matter. Domestic while others have argued that controls of any type can residents may be more prepared than foreign never be effective. Neither argument is valid. agents to evade restrictions, making controls 2. This means, in particular, that the size of capital agents to ~~~~~~~~~~flows in the presence of controls is nor a reliable indicator on outflows less effective than controls on of the effectiveness of controls, since large flows may sim- inflows. ply indicate large arbitrage opportunities. able external financial shocks-that is, they can limit the incidence of volatility by preserving some degree of financial autarky. Since controls directly address the source of the shock, they may, it is ar- gued, succeed in stabilizing the economy without introducing new distortions. This argument for quantitative controls, however, is less persuasive than the first. From the country perspective, fluctu- ations in world interest rates are just an intertemporal dimension of what terms of trade fluctuations are intratemporally. If intratempo- ral price fluctuations do not require insulation, it is not dear why intertemporal fluctuations should do so. a Preservation ofshort-run monetary autonomy. Under perfect capital mobility, the effectiveness of monetary policy can be preserved only if exchange rate policy is flexible. In this case, however, un- pleasant tradeoffs may arise between internal (inflation) and ex- ternal (current account) balance targets. Tight monetary policy 199 Qop,W-purl`TAL FLOWS TO DEVELOPING COUNTRIES adopted to combat inflation, for example, would increase the do- mestic real interest rate while causing the real exchange rate to ap- preciate, possibly resulting in a deterioration of the trade balance. Capital controls would enable the monetary authorities to avoid this tradeoff between targets by making it possible to preserve monetary autonomy with an officially determined exchange rate. Thus controls would, in effect, provide a second independent macroeconomic instrument-either monetary policy or the ex- change rate-to simultaneously address the two targets of infla- tion and the external balance of payments. The benefits for macroeconomic management are large, however, only if other stabilization instruments are not available. If fiscal policy is suffi- ciently flexible to be used for stabilization purposes, for example, then price stability and satisfactory current account performance can be pursued through the combined use of fiscal and exchange rate policy.i4 * Changing the composztion of inflows. Finally, there is the argument that controls can be used to alter the composition of capital in- flows. Even if controls can be used for this purpose, whether or not they should be depends on whether the composition of in- flows matters. Unfortunately, there is currently no consensus on this issue. While researchers have provided some support for the idea that reliance on short-term flows is associated with enhanced vulnerability and volatility, the evidence is mixed. * A separate argument for being concerned with the composition of flows is that certain types of inflows may be driven by implicit government guarantees (for example, external borrowing by do- mestic financial institutions) and thus may not be welfare en- hancing. But even if such flows could be discouraged by capital controls, it is not clear that the same implicLt guarantees do not also apply to other types of flows (such as direct lending to do- mestic firms) through less direct means (Dooley 1996). If they do, then the case for altering the composition of flows would have to rely on the differential welfare effects olf direct government guarantees extended to financial institutions For different types of flows. This case seems plausible a priori. Summary and conclusion. Empirical assessment of the effectiveness of capital controls is an inexact science (box 4.3). AlLhough many devel- 200 CHALLENGES OF MACROECONOMIC M}A5 Box 4.3 Effectiveness of Capital Controls in Chile, Colombia, and Malaysia THE NATURE OF THE CAPITAL CONTROLS THAT Colombia and found no statistically significant im- were imposed in each of these three countries was de- pact for the tax term. At the same time, however, they scribed in box 4.1. Since these controls were imposed found that the tax term affected domestic interest relatively recently, evidence on their effectiveness is rates (suggesting an enhanced degree of monetary au- limited and somewhat contradictory. Net private tonomy) and the black market premium. They also capital inflows contracted in Chile in 1991 and found a substantial change in the term structure ofex- (rather drastically) in Malaysia in 1994, the years in ternal borrowing at the time that regulations were which the controls were implemented. Inflows accel- changed. They interpret this evidence as suggesting erated in Colombia in 1994, however, despite the that controls did not affect the total magnitude of imposition of controls in September of the previous capital inflows into Colombia but altered their com- year. position toward longer maturities. This conclusion is The problem with such before-after assessments, supported by Quirk and Evans (1995), who based of course, is that other factors could have accounted their judgment on an analysis of changes in flow com- for these outcomes in all three countries. Unfortu- position following the implementation of controls. nately, there is no more systematic evidence avail- For Chile, analysts have examined the change in able for Malaysia. (The consensus among observers, the composition of flows and have estimated capital however, is that the Malaysian controls were quite flow equations. Le Fort and Budnevich (1996) and effective in stemming short-term capital inflows; Quirk and Evans (1995) all conclude, from their see, for instance, Quirk and Evans 1995.) The analyses of changes in the composition of Chile's ex- Chilean and Colombian experiences have been stud- ternal financing after the imposition of the unremu- ied in more detail. nerated reserve requirement, that the composition One insight from these studies is that the effective- of flows was altered in the direction of longer matu- ness of capital controls in altering the magnitude of rities. On the other hand, Valdes-Prieto and Soto total capital flows, or of particular types of flows, can (1996) estimated capital flow equations for short- be measured by capital flow equations in which the term capital and found that the tax variable did not interest rate differential is adjusted by adding an esti- have a statistically significant effect on aggregate mate of the implicit tax rate imposed by measures flows of short-term capital, although a negative and such as unremunerated reserve requirements. Carde- statistically significant effect appeared for at least nas and Barreras (1996) followed this procedure for one important component. oping countries that maintained controls during the initial period of financial integration seem to have preserved a meaningful degree of monetary autonomy, the effectiveness of controls is likely to decrease as the countries become more integrated. As conditions change in these countries, the evidence from industrial countries may become more relevant for them, and controls will be able to preserve a degree of monetary autonomy for only a limited period of time. In addition, controls have not been able to prevent large capital outflows and 201 £PdA'ITAL FLOWS TO DEVELOPING COUNTRIES inflows in response to large prospective arbitrage profits. Finally, the available evidence suggests that controls can be effective in altering the composition of flows. With regard to the desirability of implementing controls, we can conclude that: * Prudential restrictions on external borrowing and lending by do- mestic financial institutions may plausibly be warranted. These restrictions on capital movements have a clear second-best ratio- nale, since they are directed at a specific distortion: the inability of the government to credibly commit to removing certain im- plicit guarantees. * Temporary restrictions on capital movements designed to address "incredible liberalization" or "incredible stabilization" problems are also warranted if credibility is not achievable, again on sec- ond-best grounds. Such restrictions may be of limited effective- ness, however, if the size of the perceived arbitrage margin is large. * Restrictions on capital inflows for more general macroeconomic stabilization purposes may be motivated by Lhe desire to insulate the domestic economy from "pure" external financial shocks or to preserve monetary autonomy in an effort to free up an addi- tional policy instrument. The first goal is questionable, since it is based on the premise that the government is better informed about the duration of shocks than the private sector. With re- gard to the second objective, the best solution to this problem may be to increase the flexibility of fiscal policy rather than adopt capital controls. At best, this objective provides an argu- ment for retaining restrictions as a transitory device, until the fiscal system can be reformed to make fiscal policy an effective stabilization tool. Exchange Rate Policy Exchange rate regimes. Different exchange rate systems can be and have been used by countries while becoming integrated with the rest of the world-in particular, pure floating; managed, or dirty, floating; fixed exchange rate; or a currency (or quasi-currency) board. Each system has consequences for the effectiveness of monetary and fiscal policy. Thus, while in a pure floating system fiscal policy is less effective and 202 CHALLENGES OF MACROECONOMIC MANA monetary policy more effective, in a fixed exchange rate system the opposite results (Mundell 1968). Therefore, the advantages or disad- vantages of each system vary from country to country depending on the nature-nominal versus real-of the shocks affecting each coun- try and the ability of the authorities to have a flexible fiscal policy. Countries in which fiscal policy cannot be modified effectively in the short run should not relinquish monetary policy completely by adopt- ing a fixed exchange rate system. In recent years an increasing number of economies have adopted "flexibly managed" exchange rate systems, which give them the option of effectively using monetary policy at the cost of eroding their credi- bility regarding inflation targets. The few exceptions have been economies where the credibility of the government was extremely low-Hong Kong during the 1980s and Argentina and Estonia during the 1990s-all of whom adopted currency boards because of the need to regain market confidence.15 The major difficulty facing these economies, however, is to maintain a flexible enough fiscal policy and build cushions-such as a large stock of international reserves-to im- prove the resilience of the economy to shocks. In the aftermath of the Mexican crisis, many thought that the days of managed exchange rate systems were over. However, most countries were able to go through the Mexican crisis without modifying their ex- change rate systems. The evidence to date shows that international cap- ital markets have been selective since the Mexican episode and that fundamentals matter (Sachs, Tornell, and Velasco 1996). Thus, there are indications that officially determined exchange rates can be man- aged successfully within a consistent macroeconomic policy frame- work. In particular, the success of flexibly managed systems depends on two things: * whether policymakers can actually succeed in tracking the equi- librium real exchange rate, so that fundamentals-driven specula- tive attacks can be avoided * whether governments can achieve enough credibility to cause market expectations to coalesce around the "good" equilibrium in a multiple-equilibrium situation, thereby avoiding speculative at- tacks driven by self-fulfilling expectations. These are the two most important sources of volatility associated with a managed exchange rate regime. The experiences of Chile, 203 fALHTAL FLOWS TO DEVELOPING COUNTRIES Colombia, and the four East Asian countries that were not affected by the Mexican crisis suggest that both of these objectives are within the reach of policymakers in developing countries. Exchange rate management. Since managing the exchange rate indeed seems to be the preference of most developing countries, the issue of how to manage the rate in a financially integrated world becomes an important one. One way to approach this problem is to consider the exchange rate regime as consisting of a band around a moving central Box 4.4 Recent Country Experience with Exchange Rate Bands THE BANDS THAT HAVE BEEN ADOPTED IN bound, implying a gradually increasing band width. developing countries to date typically have been im- The daily change in the peso-dollar parity was imi- plemented on the heels of exchange rate-based sta- tially set at 20 cents per day, when the band was bilization programs. They thus represent the adopted in November 1991, and was later increased 'flexibilization" stage of such programs. While the to 40 cents per day in October 1 992, allowing for bands share this common background, the choices an accelerated annual rate of depreciation of the that individual countries made in their implemen- peso. By the time of the Mexican crisis, in Decem- tation have been rather different. All of these coun- ber 1994, the width of the band had increased to 14 tries h2ve adopted a crawling central parity, a key percent. Colombia formally introduced a crawling feature of the developing-country bands that distin- band in January 1994, although a de facto band ex- guishes them from the European ERM, but the rules isted under the certificados de cambio system that governing the behavior of the central parity have was adopted in the attempt to ieduce the cost of differed across countries. In Chile, the central par- sterilizing capital inflows.' The i riitial rate of crawl ity has been adjusted continuously, with an- was set at 1 Ipercent and was increased to 11.5 per- nounced daily depreciations for the coming month, cent at the end of 1994. based on a forecast of foreign and domestic mnfla- Band widths tended to increase over time in tion over the coming month. Periodic revaluations each of these cases. As already indicated, the Mexi- reflectinig perceived changes in "fundamentals" (as- can band widened automaticallv over time, since sociated with capital inflows) have been superim- its upper bound was depreciased continuously, posed on this gradual nominal depreciation. In while its lower bound was fixed. Band widths also Israel, by contrast, the central parity was initially increased over time in Chile and Israel, but in dis- fixed relative to a basket of currencies in January of crete fashion and relative to an announced parity. 1989, 2nd then in December 1991 it was allowed to The width of the Chilean band siarted out at 2 per- crawl at an annual rate determined by the difference cent around the central parity in 1985 and had in- between the government's targeted rate of inflation creased to 10 percent by 1992. Similarly, the width over the coming year and a forecast of foreign infla- of the Israeli band was increased from 3 percent to tion. MLexico announced no central parity but fixed 5 percent around the central parity in March 1990. the lower bound of the band and announced a pre- Colombia began with a total band width of 12.5 determined rate of daily depreciation for the upper percent under the certificados cle cambio system 204 CHALLENGES OF MACROECONOMIC MM parity. The decisions that need to be made then consist of how to adjust the parity, how wide to make the band, and how to intervene within the band. The experiences of Chile, Colombia, and the East Asian countries (box 4.4) suggest that the objective of setting the central parity should be to maintain competitiveness-that is, the central parity should track the long-run equilibrium real exchange rate-to prevent expectations of discrete realignments. This means adjusting the parity not only in but widened the band to plus or minus 7 percent which distinguishes these bands from their Euro- when a crawling peg was formally adopted at the pean counterparts-is an important clue to their beginning of 1994. Concerning intervention inside survival up to the present in three of the four coun- the band, only Chile appears to have made use of tries that have adopted them. the full band width, with the exchange rate regu- Finally, the adoption of a band does not larly approaching the upper and lower bounds. Is- represent a magic solution to credibility problems. rael and Mexico were much more active to restrict The Chilean, Israeli, and Mexican bands were all fluctuations to a narrower zone inside the band. characterized by periods in which expectations of The relevant lessons from experience with ex- realignment-associated with the behavior of change rate bands in developing countries can be "fundamentals"-emerged, even before the Mexi- summarized as follows (see Helpman, Leiderman, can crisis. The long cycles in interest rate differen- and Bufmnan 1994 ). First, moving to a band from tials associated in timing with the renewal of the a fixed rate, or to a band with a crawling central Pacto agreements in Mexico, as well as the behav- parity from one with a fixed parity, has not obvi- ior of interest rate differentials in Israel around the ously been associated with a loss of price stability. time of realignments, suggest that markets identify Second, bands have been associated with a variety episodes of misalignment and act quickly on expec- of real exchange rate experiences. In Chile, the real tations of devaluation. In Colombia, revaluations exchange rate depreciated during the early years of occurred twice (at the beginning and end of 1994), the band, and it appreciated after capital inflows when market pressures were strong and the ex- began to arrive in 1989. In Mexico and Colombia, change rate was at the lower end of the band. Sim- by contrast, the exchange rate was associated with ilar events occurred in Chile during 1991-94. fairly continuous real appreciation. Israel's real ex- change rate has been relatively stable since the 1. Under this system, sellers of foreign exchange re- adoption of the band. These differences reflect dif- ceived dollar-denominated claims on the central bank, ferent weights attached to competitiveness and denoted certificados de cambio. These could be re- price stability by the authorities in the manage- deemed at maturity (initially three months, but extended men.ofthecenralpartynduggsthat ro tO one year in August of 199 1) for the full face value at men fthcenralpaitadue the then-prevailing exchange rate, redeemed before ma- crawling nature of the central parity-which allows turity at 87.5 percent of face value, or sold freely in the this variety of real exchange rate outcomes and market. 205 APJTAL FLOWS TO DEVELOPING COUNTRIES accordance with the domestic-foreign inflation differential but also in accordance with changes in the underlying equilibrium real exchange rates, which are driven by permanent changes in fundamental factors such as the terms of trade, commercial policy, fiscal policy, and condi- tions in external financial markets.16 Recent evidence indeed suggests that five East Asian economies-Indonesia, Malaysia, the Philippines, Singapore, and Thailand-have all tracked their long-run equilibrium values fairly closely during recent periods of substantial capital inflows in those countries (Montiel 1996). On the assumption that the central parity seeks to track the long-run equilibrium real exchange rate, the next issues are how wide the band should be and how much intervention should take place within it. The desirable width of the band depends on the value to the domestic econ- omy of an independent domestic monetary policy. The larger the scope for the exchange rate to deviate from its central parity-that is, the wider the band-the greater the scope for an independent domestic monetary policy. In turn, the usefulness of an independent monetary policy for re- ducing volatility depends on the availability of alternative stabilization instruments (a flexible fiscal policy) and on the sources of shocks to the economy. The traditional analysis of this issue focuses on the implica- tions of shocks in the domestic money (nominal) and goods (real) mar- kets, the standard prescription being that, holding fiscal policy constant, domestic real shocks call for exchange rate flexibility, while domestic nominal shocks call for fixed exchange rates. This suggests that if fiscal policy is not available (or is costly to use) as a stabilization instrument, countries in which domestic real shocks predominate should adopt fairly wide bands, while those in which domestic nominal shocks are dominant should keep the exchange rate close to its central parity. With a flexible fiscal policy, however, domestic real shocks can be countered through fis- cal adjustments, thereby diminishing the value of irndependent monetary policy as a stabilization instrument. Thus the adoption of a fairly narrow band is more likely to be consistent with the stabilization objective if fis- cal policy is available as a stabilization instrument. Regarding intervention within the band, the logic of this analysis suggests that for a given band width, active intervention should accom- pany nominal shocks, whereas real shocks instead call for some combi- nation of exchange rate and fiscal adjustment. Consider, however, an alternative source of shocks-that is, "pure" external financial shocks, say, in the form of fluctuations in world interest rates. In this case, ex- 206 CHALLENGES OF MACROECONOMIC MA change market intervention policy determines the form in which the shock is transmitted to the domestic economy. If world interest rates fall, for example, and active (unsterilized) intervention keeps the ex- change rate fixed at its central parity, domestic monetary expansion and lower domestic interest rates will ensue-an expansionary shock. If, on the other hand, the central bank refrains from intervention, the domes- tic currency will appreciate in real terms and the domestic real interest rate will rise-a contractionary shock. Purely from the perspective of stabilizing domestic aggregate demand, the appropnate intervention policy within the band in this case depends on the direction in which it is feasible or desirable to move fiscal policy. If fiscal policy is literally in- flexible, then the choice confronting monetary authorities is between real appreciation (under no intervention) or overheating (under full un- sterilized intervention). To avoid both overheating and real appreciation requires a mix of fiscal contraction and intervention to keep the nomi- nal exchange rate close to its central parity. Conclusions. Experience suggests that crawling pegs remain a viable exchange rate option for developing countries that become integrated into world capital markets. However, the scope for deviating from fundamentals, as well as from commitment to the announced regime, is much reduced under these circumstances. Managed rates are cer- tainly capable of generating macroeconomic volatility in this environ- ment, but they have not generally done so. The following conclusions can be drawn with regard to exchange rate management: * The central parity should be managed so as to track, to the extent possible, the underlying long-run equilibrium real exchange rate. This means not just offsetting inflation differentials but also ad- justing the real exchange rate target to reflect permanent changes in fundamentals. Large and persistent temporary misalignments should be avoided, primarily because they threaten the sustain- ability of the regime and make speculative attacks more likely. * Small temporary deviations from the central parity can play a useful stabilizing role when fiscal policy is inflexible and either real domes- tic shocks dominate or exchange rate flexibility is required to ensure that external shocks affect domestic demand in the right direction. * In general, fiscal flexibility can substitute for exchange rate flexi- bility as a stabilizing device. 207 thPAP1TAL FLOWS TO DEVELOPING COUNTRIES Monetary Policy Fixing an exchange rate target in the face of capital movements implies central bank intervention in the foreign exchange market, which has the effect of altering the stock of base money. Sterilized intervention is the indicated policy for a government attempting to simultaneously run an independent monetary policy (targeting either some monetary aggre- gate or some domestic interest rate) and fix the nominal exchange rate. As already indicated, however, if capital mobility is high, such an at- tempt may not be successful in the absence of capital controls. The purpose of sterilized (as opposed to unsterilized) intervention is to prevent a change in the demand for domestic interest-bearing assets from causing too large a change in the price of those assets, essentially by meeting the demand shift with a supply response. Thus, sterilization in response to capital inflows involves an increase in the supply of do- mestic debt in one form or another. As in the case of capital controls, the general issues that arise in this context are not only whether steril- ized intervention can work to stabilize domestic aggregate demand but also, if it can, whether it is desirable. As shown in annex 4.2, even if sterilization remains possible for financially integrating developing countries, its effectiveness in insulating domestic demand from exter- nal financial shocks is questionable. Sterilization is most effective when domestic interest-bearing assets are close substitutes among themselves but are poor substitutes for foreign interest-bearing assets. Under these circumstances, sterilized intervention can insulat. domestic aggregate demand from transitory portfolio shocks. Howxever, the conditions necessary for sterilized intervention to be effective imply that its effec- tiveness may depend on how it is attempted. Because bank borrowers may not have access to securities markets, for e, ample, sterilizing by raising reserve requirements on banks is likely to be less effective in in- sulating the domestic economy from portfolio disturbances than is sterilizing through open market bond sales. If sterilization is possible, when is it beneficial? The answer is that sterilization is beneficial whenever the prices of domestic assets need to be insulated from shocks; that is, whenever the economy experiences transitory shocks to portfolio preferences-domestic nominaL shocks or external financial shocks-or when the authorities seek to accom- modate a permanent change in portfolio preferences in a gradual fash- ion. When this happens, domestic aggregate demand can be stabilized 208 CHALLENGES OF MACROECONOMIC MANA: by preventing changes in portfolio preferences from being transmitted to the real sector through changes in exchange rates and asset prices. On the other hand, domestic real shocks do not call for sterilized inter- vention, since in this case the asset price adjustments triggered by the shock are likely to prove stabilizing. Conclusions. The point for policymakers is that sterilized interven- tion may be indicated when an economy experiences shocks to port- folio preferences. Even when it is desirable for stabilization purposes, however, sterilization may carry a fiscal cost, particularly for govern- ments whose claimed intentions not to devalue or default on debt are not fully credible, resulting in high domestic interest rates.17 Thus, the use of this tool without impairing the government's solvency requires fiscal flexibility. Countries that lack such flexibility may be tempted to sterilize through changes in reserve requirements, despite the likelihood that the use of this tool will result in imperfect insula- tion, because the fiscal implications of doing so are less adverse than those of open market operations. This advantage, however, is at the cost of implicit taxation of banks and their customers. Fiscal Policy From the standpoint of reducing volatility, the key characteristics of fis- cal policy are short-run flexibility, the perceived solvency of the public sector, and its vulnerability to liquidity crises. Short-run fiscal flexibility plays an important role in neutralizing shocks. The importance of short-run fiscal flexibility was emphasized earlier in this chapter. Recall, for example, the reference case of a "pure" external financial shock for a country with well-integrated financial markets. In this case, sterilized intervention is not an option, and thus the country has only one independent monetary policy instrument. Faced with an external financial shock, the domestic monetary authorities can choose a value for either the exchange rate or the domestic money supply (and thus the domestic interest rate), but not for both. There- fore, when both the level and the composition of aggregate demand are important, the authorities will find themselves one instrument short and may face an unpleasant choice between, say, stabilizing domestic demand and safeguarding the competitiveness of exports. This tradeoff suggests an important role for short-run fiscal flexibil- ity. If fiscal policy can be counted upon to sustain the level of aggregate 209 ,,AkTITAL FLOWS TO DEVELOPING COUNTRIES demand at its preshock value (by adopting a more or less expansionary stance as needed), then the choice of monetary response can be based on the desired composition of aggregate demand. In the absence of short- run fiscal flexibility, however, the nature of the monetary response may depend on tradeoffs between the level and composition of demand. How can fiscal flexibility be achieved? There are at least two impor- tant constraints tending to limit the flexibility of fiscal policy. The first concerns inflexibility of fiscal instruments, arising from inefficient tax systems that associate large excess burdens with po [icy-induced changes in tax revenues, as well as from political imperatives that tend to drive up the level of public expenditures. Structural measures that remove rigidities on either the expenditure or revenue side of the government's budget-such as privatization of state enterprises and tax reform designed to widen the tax base and remove egregious distortions in marginal tax rates, as well as to improve the efficiency of tax adminis- tration-can thus make an important contribution to enhancing fiscal flexibility. The second restriction on fiscal flexibility arises from the be- havior of creditors. If fiscal profligacy during good times causes credi- tors to question fiscal solvency during bad times, then countercyclical fiscal measures will be ruled out by an inability to finance deficits dur- ing downturns. The upshot is that overly expansionary fiscal policy in response to positive shocks is likely to make fiscal policy procyclical in both directions by constraining fiscal flexibility when the economy is hit by adverse shocks (Gavin and others 1996). Thus a key step in achieving symmetric fiscal flexibility is the implementation of mecha- nisms that permit the fiscal authorities to restrain spending when times appear to be good. Recent research suggests that institutional aspects of the budget process may play important roles in restraining the growth of expenditures during such times. Transparency in the budget process, as well as a hierarchical, rather than collegial, process of budget formu- lation that permits the finance ministry to restrict the growth of spend- ing by line ministries when budget constraints appear to be eased, have been associated empirically with a more restrained fiscal stance, on aver- age. Thus, broader institutional reforms designed to safeguard the per- ceived solvency of the public sector may be needed in some cases to complement more conventional structural reforms if fiscal policy is to achieve the desired degree of flexibility. Public sector solvency is important to creditors. Beyond the role of short- run stabilization in response to external financia] shocks, fiscal policy 210 CHALLENGES OF MACROECONOMIC MA;Ah plays a more fundamental role in the context of increased financial integration, when both the direction and the magnitude of capital flows are likely to become very sensitive to perceptions of domestic public sector solvency. Potential insolvency can generate large capital outflows and large interest rate premiums, as creditors try to avoid taxation of domestic assets while demanding compensation for expos- ing themselves to the risk of taxation they face by continuing to lend in the domestic economy. This situation is aggravated when the gov- ernment makes explicit guarantees to creditors, however, since the value of the guarantees will fluctuate with the government's perceived financial ability to back them. The key point is that in the context of high financial integration, the stock dimension of fiscal policy, in the form of changes in the government's perceived net worth, may itself represent an important source of shocks to the domestic economy, transmitted through the terms on which both domestic and foreign creditors are willing to hold claims on the domestic economy. The stock and flow dimensions of fiscal policy are not independent. A critical link between them is created by the fact that what matters for creditors is the perceivedsolvency of the public sector. For a government whose long-run fiscal stance is uncertain, short-run policy changes will be scrutinized for information about the government's longer-run in- tentions. Knowing this, governments may be reluctant to act in ways that may be perceived as sending the wrong signal to creditors, and this reluctance may limit the government's short-run policy flexibility. Thus, achieving a reputation for fiscal responsibility may maximize the government's short-run policy flexibiLity. Debt management policies will determine the likelihood of a debt run. Public sector solvency requires that the public sector's comprehensive net worth be positive. However, the need to preserve macroeconomic sta- bility in a financially integrated environment may impose stricter con- ditions on the public sector's balance sheet than simply maintaining a positive value of comprehensive net worth. The composition of assets and liabilities may matter as well. In particular, a public sector that is solvent (that is, one that can credibly honor its obligations over a suf- ficiently long horizon) may nevertheless be vulnerable to short-run liquidity crises. If the public sector is perceived as unlikely to honor its short-term obligations, then creditors will be reluctant to take on the government's short-term liabilities, and in a vicious cycle, the gov- ernment may then be unable to meet its short-run obligations. The 211 S CAPITAL FLOWS TO DEVELOPING COUNTRIES likelihood of such a debt run depends on the maturity and currency composition of the public sector's liabilities relative to that of its assets-that is, on the government's debt managernent policies. In managing the composition of its debt, the government faces a dif- ficult tradeoff between enhancing its credibility, on the one hand, and exposing itself to liquidity crises, on the other. The existence of long- term (fixed-interest) domestic-currency-denominated (nominal) debt provides the government with some financing options that it does not have if its debt is short-term and denominated in foreign currency- that is, it can effectively repudiate the long-term debt by inflating or devaluing, thereby reducing the debt's real value. G iven the nominal in- terest rate on this debt, it may indeed be sensible for a welfare-maxi- mizing government to do so, since by acting in this way it would have the option of sustaining productive expenditures or reducing distor- tionary taxes. However, the prospect of the government exercising this option would raise domestic nominal interest rates, making it expen- sive for the government to borrow long-term in nominal terms. And even if the government never intends to behave in i:his fashion, the time inconsistency problem involved may make it very difficult for the gov- ernment to convince its creditors of its honorable intentions. To reduce its borrowing costs, the government may therefore be induced to bor- row short-term and in foreign currency. By doing so, it eschews the op- tion of gaining from devaluation or inflation at the expense of its creditors, and thus enhances the credibility of its promise to do neither. The problem is, of course, that in doing so it incurs liquid foreign-cur- rency-denominated liabilities, thereby making itself vulnerable to debt runs, as happened in the Mexican crisis, described in box 4.5. The way out of this dilemma is to note when it arnses in acute form-that is, when the government actually retains the discretion to act as creditors fear, when it lacks credibility on other grounds (for ex- ample, when it has a reputation for acting in a discretionary fashion), and when the government's revenue needs are higyh and conventional taxation is highly distortionary. In other words, the existence of long- term nominal debt is only one factor in the government's decision to devalue or inflate, and creditors can rationally expect the government to forgo the option of inflating away the real value of their assets if it is institutionally unable to do so, if it is perceived as placing a high value on the credibility of its policy announcements, or if inflatng creates few net benefits from the government's perspectixe. Thus, the govern- 212 CHALLENGES OF MACROECONOMIC MANA Box 4.5 The Mexican Tesobono Crisis THROUGH A COMBINATION OF DEBT RESTRUC- percent by January 1995. By the end of 1994, when turing, privatization, and a succession of overall fiscal the balance of payments crisis hit, the Mexican gov- surpluses, the Mexican government achieved a dras- ernment faced payment obligations on the stock of tic reduction in its total stock of outstanding debt Tesobonos amounting to $17 billion for the first six over the period from 1988 to 1994. By the beginning months of 1995. of 1994, the stock of Mexico's outstanding public This stock of obligations exceeded Mexico's for- sector amounted to about 30 percent of GDP, most of eign exchange reserves-$6 billion at the end of it in long-term, domestic-currency-denominated in- 1994-as well as credit lines of $8 billion available to struments. However, events during the first four the country under NAFTA (North American Free months of 1994 (the Chiapas rebellion in January, Trade Agreement) arrangements. Consequently, the increase in U.S. interest rates beginning in Feb- there was no prospect of repayment in the event of a ruary, and the assassination of presidential candidate debt run unless exceptional arrangements were made Luis Donaldo Colosio in March), coupled with an with official creditors. The events surrounding the ongoing and substantial real appreciation of the peso, balance of payments crisis in December, as well as the created expectations of devaluation that dried up governmenes failure to put in place adequate excep- capital inflows and built a substantial devaluation tional arrangements until February of 1995, coordi- premium into domestic interest rates. Since the cur- nated the expectations of creditors around the debt rent account deficit was quite large, the central bank run equilibrium, and in the early weeks of 1995 the incurred substantial reserve losses. The government Mexican govemment proved unable to roll over its responded by sterilizing the monetary impact of the Tesobono debt. The effect was to transform a cur- reserve losses and reducing the cost of refinancing rency crisis into a broader crisis of confidence in maturing government debt by issuing short-term macroeconomic policy, which required much stricter dollar-denominated instruments (Tesobonos). The and more costly adjustment policies than most ob- share of Tesobonos in Mexican government debt in- servers would have predicted as a result of the cur- creased from 5 percent in January 1994 to almost 55 rency devaluation alone. ment can avoid making its borrowing costs overly sensitive to the com- position of its debt by creating institutions that limit its discretion (for example, by increasing the independence of the central bank), by es- tablishing a reputation for nondiscretionary behavior, and by choosing levels of expenditure and mobilizing sources of taxation that minimize distortions.19 Under these circumstances, the option to borrow long term in domestic currency terms may be retained, and the likelihood that macroeconomic stability will be impaired by runs on government debt would be minimized. In sum, when financial integration is high, short-run fiscal flexibil- ity is very important, since it provides an additional instrument for sta- bilizing domestic aggregate demand in response to external financial 213 TL&YTAL FLOWS TO DEVELOPING COUNTRIES shocks, thereby freeing monetary or exchange rate policy to address other macroeconomic objectives. Moreover, a stable perception of fiscal solvency becomes crucial for preventing the domestic public sector it- self from becoming an important source of macroeconomic shocks, transmitted through the terms on which creditors are willing to hold claims on the domestic economy. Indeed, the perception of public sec- tor solvency may itself be the most important component in freeing up fiscal policy to play a short-run stabilizing role. Finally, under high fi- nancial integration, institutional arrangements that limit the govern- ment's ability to act in a discretionary fashion, or that enhance its incentives to avoid doing so, may preserve the government's access to low-cost long-term finance, thereby preventing the emergence of debt runs that could introduce an important source of macroeconomic in- stability even when the public sector is solvent. Lessons for Macroeconomic Management I-NCREASED FINANCIAL INTEGRATION IN DEVELOPING COUNTRIES has been driven in part by improvements in macroeconomic man- agement. Continued good performance in this area will be crucial for many of these countries to retain the strong Lies to world capital markets that they have forged over the last several years. In spite of the progress that has been made, however, integration continues to pose important macroeconomic challenges. In this chapter, we have con- sidered three of these challenges: the possibility of overheating, partic- ularly in the initial stock adjustment phase of financial opening up; the potential vulnerability of financially open economies to sharp reversals in capital flows; and the need to cope with new sources of macroeconomic volatility. To address these challcnges, policymakers have four types of broad policy instruments at their disposal: controls on the free movement of capital, exchange rate policy, monetary pol- icy, and fiscal policy. On the basis of experience, a successful strategy for addressing all three of these issues, using a combination of these tools, can be summarized as follows: * The central component of good macroeconomic management under increased financial integration is the same as under finan- cial autarky: responsible and flexible fiscal management. However, 214 CHALLENGES OF MACROECONOMIC MAN fiscal policy management becomes even more important under financial integration. Fiscal policy can be critical to avoiding overheating when capital inflows materialize, it can help to ad- dress the primary causes of vulnerability, and it can provide a valuable instrument to cope with volatility. * While the avoidance of overheating implies relying on tight mon- etary policy as well as on fiscal policy if a country adheres to offi- cially determined exchange rates, the monetary-fiscal mix seems to matter. Countries that rely more heavily on tight fiscal policy tend to generate a mix of absorption in response to capital inflows that favors investment over consumption and to generate both less real appreciation and smaller current account deficits. In the sample of countries examined in this chapter, the orientation to- ward investment was associated with faster economic growth.20 * This mix of macroeconomic outcomes was also associated with reduced vulnerability to the tequila effect that accompanied the Mexican crisis. Thus, an active fiscal policy may play an impor- tant role not only in securing the growth benefits of financial in- tegration but also in reducing one of the dangers to which financially integrated countries can become exposed: vulnerabil- ity to external financial crisis. * Finally, fiscal policy has an important analytical role to play in managing volatility more generally, for several reasons: (a) Flexible fiscal policy provides a short-run stabilization instru- ment that reduces the incentive to restrict capital movements to stabilize the economy while using exchange rate policy to maintain competitiveness. (b) Fiscal flexibility and solvency are likely to reduce the inci- dence of "incredible" reforms, which create distortion-in- duced cross-border capital flows. (c) Fiscal flexibility is necessary to ensure that the equilibrium real exchange rate tracks the actual rate under exchange-rate- based stabilization in a world of high capital mobility. (d) Under managed exchange rates, the availability of fiscal pol- icy as a stabilization instrument determines the desirability of exchange-market intervention in response to domestic real shocks. It also determines the feasibility of simultaneously adjusting the level and composition of aggregate demand in response to "pure" external financial shocks. 215 *',PA;APITAL FLOWS TO DEVELOPING COUNTRIES (e) Because sterilized intervention has fiscal effects, the feasibil- ity of this policy under imperfect capital mobility depends on whether these effects can be accommodated by fiscal policy. (f) Sterilized intervention may also, through its effects on the stock of public sector debt, affect the perceived incentives facing the fiscal authorities. The consequences of this for do- mestic macroeconomic stability depend on the credibility of the authorities' announced fiscal stance. (g) Perhaps most significant of all, fiscal policy affects the per- ception of public sector solvency. This rnay be the most im- portant determinant of country risk, and thus of cross-border capital movements. (h) Finally, even a solvent government may be vulnerable to debt runs if it issues a large stock of short-teim foreign-currency- denominated debt. Retaining the option to borrow long-term in domestic currency requires that the government establish a reputation for nondiscretionary behavior and structure its ex- penditures and revenues so as to limit its perceived incentives for debt repudiation through devaluation or inflation. • Another important policy lesson is that exchange rate management appears to remain possible under the degree of financial integra- tion that developing countries have achieved up to this point. However, as in the case of fiscal policy, appropriate exchange rate management is particularly important when r.here is a high degree of financial integration. In particular, an active exchange rate pol- icy that prevents the emergence of misalignment becomes crucial, particularly in avoiding vulnerability to external financial crises. This does not imply a loss of control over the domestic price level. The countries in our sample that have managed their ex- change rate with the objective of maintaining competitiveness, thereby achieving a stable or depreciating real exchange rate, did not exhibit deteriorating inflation performance in the context of capital inflows. In short, a policy mix that rzlies on nominal ex- change rate management to maintain compe cLitiveness while plac- ing significant weight on fiscal policy to achieve internal balance is fairly effective in both avoiding overheating and reducing vulnerability. * If the nominal exchange rate is to be official]ly determined, then monetary policy will have a role in macroeconomic management 216 CHALLENGES OF MACROECONOMIC MANAtf40 only if natural or policy-imposed barriers to capital movements remain strong enough to preserve some degree of monetary au- tonomy. The evidence to date suggests that they do, despite in- creased financial integration. Sterilized intervention therefore continues to be a feasible policy to preserve macroeconomic sta- bility in response to cross-border capital flows. The flexibility of this policy has made it the instrument of choice for the majority of developing countries receiving capital inflows. These countries have sterilized aggressively and in a variety of ways, in a largely successful effort to prevent inflows from generating macroeco- nomic overheating. Use of this policy to control the economy is subject to many caveats, however. When potential arbitrage mar- gins are large, for example, the sterilization of inflows may turn out to be so expensive that it impairs fiscal credibility, thus prov- ing counterproductive, while the sterilization of outflows may provoke Mexican-style balance-of-payments crises. Even in less extreme circumstances, the domestic asset substitutability condi- tions necessary for sterilization to effectively stabilize aggregate demand are unlikely to be met, whatever form of sterilization is adopted. Thus, sterilization is, at best, an imperfect instrument for stabilizing domestic demand. And in the specific case of capi- tal inflows, the evidence suggests that a policy mix which restrains demand with less weight on sterilization and more on restrictive fiscal policy has been more beneficial than one which does the opposite. * Finally, the role of restrictions on capital movements as a tool for in- tegrating countries to achieve stabilization needs to be carefully circumscribed. While capital controls have not been able to pre- vent large outflows or inflows in response to large prospective ar- bitrage profits, they have been found to be effective in the short term in reducing the magnitude and altering the composition of net capital inflows. The most compelling macroeconomic argu- ments for the use of controls are that they are second-best pru- dential regulations on the behavior of financial intermediaries, as discussed in the next chapter, and they are a transitional device to preserve an enhanced degree of monetary autonomy pending the reforms required to permit the use of fiscal policy as a stabiliza- tion instrument. However, in the longer term, both the effective- ness and the desirability of capital controls are questionable. 217 *>iw CtAPITAL FLOWS TO DEVELOPING COUNTRIES Regarding desirability, the central problem, of course, is that by restricting the degree of financial integration, controls limit the gains that can be derived from integration- f for no other reason than they create perceptions of an unfriendly environment. And as developing countries become more financially integrated, the evidence from industrial countries suggests that controls will lose effectiveness. 218 CHALLENGES OF MACROECONOMIC MANA.'4 Annex 4.1 Table 4.7 Monetary and Fiscal Policies during Capital Inflow Episodes (percent) Year during episode Pre- Inflow inflow Country and indicator epis_ed perioda 1 2 3 4 5 6 7 8 Argerntzna GrawtbofM2 1991-94 988.6 141 3 62.5 46 5 17.6 Growth of monetary base 2,135.7 1163 40.7 36.1 8.5 Changeinnetforeignassets - 18.0 121.3 30.7 0.2 Change in domestc credit - 98.3 -80.5 5.5 8.3 Changem multiplier 12.0 11.6 15.5 76 8.4 Central government balance -5.6 -1.2 0.4 1.7 -0 5 Inflation 1,467.0 171 7 24.9 10.6 4.2 Brazil GrowthofM2 1992-95 1,224.2 1,606.6 2,936.6 1,146.3 38.9 Growth of monetary base 1,263.8 1,148.2 2,424.4 2,241.7 11.9 Change in net foreign assets 726.7 1,763.0 5,017.1 2,019.9 57.4 Change in domestic credit 537.1 -614.8 -2592.7 221 8 -45.5 Change in multiplier 62.9 36.7 20.3 -46.8 24.1 Central government balance -1.1 -0 9 -0.1 1.6 -1.7 Inflation 1,337.0 1,008.7 2,148.4 2,668.5 84.4 Chie GrowthofM2 1989-95 30.5 31 2 23,5 28.1 23.3 23.4 11.3 25.8 Growthofmonetarybase 65.2 -3.3 58.3 226 16.2 14.6 21.9 15.8 Chatge in net foreign asses 6.9 8.7 45.5 178 27.7 13.7 21.9 8.1 Changeindomnesticcredit 58.4 -120 12.8 48 -11.5 0.9 -0.1 7.6 Charnge in multiplier -120 356 -22.0 4.4 6 1 7.7 -8.6 8 7 Central govemmentbalance -1.4 6.1 3.5 2.5 3.0 2.2 2.2 3 9 Inflanon 20.3 17 0 26.0 21.8 15.4 12.7 11.4 8.2 Colombia Growth of M2 1992-95 - 45.0 37.5 34.6 21.5 Growthofmonetarybase 34.1 36.7 25.2 26.3 10.1 Changein net foreign assets 43.8 63.4 17.5 5.6 27.9 Change in domestic credit -9.7 -26.7 7.7 20.7 -17.8 Change in multiplier - 6 1 9.8 6.6 10.3 Central government balance -0.9 -1.3 -0.3 -0.2 0.0 Inflation 28.4 27.0 22.6 23.8 21.0 (Table contintes on thefollowingpage) 219 , VCAPITAL FLOWS TO DEVELOPING COUNTRIES Table 4.7 (continued) Year duering episode Pre- fnflow inflow Country and indicator episode period' 1 2 3 4 5 6 7 8 Indonesia Growth of A42 1990-95 26.1 44.6 17.5 19.8 - - - Growth of mnoaetary base 13.7 163 3.3 19.7 - - Change in net foreign assets 16.1 58.1 62.3 76.6 - - - Change in domestic credit -2 3 -41.8 -59.0 -56.9 - - Change in multiplier 11 8 24.3 13.7 0.1 - - Central gosernment balance -1.1 1.3 0.0 -1.2 -0.7 0.0 0.8 Inflation 7.5 7.8 9.4 7.5 9.7 8.5 9.4 Korea Growth of V2 1991-95 18.6 21.9 14.9 16.6 18.7 15.6 Growth of monetary base 227 18,2 10 9 27.5 9.2 16.3 Change in set foreign assets 224 -0.9 18.2 16.5 18.2 17. Change in domestic credit 04 190 -7.3 1l.0 9.0 -1.5 Change in multiplier -1.9 3.1 3.6 -8.5 8.7 -0,6 Central government balance 0 0 -1 6 -0.7 0.3 0.5 0.4 Inflation 4.9 9 3 6.2 4.8 6.3 4.5 Malaysia Growth of M2 1989-95 9 3 15,2 10.6 16.9 29.2 26.6 12.7 20.0 Growthofmonetarybase 7.6 24.3 22.7 14.5 21.8 11.6 36.2 24.7 Change in net foreign assets 13.3 27.9 36.3 18.9 80.7 115.5 -29.3 -11.5 Change in domestic credit -5.7 -3 6 -13.6 -4.4 -58.9 -103.9 65.5 36.1 Change in multiplier 1 7 -7.3 -9.9 2.1 6.0 13.4 -17.3 -3 8 Central government balance -7.3 -3.2 -3.0 -2.5 -0.9 0.2 2.8 1.2 Inflation 2.5 2.8 2.6 4.4 4.8 3.5 3.7 5.3 Mexca Growth of M2 1989-94 63.6 115 9 75.8 49.3 22.8 14.5 21.7 Growth of monetarybase 47.9 102 35.3 27.8 14.4 10.4 21.2 Change m net foreign assets 25.3 -03 3a.2 75S 16&6 49.3 -103.0 Change in domestic credit 22.7 10 5 5 1 -48.0 -2.2 -38.9 124.2 Change in multiplier 10 0 95.9 30 0 16.8 7.4 3.7 2.5 Central government balance -9.5 -5.0 -2.8 -0.2 1.6 0.4 -£.8 Inflarion 92.9 20,0 26.7 22.7 15.5 9.8 7.0 220 CHALLENGES OF MACROECONOMIC MANA{ Year during episode Pre- Inflo inflow Contsy and indicator episode perioda 1 2 3 4 5 6 7 8 Marocco Growth ofM2 1990-95 12.9 21 5 16.8 9.3 7.9 10.2 7 0 Growth of monetary base 13.6 24.3 24.8 -8 0 8.6 6.2 5 2 Changemnnetforeignassets 3.1 43.8 23.9 16.3 15.1 11.0 -148 Change in domestic credit 10.5 -19.5 0.9 -24.3 -6.5 -4.9 20 0 Change in multipioer -0.4 -2.3 -6.4 18.7 -0.6 3.8 1.8 Central government balance -6.6 -0.6 -1.0 -2.2 -2.1 -2.9 -5.0 Inflation 6.2 6.9 8.0 5 7 5.2 5.1 6.1 Philippines Growth ofM2 1989-95 15.0 30.1 22.5 173 13.6 27.1 24.4 24.2 Growth of monetasybase 21.2 389 17.7 201 13.11 18.9 5.1 17.2 Change in net foreign assets -67.7 19.3 -33.7 63.7 44.8 42.1 19.2 13.9 Change in domestic credit 88.9 19.6 51.3 -43.6 -31.8 -23.2 -14.1 3.3 Change in multiplier -4.1 -6.3 4.1 -23 0.5 6.9 18.4 6.0 Central government balance -3.2 -2 1 -3.5 -2.1 -1.2 -1 6 -1.6 -1.4 Inflation 15 3 12.2 14.1 18.7 8.9 7.6 9.1 8.1 Sri Lanka Growth ofM2 1991-95 12.8 224 16.4 23.1 192 19.4 Growth of monetary base 15.4 279 8.5 25.2 19.1 16.1 Change in net foreign assets -11.2 37.3 24.4 54.8 23,9 8.3 Change in domestic credit 26.6 -9.5 -15.9 -29.7 -4.8 7.8 Change in multiplier -1.9 -4.2 7.3 -16 0.1 2.9 Central government balance -9.6 -9.8 -6.1 -7.1 -9.1 -8.8 Inflation 10.7 12.2 11.4 11.7 8.4 7.7 Thailand Growth ofM2 1988-95 18.8 18 2 26.2 26.7 19.8 15.6 18.4 12.9 17.0 Growthofmonetarybase 11.2 14.9 16.9 186 13.3 17.9 16.1 14.5 22.6 Change in net foreign assets 5.6 48 4 74.6 62 5 56.2 35.1 44.3 38.1 51 8 Change in domestic credit 5 6 -33 6 -57.7 -43 9 -43.0 -17.3 -28.2 -23.6 -29 2 Change in multiplier 69 2.9 8.0 6.8 5.8 -1.9 1.9 -1.5 -4.5 Central governmentbalance -3.7 2.6 4.2 4.6 4.1 2.5 2.0 2.0 2.6 Inflation 6.1 3.8 5.4 6.0 5.7 4.1 3.4 5 2 5 7 (Table continues on the following page) 221 __ALjTAL FLOWS TO DEVELOPING COUNTRIES Table 4.7 (continued) Year during episode Pre- Inlow inflow C2ountsy,endindicator episode- peri-da 1 2 3 4 5 6 7 8 T7unisia Growth of M2 1992-95 11 6 8. 6.1 8.1 6.6 Growrh of nonetary base 1210 7.2 4.8 7.2 9.4 Change in niet foreign assets 9.4 10.6 2.5 374 6.3 Change in domestic eredit 2.5 -3.4 2.3 -30.1 3.1 Change in mnultiplher -02 1.1 13 0.8 -2.6 Central government balance -4.5 -2.4 -2.9 -2,7 -4.2 Inflation 7.4 5.8 4.0 4.7 6.2 - Not available Note: M2 refers to the sum of Currency. demand deposits, time deposits, savings deposits, and foreign currency deposits (IFS, line 35). Domestic credit is calculated as the difference of the mnonetary base (IFS, line 14) and the central bank's net foreigrL assets (IFS, line I I minus line 16c). Chanages in net foreign assets and domestic credit are expressed as a percentage of the monetary base. The multipher is defined as the ratio of M42 and the monetary base. Central government balance is expressed as a percentage of GDP Iaflation is based on consumer prices (IFS, line 64). a. Denott s the period of equal length immediately preceding the capital inflow stage. Figures are averages of annual figures. Source, ilv[, Worldconomic Outlook data base; i4F, Internartional Finaucial Statsties data base 222 CHALLENGES OF MACROECONOMIC MANA W Annex 4.2 Can Sterilization Be Effective? W HE HETHER STERILIZATION CAN WORK this experience is that external creditors wished to essentially depends on relative substi- acquire equity in Mexican firms, while the steriliza- tutability among assets, both between tion instrument issued by the central bank consisted domestic and foreign interest-bearing assets and of claims on the Mexican government. In this case, among different types of domestic assets. If domes- the maintenance of portfolio equilibrium in the face tic interest-bearing assets are perfect substitutes of increased external demand for Mexican equities among themselves, then the issue of whether steril- would have required higher equity prices in Mexico, ization can work amounts to whether it is possible even if the sterilization of capital inflows had been to prevent a change in demand for domestic inter- complete. est-bearing assets from causing a change in their Because sterilization involves the issuance of ad- price through a suitable quantity response. Clearly, ditional domestic debt, and because its effectiveness this will depend on the degree of substitutability depends on substitutability among domestic assets, between domestic and foreign interest-bearing the form that it takes may matter for both its effec- assets. If they are close to perfect substitutes, then tiveness and its desirability. Capital-importing the quantity response would have to approach countries have used three alternative sterilization infinity to prevent a price response. This is what is techniques: typically meant by the impossibility of sterilization. T - . , . , * ~~~~~~~* transferring public sector deposits out of the However, it is also possible for sterilization to be commercial banking system and into the cen- ineffective in preventing a change in the price of do- tral bank mestic assets even if foreign and domestic interest- bearing assets are imperfect substitutes. This is the seokecd case when domestic assets are imperfect substitutes markets among themselves and a capital inflow represents an ticreas. increased demand for a particular type of domestic tic banks. asset that the central bank cannot provide, either di- While these techniques share the common objec- rectly or indirectly. In this case, sterilized interven- tive of stabilizing the domestic money supply, the tion that keeps the monetary base constant, by analysis below suggests that their macroeconomic issuing an asset other than that demanded by the effects may be quite different. agents generating the capital inflow, could not pre- Sterilization through transfers of public sector deposits. vent relative price adjustments among domestic as- Consider the effect of a capital inflow triggered by sets as portfolio equilibrium is restored. If such price a shift in private sector (domestic or foreign) pref- adjustments cannot be avoided, then it may not be erences from foreign to domestic interest-bearing feasible to insulate aggregate demand. For example, assets. To effect this portfolio reallocation, the pri- aggressive sterilization in Mexico did not prevent vate sector has to sell foreign exchange to the the arrival of capital inflows from being associated domestic central bank. When the public sector off- with a stock market boom. One interpretation of sets a purchase of foreign exchange by transferring 223 tSAtBCAPITAL FLOWS TO DEVELOPING COUNTRIES public sector deposits from commercial banks to increasing the stock of outstanding domestic pub- the central bank, it leaves the stock of base money lic sector debt. The amount cf new debt is there- unchanged but exchanges a claim on the domestic fore equal to the increase in dlemand for interest- banking system for an external claim. At the same bearing claims on the domestic economy. The time, the private sector changes its portfolio in the effect of the transaction on the central bank's bal- opposite direction. ance sheet is to leave its liabilities (the base) There are two ways that the macroeconomic unchanged, but to change the composition of its equilibriuni can be affected by this transaction, even assets, reducing its claims on the domestic govern- if the monetary base is unchanged. First, if interest- ment and increasing its international reserves. bearing deposits in the domestic banking system are From the standpoint of the nonfinancial public imperfect substitutes in private portfolios for other sector as a whole, sterilized intervention amounts domestic interest-bearing assets, then private port- to a portfolio transaction in which the domestic folio equilibrium will be disturbed unless the do- nonfinancial public sector issues interest-bearing mestic asset for which there is increased demand debt denominated in domestic currency in order to happens to be deposits in the domestic banking sys- acquire a foreign interest-bearing claim. tem. If that is the case, then the private sector and Sterilized intervention through open market op- the government simply exchange claims on the erations is not likely to be costless, however. The banking system and there are no price implications portfolio reallocation implied for the public sector to the transaction. But if it is not the case, then rel- involves issuing a high-yielding liability (domestic ative domestic asset prices will have to change to currency debt) in exchange for a lower-yielding asset maintain portfolio equilibrium. If the initial asset (international reserves). This interest differential shift was toward domestic securities, for instance, leaves the public sector in a weakened financial po- the yield on such securities would presumably have sition.2" The net adverse effect of such transactions to fall and interest rates on bank assets and liabilities on the public sector's solvency is overstated by the would have to rise. interest differential, however, because this differen- The second effect is a fiscal one. To the extent tial presumably exists in part to compensate credi- that the yield on domestic deposits differs from the tors for the currency and country risk associated yield on foreign exchange reserves, the solvency of with holding domestic public debt. By issuing such the domesi:c public sector will be affected by the debt in exchange for reserves, the public sector is re- transaction. In the particular case of sterilization ceiving a benefit that partly offsets the interest through shifts in public sector deposits, the liquid- penalty-that is, the option to reduce the real value ity services provided by such deposits suggest that of its obligations by devaluing or defaulting. the public sector's net interest receipts could rise or Nonetheless, these benefits may be worthless to a fall as a resalt of the sterilization operation. government that never intends to exercise these op- Sterilization through open market operations. Open tions (but is unable to convince its creditors of this market sterilized intervention requires the central fact). In this case, sterilized intervention carries a fis- bank to sell enough domestic bonds to purchase cal burden, the cost of which depends on the ease the foreign exchange associated with the inflow, with which an offsetting fiscal adjustment can be thereby leaving the monetary base unchanged but effected. 224 CHALLENGES OF MACROECONOMIC MANAG1 As in the previous case, the conditions required However, bank borrowers squeezed out of credit for sterilization through open market operations to markets by higher reserve requirements may not be effective in insulating the domestic economy are have access to securities markets and thus may be that domestic interest-bearing assets must be perfect unable to supply the assets that are in higher de- substitutes among themselves or, if they are not, that mand. An extensive literature on precisely this prob- sterilization operations must be capable of supplying lem for the United States claims that because of precisely those assets that are in increased demand. asymmetric information, borrowers with low net Sterilization through restrictions on domestic credit worth who lack other forms of collateral are able to growth. Sterilized intervention through transfers of obtain credit only from institutions that are highly public sector deposits and open market bond sales specialized in evaluating and monitoring loans operates by fixing the size of the monetary base. An (banks). Such borrowers cannot securitize their lia- alternative is to allow the base to expand as a result bilities. If this is the case, then there is a prima facie of central bank intervention in the foreign exchange case for imperfect substitutability among the rele- market, but to restrict expansion of the money sup- vant domestic assets when sterilization is pursued ply by raising reserve requirements on banks, thus through a policy of altering reserve requirements, causing the money multiplier to contract. suggesting that insulation of the domestic economy Again, the objective is to fix the price of domes- is less likely to be achieved under this policy than tic interest-bearing assets in the face of an increase through open market operations.22 in the demand for such assets. From a portfolio al- An additional difference between sterilization location perspective, this policy works, in principle, through restrictions on domestic credit and through by having the private sector rather than the central open market operations is fiscal cost. In the case of bank issue the domestic interest-bearing assets that domestic credit restriction, the public sector still ac- are in increased demand. When commercial banks quires interest-bearing foreign assets but does so by face higher reserve requirements, they are forced to emitting noninterest-bearing liabilities (that is, the absorb the additional monetary base emitted by the monetary base). Thus, the public sector's fiscal situ- central bank in the course of its foreign exchange ation actually improves when sterilization is effected operations, rather than acquire domestic interest- by increasing reserve requirements. This approach, bearing assets in the form of credit extended to do- however, implicitly taxes private agents. The re- mestic agents. The contraction of credit on the part quirement that banks hold a larger stock of nonin- of commercial banks causes domestic agents who terest-bearing reserves imposes an explicit tax on would otherwise have borrowed from these banks to them, which must be apportioned in some way instead issue securities. Since these securities are among bank depositors, borrowers, and sharehold- precisely the assets that are in increased demand by ers. The incidence of the tax will depend on the the nonbank public, the supply of domestic inter- elasticity of supply of bank deposits, as well as on est-bearing assets expands to meet demand, so there the demand for bank loans. is no change in the price of such assets. 225 .P>PITAL FLOWS TO DEVELOPING COUNTRIES Notes 1. Costa RIca has been receiving large amounts of cap- depreciated than during the preceding six-year period, as ital relative ro the size of its economy at least since the indicated in table 4.4. mid- 1970s. 11. The debt-export ratio can be interpreted as an in- 2. In principle, causation can run both ways between dicator of past fiscal policies as well as of future fiscal private and official flows constraints. 3. Econometric tests conducted for Chile and Indone- 12. In contrast to the issues of overheating and vulner- sia as part of this study are consistent with essentially full ability, the management of capital account volatility sterilization of changes in net foreign assets during the re- cannot readily be analyzed by reviewing country expern- cent capital inflow episodes in these countnes. ence-at least not the experience of developing countries, which has been dominated by ore-way flows associated 4. Several countries exhibit a pattern in which with the process of integration itself (inflows) and the re- growth of the base accelerates sharply when inflows cent Mexican cnsis (outflows). I'hus this section relies begin to arrive, then decelerates as sterilization is imple- heavily on analytical principles and related empirical work. mented Examples are Chile in 1990-92, Indonesia in 1990-91, K.orea in 1991-92, and the Philippines in 13. Strictly speaking, this is a domestic rather than an 1991-92. external financial shock, but it has an important external dimension. 5. In Chile, while the increase in net private inflows was offset ori average by a decline in official flows, im- 14. The lack of fiscal flexibility .ias been cited as an im- provement in the current account balance led to a large portant motivation for the adoption of capital controls in overall balance-of-payments surplus and hence upward Chile, whose central bank has simultaneously pursued in- pressure on the money supply. flation and real exchange rate targets since 1985. 6. More recently, there may have been some buildup 15. The main advantage of a currency board is its pos- of inflationary pressures in those countries that received itive effect on the credibility of the government's commit- the most flows in East Asia: Indonesia, Malaysia, and ment to low inflation (for a more detailed discussion see Thailand. annex 5.4 in chapter 5). 7. This abstracts from changes in net factor income 16. This does not rule out using the exchange rate as a and net current transfers. nominal anchor for stabilization purposes in a world of high capital mobility Doing so without restricting capital 8. While it may be natural to interpret causation as movements, however, requires that fiscal policy be capa- running frorn investment to growth, the two other logical ble of adjusting the fundamentals that determine the alternatives (from growth to investment and both reflect- equilibrium real exchange rate in such a way as to cause ing the influence of a third variable) are also plausible. the equilibrium real exchange rate to track the actual rate implied by the nominal exchange rate rule and any re- 9. While it may seem odd to include Mexico, recall maining inertial inflation. that the objective is to identify the fundamentals associ- ated with vulnerability, not to determine spillover effects. 17. The fiscal cost of sterilization arises from the gap between the interest rate on domestic public sector (gov- 10. While Chile experienced real appreciation during emient or central bank) liabilities and that on reserve as- 1988-93, its real exchange rate during 1989-94 was more sets. This has been estimated at about 1 percent of GDP for 226 CHALLENGES OF MACROECONOMIC MAA Chile, Colombia, and Indonesia during periods of heavy countries have not faced such severe credibility problems sterilization (Peak costs of sterilization have also ap- in recent years. proached this value in Malaysia.) 20. While the possibility is not explicitly examined 18. Instances of procyclical fiscal policy are most evi- here, it is also likely that a policy mix emphasizing tight fis- dent in countries subject to long terms-of-trade shocks. cal policy would have exerted feedback effects that would Nigeria and Venezuela, two countries in which oil rev- moderate the pace of capital inflows by keeping domestic enues loom large in public sector budgets, provide vivid interest rates lower than they would have been otherwise. illustration of the problem. 21. Calvo (1990) has argued that the weakening of the 19. When credibility problems become extreme, the government's financial position creates an incentive for it government's financing problems may become severe to inflate. The extent to which this is so depends on the enough to warrant the adoption of institutional devices currency denomination of the debt as well as on the prop- that will improve credibility but may greatly circumscribe erties of the government's loss function. the government's freedom of action-that is, adopting a currency board or joining a currency union. The position 22. For evidence on similar segmentation in Latin taken previously in this chapter that such arrangements American credit markets, see Rodriguez (1994). Chinn and are not necessarily mandated in a highly financially inte- Dooley (1995) provide similar evidence of segmentation in grated environment reflects the judgment that most the bank credit markets of East Asian developing countries. 227 PTER 5 The Effects of Integration on Domestic Financial Systems HE BANKING SYSTEM PLAYS A LEADING ROLE IN THE process of financial integration and is one of the main channels through which the benefits of inte- gration materialize. This is so because banks domi- nate financial intermediation in developing coun- tries (table 5.1) and therefore they end up directly or indirectly intermediating a large proportion of private flows. Since an increasing share of private flows is being used by the private-as opposed to the public-sector in recipient countries, banks are increasingly responsible for its allocation. In addition, as discussed in chapter 3, in the medium and long term, banks in developing coun- tries can benefit from financial integration by adopting more advanced financial technologies, achieving greater diversification in their portfolios, having access to a larger supply of funds, and realiz- ing efficiency gains derived from economies of scale and scope and a higher degree of market competition. However, the transition toward greater financial integration also in- volves risks for the economy in general (chapter 1) and the banking sec- tor in particular. During the process of financial integration banks will be adversely affected by increased macroeconomic volatility and by structural changes in banking. The main structural changes affecting the banking sector that result from integration are an increase in com- petition, which can erode banks' worth, and exposure to new sources of risk that banks may not be prepared to manage properly. 1 Because con- ditions can change more swiftly and markets can react faster and with increased amounts of funds, the room for maneuver available to policy- makers is significantly reduced in an integrated environment. The weaker the initial conditions in the banking sector-and in the macro- 229 F'APITAI FLOWS TO DEVELOPING COUNTRIES Table 5.1 liidicators of Financial Intermediation, Selected Countries Bank Percentage of GDP interimediation Securities autstandinga Deposit maney bank assets ratio, 1990 1994 1990 1994 1994b Asia India 35.0 93.4 41.7 45.3 80 Indonesia 13.5 32.6 62.3 64.9 91 Korea 91.4 105.4 64.6 72.3 38 Malaysia 196.5 352.6 95.8 104.6 64 Taiwan (China) 83.2 134. 148.5 196 0 80 Thailand 37.8 113.8 78.3 106.8 75 Latin Amne.ica Argentina 9.4 30.4c 29.3 26.1 98 Brazil 11.9 66.6 68.2 64.0 97 Chile 54.5 155.8 52.2 52.8 62 Colombia 22.8 34.4 23.7 27.8 86 Mexico 44,6 73 9 30.8 47.1 87 Venezuela 22,4 18.2 29.3 27.0 92 IndustriaInations Germanyd 88.9 132.7 149-1 152.0 77 Japan 189-5 178.2 164.1 156.8 47 United States 203.7 244.5 87.0 74.4 23 a. This mm hides, where available, shors-ternimoney market instrunents, government bonds, corporate bonds, and equities at market value. b. This is ihe ratio of the banking sector's assets to the assets ofail fmancial institutions. Nonbank financial instosutions are broadly defined to isiclude insurance companies, investment flinds, finance companies, and so on The definitions, however, tend to vary somewhat from country to country. c. This figiare is for 1993, d The banking sector in Germany is larger than in other counitries because of the universal banking system adopted in thLs country. Source: Bas (1996); IMF, InternationalFinancial Staustics data base; World Bank data. economy-are in integrating economies, the greater the challenge for policymakers. One of these challenges arises because during integration banks are given easier access to funds and are therefore able to expand lending much more quickly, so they can finance a longer-lasting boom in expenditures and asset prices. If initial conditions in the banking sec- tor are weak, banks are likely to extend credit in excess and to more risky sectors. The process of financial integration is usually accompanied in its early stages by a surge in private flows that may in crease bank lending and exacerbate macroeconomic and financial sector vulnerability. The surge in private inflows is partly due to improved economic prospects 230 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANC1IVI. in the recipient country, which, in turn, will increase economic agents' expectations. The inflows also often finance a rapid expansion in bank lending and aggregate expenditures, accelerating economic growth and validating agents' expectations, and cause an increase in asset prices. These effects can reinforce one another, leading to additional borrow- ing, higher expectations and output growth, and increased levels of ex- penditures and asset prices (chapter 1). In developing countries with weak infrastructure and regulatory institutions in banking-a common condition in many emerging economies-the surge in bank lending and the rise in asset prices will be associated with a deterioration in banks' portfolios. Poorly managed and supervised banks will tend to in- vest in highly profitable although risky activities. For example, poorly regulated banks may finance consumption booms and speculative ac- tivities, such as a boom in construction and real estate, that increase macroeconomic vulnerability. Also, because of poorly diversified port- folios and lack of adequate provisioning, the surge in bank lending usu- ally exacerbates financial sector vulnerabilities. This chain of events is potentially costly not only because of resource misallocation but also because in the extreme it can lead to financial distress and crisis. An integrating economy is more likely than an autarkic one to experi- ence a boom-bust cycle in bank lending, and the potential cost associated with such a course is higher This is so because, as argued earlier, in an in- tegrated environment banks can expand lending more quickly and in larger amounts, circumstances can change more swiftly, and markets react faster. All these factors can lead to swift changes in market senti- ment and cause large reversals of flows, creating significant macroeco- nomic and financial sector distress and carrying high economic and social costs. In addition to the direct costs of bailing out failed banks, the sudden loss of liquidity in the banking sector, which is more difficult to contain in an integrated environment, can amplify economic downturns. Such distress can set back economic reforms severely, as was the case with several Latin American countries following the debt crisis of 1982. An important challenge, therefore, is for a country to maintain or improve the health of its banking system. Doing so will determine to a large extent whether it is able to realize the benefits of financial inte- gration and avoid its pitfalls. To distill lessons for financial sector reform in the face of growing in- tegration, this chapter looks at the role that banking systems have played in the process of financial integration, and how they have, in turn, been 231 CAPITAL FLOWS TO DEVELOPING COUNTRIES affected by this process. It draws on the experience of a wide range of country episodes during the 1980s and 1990s, periods when these coun- tries received substantial capital inflows as a result of integration. The episodes studied here are historical and do not reflect current country conditions. The sample analyzed here differs from that used in chapter 4 in that it includes both industrial and developing countries, and in the case of some developing countries more than one episode is studied.2 The purpose of using this sample is to highlight the role of the banking sector in the process of financial integration. A detailed description of the sample of country episodes is provided in annex 5.1. A number of important conclusions emerge from this wide-ranging country experience: * First, countries that received substantial capiLal inflows as part of the process of financial integration also typically experienced a lending boom, reflecting higher levels of direct and indirect in- termediation of capital flows by the banking system. * Second, countries in which the lending booms were greater and in which no compensatory macroeconomic policies were imple- mented typically experienced a significant increase in macroeco- nomic vulnerability, as indicated by a widening of the current account deficit above what one would expect on the basis of the size of the inflows, relatively large consumption booms, and rela- tively small investment booms-relative to the size of the inflows. * Third, lending booms were often associated with an increase in financial sector vulnerability despite the fact that booms tended to improve bank profitability in the short term. * Fourth, countries in which lending booms were associated with an increase in both macroeconomic and financial sector vulnera- bilities typically experienced banking crises. * Fifth, countries that took actions to restrain the lending boom, or pursued prudent macroeconomic policies, fared well and avoided major crises. Based on these findings, the chapter draws the following main pol- icy conclusions: * Extremely weak initial conditions in both the macroeconomy and the financial sector may warrant a cautious approach toward external financial liberalization. 232 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL.' * Financial integration makes it imperative for countries to move aggressively to create a prudent incentive and institutional frame- work for their banking systems. Given the larger volume of funds that is intermediated during the early stages of financial integra- tion, and the greater risks of vulnerability from the increased magnitude and speed of market reaction, countries should move forward in implementing policies aimed at strengthening the su- pervision and prudential regulation of banks. * A strong macroeconomic stance, and especially a strong fiscal po- sition, can help prevent or postpone financial sector distress. * Because of the high costs of a banking crisis (box 5.1), and the time needed to improve the macroeconomic position and strengthen the banking system, a strong case can be made for containing lending booms associated with the early stages of fi- nancial integration. This would help mitigate an increase in macroeconomic and financial sector vulnerabilities associated with large lending booms, especially when macroeconomic and financial sector conditions are weak. = Finally, given the increased susceptibility to banking crises, coun- tries should put in place mechanisms that enable policymakers to deal with such crises promptly and effectively. Delaying the ac- tions necessary to contain a crisis will only increase its costs. The rest of the chapter is organized in three sections. We first ex- amine the extent to which the surge in private capital flows following the opening of the capital account causes an increase in bank lending, and how the lending boom exacerbates macroeconomic vulnerability. We then analyze the reasons for increased vulnerability of the banking sector following a surge in capital flows and the subsequent lending boom. Next, we integrate the analysis of the two previous sections by comparing the changes in macroeconomic and financial sector vulner- abilities in countries that have experienced banking crises with those that have not, and summarize the major policy lessons relevant for countries undergoing financial sector reforms along with integration. These lessons extend previous work on financial sector reform by adding the factor of integration, which makes financial sector reform significantly more complicated. The discussion ends with an analysis of the most effective measures for managing a banking crisis in an in- tegrated environment. 233 W-ItCAPITAL FLOWS TO DEVELOPING COUNTRIES Box 5.1 The Costs of Banking Crises POLICYMVAKERS MUST BE CONCERNED ABOUT crisis may result from a decrease in economic activ- banking crises because these crises are costly in terms ity caused by other factors, with the banking sector of lowes economic growth, and the political conse- playing a key role in amplifying the effect of the ex- quences of allocating losses among bank stockhold- ogenous shock on aggregate output. Figure A shows ers, creditors, depositors, and taxpayers can be the average output growth for the precrisis, crisis, significant. It is important to note, however, that the and postcrisis years for country episodes where a direct economic cost of a banking crisis derives from banking crisis occurred. Although the decline in the decrease in economnic activity and growth that re- economic growth overstates the cost of banking sults frem a reduction in the total volume-or a crises, the magnitude of the dec]ine suggests that more inefficient intermediation-of loanable funds, these costs are significant.2 The country episodes while the allocation of losses has primarily distribu- shown in the figure experienced, an average, a posi- tion effects (usually very important) but on its own tive growth of about 5 percent a year in the precrisis does nol imply an economic (deadweight) loss. Nev- period, and slightly negative economic growth in the ertheless, the allocation of losses is important because crisis and postcrisis years, implying a decrease in it changes the incentives of the different groups of output growth of about 5.2 percent per year. economic agents and, through this, can lead to lower The allocation of losses, often used as a measure investment and growth. For example, forcing the of the cost of banking crises, is alsco quite significant. 'corporate sector to repay its past-due loans, although Figure B shows the cost of restracturing banking it may be fair practice (depending on the primary systems after the crisis, measured by loans from the cause of the sector's financial difficulties), may be central bank to commercial banks and other rescu- counterproductive for the purpose of facilitating the ing measures, for some of the countries in our sam- economic recovery.1 Indeed, in the presence of a ple. The average cost for the episodes shown is about debt overhang problem entrepreneurs may decide 10 percent of GDP and 23 percent of total loans. not to start new-or to discontinue old-investment These figures appear strikingly higi when compared projects, because from a private point of view it may with the cost of the savings and loan restructuring not be vworth investing in projects whose profits will program in the United States, wlhich by 1991 had benefit the firm's creditors (banks). Similarly, allow- reached 5.1 and 7.8 percent of GDP and total loans, ing bariks to recover their losses by increasing respectively (Rojas-Suarez and W,isbrod 1996; see spreads--thereby taxing depositors and debtors in also Goldstein and Turner 1996). good stsnding-will reduce the volume of finds in- termediated through the banking system and may 1. If the ultimate cause of the banlkng crisis is the im- lead to lower investment and growth. plementation of "wrong" economic policies by the author- Quantifying the economic losses associated with ities, then it can be argued that the corporate sector is not a banking crisis is a cumbersome task, mainly be- fully responsible for defaulting on its debt. cause it is not clear how to separate the decrease in 2. In addition to overstating the cost ofcrises because of economic a due tO the banking crisis per se the difficulty in separatng other exogenous factors that re- economic activity due to the banking crisis per se duce output growth, the figure overstates these costs by im- from that which is due to other exogenous factors plicirly assuming that growdt in the precrisis (boom) (such as a terms of trade shock). In fact, a banking period is normal. 234 THE EFFECTS OF INTEGRATION ON DOMEST1C FINANCMI3 Box Fiure S.1 Cost of Bankdng Cises A. Banking Crises and Economic Growth Annual percentage m Po6tclsis yeats 4 2 -2 -4 S~~~~~~~~~~ : -2 t tS51| v*S t E iv v*4 SSource- I4mp, International FnanczilStatistiss data base, iMP, World Economic O,atlook dara base. B. Cost of Restructurlng the Banking System Percentage of QDP Percentage of total loans 20 ~~~~~~~~~~~~~~~~~~~~~~~so Total cost/GDP s 15 R- Total cost/total loans ±0~~~~~~~~~~~~~~~~~~~~~~4 0~~~~~~~~~~~~~~~~~~~~~~~2 ArgentIna Chile Finland Norway Sweden Venezuela 1980 1981 1.991 1988 1991 1994 Soure. Rojas-Sukrez and Wejsbrod (1996). 235 LW_K -CAPITAL FLOWS TO DEVELOPING COUNTRIES Bank Lending and Macroeconomic Vulnerabilit T HIS SECTION FOCUSES ON THE INTERACTION BETWEEN capital flows and bank lending. We explore the role of bank lending in propagating and amplifying the effects of interna- tional financial integration on macroeconomic variables. In particular, we focus on how bank lending can make the economy vulnerable to shocks that can affect market sentiment toward a country and ulti- mately trigger a banking-and sometimes a balance-of-payments- crisis. We begin with a general discussion of the role of banks in the macroeconomy in which we highlight the interaction between bank credit and capital flows. Next, the section looks at the problems a number of countries have experienced as they become more financial- ly integrated. The Role of Banks in the Macroeconomy Financial intermediaries play a crucial role in the process of economic development and growth by channeling savers' inoney to more pro- ductive uses in an economy; they do this by screerLing and selecting in- vestment projects and transforming assets, usually from short-term deposits to long-term investments, but more generally from illiquid to liquid assets.3 The fact that banks overcome informational problems between lenders and borrowers and facilitate the creation of liquidity makes them vulnerable to confidence crises and, therefore, susceptible to runs and liquidity crises (Diamond 1984, Diamond and Dybvig 1983, Ramakrishnan and Thakor 1984). There has been extensive research on the role of the banking sector in the macroeconomy and its importance in propagating and amplify- ing business cycles; it has led to the conclusion that bank credit, and not just monetary balances, affects macroeconorrmic performance-far beyond what standard macroeconomic models suggest.4 Past studies have asserted that the banking sector amplifies the mag- nitude of the business cycle because bank credit behaves procyclically. For example, a positive development in the real sector that raises ex- pectations about the future will increase firms' wilLingness to invest, in- duce them to borrow more, and cause an expansion in bank credit. This process, if accommodated by the monetary authorities, will lead 236 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL S 51 to a lending boom, which, in turn, will enable market participants to increase their level of expenditure and will accelerate economic growth. The increase in economic growth reinforces the expansion of lending and spending by validating economic agents' expectations about the fu- ture. The state of euphoria triggered by the initial increase in spending also leads to an increase in asset prices and financial wealth, which raises the value of loan collateral, increases households' aggregate con- sumption, and further reinforces the process. At the same time, how- ever, the economy as a whole becomes more vulnerable because of the increase in indebtedness of firms and households and the risk of a col- lapse in asset prices. A plunge in asset prices would reduce borrowers' financial wealth and thereby affect the financial health of banks. A shock that reduces economic growth, or that leads agents to believe that the current conditions are no longer sustainable, will induce agents to cut their spending on consumption and investment and thereby will slow economic growth. Firms and households will then have difficulty servicing their debts, banks will start calling loans back in and liquidat- ing the assets used as collateral, and asset prices will plunge, all of which will deepen the slowdown in economic activity. Because the banking system is vulnerable to confidence crises, this sequence of events may lead to a bank run, further reducing total credit and liquidity and ag- gravating the slowdown in economic activity.5 In addition to amplifying the magnitude of the business cycle, a poorly regulated and supervised banking industry will tend to misallo- cate resources, increasing the economic cost of the boom-bust cycle in bank lending. Poorly capitalized and regulated banks may, for instance, invest excessively in risky projects, such as real estate, or in other sectors prone to suffer boom-bust cycles. Also, poorly managed banks operat- ing under distorted incentives will not diversify their portfolios ade- quately, thus exacerbating financial sector vulnerability. These banks will suffer greater losses during the contractionary phase of the cycle. The weaker initial conditions in the banking sector are, the more vul- nerable countries will be to large downturns and costly banking crises. In an integrated environment these effects can become even more pronounced (for a review of recent developments in the literature see box 5.2). This is so for several reasons. First, integration gives banks ac- cess to a larger supply of funds to intermediate, either directly or indi- rectly, which allows them to increase credit rapidly.6 Second, financial integration usually occurs with the implementation of economic re- 237 WKPITAL FLOWS TO DEVELOPING COUNTRIES Box 5.2 International Financial Integiation and Boom-Bust Cycles DESPITE THE EXTENSIVE LITERATURE DOCUMENTING BOOM- bust cycles in bank lending and the associated fluctuations in eco- nomic activity, these issues have only recently been investigated in the context of the financial integration process. Ir general, the con- clusions that emerge from reviewing this literature appear to be straightforward. First, although the presence of foreign credit may result in larger cyclical fluctuations than could otherwise occur, it is the domestic banking sector that amplifies the magnitude of the macroeconomic cycle. And second, the domestic banking sector ex- acerbates the vulnerabilities in the emerging economies. For example, in a recent paper Sachs, Tornell, andVelasco (1996) conclude that the impact of the Mexican peso crisis of December 1994 on other emerging economies-the so-called tequila effect- can be partly explained by the level of private sector debt held by the countries' banking systems. In other words, developing countries in which bank credit was growing rapidly during the surge in private flows in the early 1990s were more vulnerable to the shocks of the Mexican crisis. Similarly, recent work by Kamirsky and Reinhart (1996) and Hausmann and Gavin (1995) provides empirical evi- dence that the boom-bust cycles in domestic bank credit, asset prices, and economic activity resemble and precede those in the ex- ternal accounts. This confirms that in financially integrated economies the domestic banking sector plays an important role in amplifying cyclical swings. Finally, Goldfajn and Valdes (1996) use a theoretical model to show that in a financially irntegrated economy the existence of a domestic banking system exacerbates capital move- ments from abroad and, therefore, amplifies the magnitude of exter- nal shocks. forms-in the financial sector, the real economy (trade, fiscal), or both-that improve the country's economic prospects and raise agents' expectations. Third, integration increases the sources of risk and the speed of market reaction. The first two reasons above, the implementation of economic re- forms and the easier access to funds, create the necessary conditions for a lending boom to start. In particular, the surge in private flows that usually accompanies financial integration provides the additional re- sources needed to finance an increase in aggregate demand and output, 238 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANC_IAWI validating agents' expectations. Given the size of the capital inflows for the average recipient country observed in recent years, the potential for experiencing a large increase in asset prices and a lending boom-one lasting several years-is faster in an integrated environment. The third reason above, the increased sources of risk and speed of market reaction, makes it more likely that a banking industry operating under suboptimal conditions-a common occurrence among develop- ing countries-will misallocate resources, exacerbating the country's macroeconomic and financial sector vulnerabilities. In particular, if an increasing amount of funds flowing into a country is intermediated by an undercapitalized and poorly regulated banking sector, one that is too prone to risk taking, the economy is likely to end up investing ei- ther insufficiently or in risky projects. This could be the case if the banking sector finances a consumption and imports boom, which in turn will make the economy more vulnerable to shocks. Because of the additional sources of risks and faster market response, the likelihood and the cost of a banking crisis increase with integration. Capital Inflows, Bank Credit, and Macroeconomic Vulnerability: Country Experiences A lending boom has usually accompanied a surge in private capital flows in countries that have recently undergone financial integration. Figure 5.1 shows the ratio of bank lending to the private sector as a share of GDP, in the years prior to and during the surge in capital in- flows for a numnber of countries in the 1980s and 1990s. In all except two episodes in the sample (Chile and Venezuela in the 1990s), the share of bank lending to GDP was higher in the inflow periods than in the years prior to the inflow. For all countries and all episodes taken to- gether, the average lending-to-GDP ratio during the years prior to the inflow surge was 29.4 percent. During the inflow periods, the average lending-to-GDP ratio was 40.9 percent. As the analysis in the previous section suggests, countries in which the banking sector intermediates proportionately larger inflows will probably become more vulnerable to macroeconomic shocks. To illus- trate this issue, we can look at the experience of countries that in recent years have received significant capital inflows while becoming more in- tegrated. For all the countries in our sample the evidence suggests that the current account deficit, on average, increased during periods of 239 C TAI FLOWS TO DEVELOPING COUNTRIES Figure 5.1 [lank Lending to the Private Sector during Inflow and Pre-inflow Periods, Selected Countries and Years (as a percenztage of GDP) Argentina, 1'J79-82 Argentina, 1992-93 Brazil, 19I92-94 _ *Inflow period Chile, 1978-81 Chile, 1989-94 Three years prior Colombia, 1992-94 Finland, 1"87-94 Indonesia, 1990-94 Malaysia, 1"80-86 Malaysia, 1989-94 - Mexico, 1179-81 Mexico, 1'89-94 Norway, 19184-89 Philippines, 1'78-83 Philippines, 1189-94 Sweden, 1'89-93 _ Thailand, 1978-84 Thailand, 1988-94 __ ___ Venezuela, 1975-80 Venezuela, 1992-93 0 10 20 30 40 50 60 70) 80 90 Source IMF, internatzonal Financial Statistics data base During periods of large capital high capital inflows, and that the increase was, in general, proportion- inflows, bank lending to the private ally larger in those country episodes where bank and nonbank credit to sector has surged. the private sector grew more during the inflow period. Conversely, the current account deficit was proportionally smaller. given the size of the capital inflow, in those country episodes that saw a smaller expansion in bank and nonbank credit (box 5.3 ). While several policies can affect the level of the real exchange rate (chapter 4), the country episodes in our sample that show the greatest real exchange rate appreciation are also, in general, the ones in which bank credit grew more.7 The banking sector plays a leading role in the adlocation of loanable funds among economic sectors, and is partly responsible for the over- 240 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL& heating and increased macroeconomic vulnerability that may result from the surge in private capital flows. The challenge for policymakers, then, is to control the effects of private capital inflows, which can finance an in- crease in either aggregate investment or consumption, or both. Although an increase in investment does not guarantee a higher rate of economic growth, it is certainly preferable to an increase in aggregate consumption, which is more likely to lead to an overheating of the economy and, in the extreme, can affect the market sentiment toward the country. In countries that have recently become more financially integrated, inflows of foreign capital have been used to finance an increase in both investment and private consumption. There was a positive association between investment and capital inflows, and in a number of the episodes in our sample, the increase in investment was smaller than that predicted by the size of the inflows. This result usually occurred in those episodes where the increase in bank and nonbank lending was among the largest. Similarly, the increase in private consumption was also proportionally larger, given the size of the capital inflow, among the episodes that had the largest increase in bank and nonbank lending (box 5.4). It is important to note that the majority of the episodes where bank and nonbank lending grew the most ended in banking crises. The conclusion to be drawn here is that one of the ways in which a poor bank intermediation process can exacerbate macroeconomic vulnerability, in the context of increased financial integration, is by biasing the increase in aggregate expenditures that would result from an increase in capital inflows toward consumption instead of investment. The discussion above suggests several things: * First, countries experiencing a surge in private capital inflows often also experience a lending boom, but the magnitude of the boom varies significantly among them. * Second, countries experiencing the highest increase in bank and non- bank lending are usually those in which macroeconomic vulnerabili- ties-measured by the increase in the current account deficit, excess consumption, and underinvestment-are exacerbated the most. * Third, countries with the largest lending booms during an inflow episode usually later experience a banking crisis.8 This implies that a rapid growth in bank lending should be avoided, or at least watched closely by the economic authorities in emerging coun- tries, especially when conditions in the banking sector are weak. 241 _W1ITL FLOWS TO DEVELOPING COUNTRIES Box 53 Excessive Bank Lending Tends to Increase the Current Account Deficit THE EXPERIENCE OF THE COUNTRY EPISODES IN To summarize, figure B seems to indicate that our samrple shows that, as expected, the current ac- the countries that exhibited the largest current ac- count deficit, on average, increased during periods count deficits following the surge in inflows were of high capital inflows (figure A). More important, those that experienced the largest increases in bank figure B below shows that, in general, the increase in lending proportional to GDP. In particular, for the the cuirent account deficit was proportionally sample of country episodes shown in the figures it larger, given the size of the capital inflow, in those follows that an increase in private capital flows of, country episodes where bank and nonhank credit to say, 10 percentage points of GDP ]Leads to a worsen- the private sector grew more during the inflow pe- ing in the current account balance of about 3.2 per- riod (the measure of "proportionality" is indicated centage points ofCDP, on average. Nevertheless, the by the solid line in figure B). Countries such as latter increases to about 4.9 percentage points of Chile (1978-81), Malaysia (1980-86), Mexico GDP (an increase of about 50 percent) for those (1979-,31 and 1989-94), the Philippines (1978-83 countries that experienced an increase in bank lend- and 1989-94), and Sweden (1989-93) exhibited ing above the sample average. This result follows larger current account deficits than expected, given from the following regression: the size of the inflows they received. These countries d 1.91 + 0.324*kafrw+ 0.I 7*(kajow*-dbank) also experienced a larger increase in bank lending s 2191 (2.43) (1.46) than the remaining country episodes in the sample. The average excess deficit ratio (that is, the ratio of where cadis the current account deficit (as a share of the acttual current account deficit to that predicted GDP), kaflow is the share of private flows to GDP, by the size of the inflows) for the country episodes dbank is a dummy that equals I for those country in- mentioned above was about 1.3, and the average flow episodes where the increase in bank lending ex- growth in bank lending was more than four times ceeded the sample average (t-statistics are in the growth in GDP. For the rest of the country parentheses; note that the coefficient on the interac- episodes shown in figure B the average deficit ratio tive term is significant at 10 percent using a one-tail was 0.74, while the average lending growth was only test). two and a half times the growth in GDP. 242 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANIA - Box Figure 5.3 Capital Inflows and the Cunent Account A. Private Capital Inflows and the Current Account Balance Cuirrent account balance (as a percentage of GDP) i * Brazil 0 nesia Ar~~~W1tin5, 1979-829-9 | - 2 sColombla- Che,198994 - Sweden a Vendz---- - nla780 -43- - . F.intan.. - -9- _ - _ - laysia,.1989-- . . Venezuela, 1992-93 Mexico, 19 . - , , Mexico, i99441ip i __ _ .Philippines, 197-83 - 8 --. . . .. _ . _ ~ ... ff . _ . . . . .._ ..... _ .. ... ..... Chile, 1978-81 U g.9 0 2 4 6 8 10 12 14 Private capital inflows Note.-The equation for the trend line is y -0.4643x -1.525. R2 = 0.4211. B. Excess Current Account (CA) Deficit and Lending Booms during Inflow Periods Bank and nonbank lending growth Excess current account deficita Is~~ ~ ~~ - . _ -4_ 14 - - - - Bank lending gowth 12-- Nonbank lendIng growth 1 -,-- Excess CA deficit ±0 J. 6 -- ._...--.---- --0.8 0- --.2 2--~~F Tl EA A t l g i gi 1'IL O ., _ . . 04 _ . e i . t_o .4 _ _, Note. Bank and nonbatlk lending growYth are divided by GvP growrth during the inflow period. a. The exces current account deficit IS uhe atauo of the actual deficit to the one predicted by the size of thecapital inflows. Source. World Bank staffestimates. 243 :CAPITAL FLOWS TO DEVELOPING COUNTRIES Box 5..4 Excessive Bank Lending Can Lead to Underinvestment and Oerconsumption THE EP[SODES IN OUR SAMPLE SHOW THAT PRI- figure (the magnitude of this increase is reduced by vate capital inflows have been used to finance an in- 4.6 percent of GDP) in those episcdes where the in- crease in both investment and private consumption crease in bank lending has been higher than average. as illustrated by the positive slope of the regression An increase in bank lending fillowing a capital lines in figures A and C below. Although the corre- inflow surge can exacerbate macroeconomic vulner- lation between investment and inflows was positive, ability by reducing potential invesltment and financ- however, the increase in investment was smaller ing a consumption boom. This was, in fact, than that predicted by the size of the inflows In a observed in the inflow episodes stu died in this chap- number of countries: Argentina (1979-82), Brazil ter. Figures C and D show that the increase in pri- (1992-94), Chile (1978-81), Finland (1987-94), vate consumption was proportionally larger, given and Mexico (1989-94). Furthermore, the increase the size ofthe inflows, in the group ofcountries that in investment appears to have been lower than it had the largest increase in bank and nonbank lend- could have been, given the size ofthe inflows, in pre- ing. Excess consumption is defined as the difference cisely those countries where the increase in bank and between actual consumption (as a share of GDP) and nonbank lending was among the largest. Underin- the consumption share predicted on the basis of the vestment, measured as the difference between the inflows received by each country. Argentina predicted and the actual investment in a regression (1992-93), Brazil (1992-94), Finland (1987-94), of the change of investment against inflow size, is Mexico (1989-94), Norway (1984-89), Sweden shown in figure B. For the episodes with the largest (1989-93), and Venezuela (1975- 80)-all episodes increase in bank lending, the underinvestment aver- that ended in a banking crisis-erhibited increases aged 2.7 percentage points of GDP, while the in- in consumption larger than those predicted by the crease in bank lending averaged about six times that size of their inflows. At the same time, the average of GDP. For the remaining countries and episodes in increase in bank lending in these countries (adjusted our sarrLple, these variables averaged -2.2 percent by GDP growth) was more than three times that ob- and 2.3 times, respectively. Moreover, for the served in the remaining countries combined. Thus episodes in our sample, regression analysis indicates for the episodes in the sample, the ernpirical evi- that an lmncrease in private capital inflows of, say, 10 dence verifies that increases in bank lending amplify percentage points of GDP leads to an increase in gross the positive effect of inflows on private consumption investment by almost the same amount. However, and, therefore, increase the vulnerability of the the increase in investment drops to about half this economy. 244 THE EFFECTS OF INTEGRATION ON DOMESTIC FIA4AL 4C 1Box Figre 54 Effects of Capita Inflows on 1nvestmn and Consumpton A. Capital Inflows and Changes in Investment Change in investment (as a percentage of GDP) 12 ... ..... . - Vneuea 0* 5Thailand, 1988-94 Vene ,Venezuela, 1975-80 Ea aysja,_1989m- 6 - ~~~~~PhilIppines 1989-94 - Venezuela, ±.992-9. ±999 4 - ------ _ - ---J,-_ _--!e........ ... .. ... .__- ----- - - Argentina, 1992-93 and, ±97884 2 _g iipplnes, 1978-83 Chile, 1978-8± 0 Mexi_o, 1989-94 Sweden in -2 - .Brazil _ ' nland - S Argentina, 1979"2 -4 - ------- 0 2 4 6 8 10 12 14 Private capital inflows Note: The equation for the trend line is y 0.8132x - 0.7138. R2 0.3762. Bank and nonbank lending growth Underinvestrment as a percentage of GDP 16 B .ank lending growth 8 14-- ._ Nonbank lending growth - -U- lJUnderinvestment 12- .,. . X . . ..... .. ... ~, .- 4 8 ~~~~~~~~~~~~~~~~~~~~~~2 0 -2 4- --4 * rlS | r S* t S C eC 00g g ¢ H S S e 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 2 ,! 14 It -t \5 2 v -i c 6 ii r ° Note. Bank and nonbank lending growth are divided by GDP growth during the inflow period Underinvestment is measured as the difference between the investment predicted by the size of the inflows and actual investment. 245 DA1?ITL FLOWS TO DEVELOPING COUNTRIES Box Figire SA Icontinued) C. Changes in Private Capital Inflows and Consumption (as a percentage of GDP) Change,- in consumption * Venezuela, 1992-93 I Vearezuela, 1975-8 Colombia Norway i a Ilaia W89-94 2 Lgn&~~~~S9~~-93 a * * Miexico, 1289-94 2 - mSweden , * o_ . Ghii~~~~~~~~~~~~~C'e' 1978-811 Indonesila * Chile, 1989-94 Thailand, 197S-84 Thalland, 1988-94 0 2 4 6 8 10 12 Changes In private capital Inflows Nofe: The equation for the trend nine isy= 0.3136x-05616. R2 0.103 D. Consumption and Lending Booms during Inflow Periods Bank and[ nonbank lending growth Excess consumption as ai percentage of GDP ME Bank lending growth - Nonbank lending growth 2 14- -U Excess consumption 12 4----- - -- 6 - w -s __t- 3--A4-4- s -@-84-tfi3 1+ -2 - I -~~~~~~~~~~- 0 10 0i 3 _ a~~~~ Co .2~~~~ .9 Note: Bink and nonbank lendiggroOvdi are divided by GDP growth during the inflow period. Excess consumption is the difference between actual consumption and that predicted by the size of the inflows. Source: WVorld Bank staffestimates. 246 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCA N C Two exceptions to these rules in our sample were Indonesia and Thailand during the early 1 990s; these countries went into their inflow episodes with a relatively low degree of initial monetization, have expe- rienced high economic growth, and have pursued prudent macroeco- nomic management. Thus they were able to avoid or delay difficulties in the financial sector (box 5.5). Capital inflows and lending booms may also exacerbate economic vulnerabilities and ultimately lead to crisis by inducing an excessive and unsustainable increase in asset prices. This is a common outcome in most country episodes, especially when the domestic and external fi- nancial sectors have been rapidly liberalized, and its most clear mani- Box 5.5 Bank Lending Booms aid Overheating: The Importance of Fiscal Policy TWO EXCEPTIONS TO THE BOOM-BUST CYCLE WERE INDONESIA (1990-94) and Thailand (1988-94). These two countries experi- enced a significant increase-20 percen-t or more of GDP-in bank and nonbank lending during the inflow period. Neither country has, however, suffered a deterioration in macroeconomic fundamentals or, until 1994, seen an overt banking crisis.' This can be attributed to at least three factors. First, both countries had a relatively low degree of monetization and shallow financial markets at the beginning of the inflow episode-private bank lending to GDP was below 30 percent prior to the inflows in Indonesia and about 45 percent in Thailand, while in Malaysia, for example, it was above 60 percent. Second, both countries grew at 7 percent or a higher rate on an annual basis during -the inflow period. Third, these countries have had careful macroeco- nomic management over time. A low degree of initial monetization implies that financial assets may increase proportionally more than aggregate output without necessarily causing an increase in aggregate spending. Similarly, careful macroeconomic management and rapid growth imply that bank credit can expand without creating the stan- dard symptoms of overheating. In fact, both countnes had a signifi- cant improvement in their fiscal stance during the inflow period, as measured by the government surplus, which increased by more than 2 and 6 percent of GDP between the pre-inflow and inflow periods in Indonesia and Thailand, respectively. 1. As discussed elsewhere, Indonesia (1990-94)-and more recently Thailand (1995-96)-have seen a worsening in financial sector vulnerability. 247 - A-PITAL FLOWS TO DEVELOPING COUNTRIES festation is through the prices of stocks and real estate. The rise in asset prices, which corresponds to a drop in the cost of funds, tends to exac- erbate vulnerabilities because it increases financial wealth and can thereby raise the level of households' indebtedness and consumption. Indeed, households often use their newly apprecial ed assets as collateral for new loans. Furthermore, the increase in asset prices can trigger bor- rowing for speculative purposes such as buying stocks and real estate. If the surge in prices proves to be unsustainable, whi ch will be the case if asset prices rise excessively-and as shown in figure 5.2 this has been a common outcome toward the end of several inflow periods-then the economy will need to adjust to a lower level of aggregate spending later on. Most important, the bursting of the asset price bubble will lead to a decline in the financial situation of banks as the stock of nonper- forming assets rises. This is so because the increase in interest rates that accompanies the drop in asset prices will cause Dverindebted house- holds and speculators to start defaulting on their debts. In addition, the value of assets used as collateral will be insufficient to cover the banks' losses. In sum, the plunge in asset prices will leave banks in a weaker fi- Figure 5.2 Stock Prices during Inflow Periods, Selected Countries (stocks' realprice index) 600 1,600 Chile 1978-81 1,400 500 - -------------------------- Venezuela /- Me xic 1,200 400 1,000 NChile 19894 300 200 4~~~~~~~~~~~~~~~~00 200 - -- ------ - -------- 800 200 Finland 0 The surge in asset prices associated with inflow periods often proves to be 0 200 unsustainable. 3 2 :1 0 1 2 3 4 5 6 7 Year with respect to start of inflow period Note The index for Finland, Mexico, and Sweden is shown on the left, the index for Chile during the 1980s and 1990s and for Venezuela is shown at the right Source IMF, International Financial Statistics data base 248 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIA IA nancial situation and may even lead to crisis.9 Capital inflows can also weaken the banking sector in other ways. These are discussed next. Bank Lending and Financial Sector Vulnerability T HE SURGE IN PRIVATE CAPITAL FLOWS THAT ACCOMPANIES financial integration in developing countries usually leads to an increase in monetization and to banks intermediating a larger volume of funds. The increased financial intermediation, espe- cially when it occurs over a short period of time, not only can exacer- bate macroeconomic vulnerability, as discussed in the previous sec- tion, but may also exacerbate microeconomic problems, especially when the banking system suffers from initial distortions and weak- nesses. These microeconomic problems are important for two reasons. First, banks could face financial distress and eventually a crisis, which in turn could mean the derailment of the entire macroeconomic and fiscal stabilization program (Velasco 1987). Second, financial fragili- ties may impose constraints on the implementation of other economic policies. For example, the weak financial conditions in the banking industry in Mexico during 1993-94 restrained the government from increasing domestic interest rates, a policy needed to prevent the loss of international reserves. Similarly, in 1994 the central bank in Venezuela had to expand liquidity to assist domestic banks despite rapid inflation and large foreign exchange losses. In this section we discuss the ways in which international financial integration has affected the banking industry in countries that have lib- eralized their capital accounts. First we present the conceptual frame- work for thinking about this issue and discuss the indicators used to evaluate the performance of the banking sector in countries receiving capital inflows. We then describe the different country episodes and fi- nally present our major conclusions from these country experiences. Lending Booms and Banking Sector Vulnerabilities A fast-growing banking industry will become more vulnerable if it is unable to evaluate the risks of increased lending (because of lack of trained personnel), and if its greater risk exposure is not accompanied by an improved ability to absorb negative economic shocks. Two com- 249 W;cAHTAL FLOWS TO DEVELOPING COUNTRIES mon indicators of banks' ability to absorb shocks are their capitaliza- tion rates, measured as the stock of capital relative to the stock of bank assets, and their level of provisioning, measured as the provisions made for future losses as a share of the stock of total loans. An increase in pro- visions against future losses reduces the probability that banks could go under if borrowers default on their loans. An increase in the capital- asset ratio reduces the likelihood that banks could default on their own borrowing if investment projects turn out to be unsuccessful. (In addi- tion, because shareholders' potential losses increase with the size of cap- ital-asset ratios, well-capitalized banks are usually better monitored by their shareholders and therefore hold safer portfolios than do poorly capitalized banks.) Two other indicators, related to banks' ability to withstand short-term shocks, are the liquidity of their assets and rela- tive maturity of their assets compared with their liabilities. A lower de- gree of asset liquidity and a shorter maturity of liabilities relative to assets make banks more vulnerable to liquidity crises. A booming banking industry is likely to appear profitable and strong. )While increased profitability may indicate improved operating conditions, however, in many instances it may simply be a consequence of the fact that banks are investing in riskier projects. A fast-growing banking industry has the opportunity to prevent or limit increased fi- nancial vulnerability if banks use these additional profits to increase their capitalization rates and provision against future losses. Based on the above, this section analyzes the irapact of increased fi- nancial integration on the banking sectors in a number of country episodes. The analysis consists of studying the behavior of several fi- nancial indicators obtained from the banks' balance sheets during the years surrounding the inflow episode. In particular, we look at prof- itability indices, capitalization ratios, the level of provisions and non- performing loans, and the magnitude of exposure to foreign exchange risk. We use these indicators, in addition to indicators of banks' port- folio composition, when available, to assess the financial health, risk ex- posure, and resilience to shocks of these banking sectors. Country Experiences Countries can be classified according to the way they behaved during inflow periods and lending booms: some of themn strengthened their banks, some experienced a deterioration in the health of their banking 250 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANC1At sector, and some of them experienced an improvement in some indica- tors and a worsening in others. Countries that strengthened their banks. Chile, Colombia, and Malaysia all used their most recent capital inflow and lending boom periods to strengthen their banking systems. U In Chile between 1990 and 1995, commercial ba-nks were able to improve the quality of their assets by reducing more than half the stock of nonperforming loans (as a share of the total) while at the same time increasing their liquidity by about 40 percent (figure 5.3). Similarly, commercial banks reduced their foreign exchange exposure by about 30 to 40 percent between 1988 a-nd 1994 (fig- ure 5.4). It is important to note that the improvement in the fi- nancial indicators of Chile's commercial banks was achieved despite the fact that asset profitability in this period-measured by net interest margins-fell by about 25 percent (figure 5.5). * Colombian banks, however, exhibited an increase of about 50 percent in net interest margins (figure 5.5), an improvement in capitalization (figure 5.6), a rise in loan loss provisions of about 70 percent (figure 5.7)-along with a worsening in asset qual- ity, as measured by nonperforming loans, of about 40 percent (figure 5.3)-and a 50 percent decrease in the return on equity (calculated after provisions). The Colombian case is a clear ex- ample of how authorities can use a lending boom period to tighten regulations and make banks more resilient by forcing them to use the additional profits to increase their capitalization and provisioning. * Malaysia also took the most recent capital inflow episode as an opportunity to strengthen its financial sector. Malaysian banks steadily increased their rate of capitalization after 1989, almost doubling it during 1989-95 despite the lack of a significant in- crease in profitability during the same period (figures 5.6 and 5.5). Countries whose banks became weaker. By contrast, a number of coun- tries saw a worsening in their banking systems' ability to withstand shocks during capital inflows and lending booms-in particular Argentina, Brazil, Mexico, and Venezuela during the early 1 990s; Sweden during the mid- to late 1980s; and Chile in an episode during the early 1980s. 251 CAITAL FLOWS TO DEVELOPING COUNTRIES * Venezuela, for example, shows a fall in loan loss provisions in the years prior to its banking crisis, in 1994, despite the fact that asset quality was deteriorating, as became evident after the crisis began (figures 5.3 and 5.7). Furthermore, banks' profitability (measured by net interest margin) increased in the years preceding the crisis (figure 5.5). * Capitalization rates in Brazil also fell by about 21 percent be- tween 1991 and 1995 (figure 5.6), even though during the same period banks' profitability increased by about 35 percent and asset quality deteriorated significantly (figures 5.3 and 5.5). Banks in Mexlco, however, although experimcing an important deterioration in asset quality (figure 5.3), along with a rise in profitability (figure 5.5) in the years preceding the 1994 crisis, in- creased their loan loss provisions only marginally between 1992 and mid-1994-the index rose 20 percenL during this period (figure 5.7). In addition, Mexican banks suFfered from a signifi- cant increase in their exposure to foreign exchange and real estate risks. Indeed, foreign exchange exposure grew about 40 percent between 1988 and 1994 (figure 5.4), while the share of banks' portfolios invested in real estate increased by about 80 percent (figure 5.8) between 1991 and 1995. * At the beginning of the capital inflow episode in Argentina, banks had relatively high levels of reserves and capital, which provided important cushions for managing the crisis later on (annex 5.4), although they were playing a minor role in intermediation.'0 As domestic inflation declined and inflows surged, the banking sector started to intermediate larger volumes of funds and the authorities began to implement reforms aimed at strengthening the supervi- sion and monitoring of banks. The increase in bank lending dur- ing 1992-93 occurred along with an increase in banks' profitability, a decrease in capitalization (an expected outcome, given the initial conditions), a fall in loan loss provisions, and a de- terioration in asset quality (figures 5.5 and 5.7). * The cases of Chile and Sweden during the 1980s are similar to those discussed above in many respects. Capital ratios of Chilean banks fell by about 50 percent between 1977 and 1980, despite the fact that banks' profitability was high in the years preceding the 1981-83 crisis (figure 5.6). Moreover, during the same period Chilean banks experienced a decline in the maturity of liabilities 252 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIA and an increase in foreign exchange exposure (figure 5.4). Con- versely, Swedish banks saw a fall in their capitalization rates of about 31 percent between 1988 and 1991, and increased their foreign exchange exposure by about 50 percent between 1986 and 1990 (figures 5.4 and 5.6). Countries where banking sectors strengthened in some aspects and weakened in others. The countries in the last group analyzed here were able to strengthen some aspects of their banking industries during the inflow and lending boom episodes while at the same time allowing a deterio- ration in other indicators of banking sector health. These countries include Indonesia during the 1990s, Malaysia during the early 1980s, and Thailand in both the 1980s and the 1990s. * In Malaysia, during the first capital inflow episode in the late 1970s and early 1980s, banks steadily increased their capitaliza- tion by about 80 percent, from an equity-over-total-assets ratio of about 3 percent in 1980 to a new high of about 5.5 percent in 1984 (figure 5.6). This increase, however, could not prevent the banking crisis that occurred in 1985-87, largely because during the inflow period banks aggravated their vulnerabilities by over- investing in real estate. Indeed, between 1980 and 1985 the share of Malaysian banks' portfolios represented by loans and advances to the real estate and housing sectors rose from 25 to 35 percent and increased further, to almost 37 percent, in 1987. * As in Malaysia, banks in Thailand increased their capitalization by about 35 percent between 1989 and 1995, while during 1988-94 their profitability increased by about 49 percent and the provisioning against future loan losses rose by about 24 percent (figures 5.5, 5.6, and 5.7). However, between 1989 and 1995 Thai banks also increased their exposure to foreign exchange risk by a factor of four (figure 5.4) and further increased their expo- sure to real estate risk, which had been on the rise since 1985. In- deed, the share of real estate and construction loans in the portfolios of Thai banks, which increased by about 20 percent between 1984 and 1988, rose by about 41 percent (figure 5.8) between 1988 and 1995. The case of Thailand during the 1990s differs from its capital inflow episode in the 1980s in several ways. First, there was only a minor in- 253 -+M)1TAL FLOWS TO DEVELOPING COUNTRIES crease in Thai banks' exposure to the real estate business during the first inflow period. Second, Thau banks decreased their exposure to foreign exchange risk between 1980 (or earlier) and 1985 (figure 5.4). And fi- nally the capitalization of Thai banks steadily decreased between 1978-79 and 1983, the first year of its 1980s banking crisis (figure 5.6). Indonesia's banking sector during its most xecent capital inflow episode (1990-94) exhibits several of the symptoms of a weakening banking system, except that banks significantly increased their provi- sioning against future loan losses (figure 5.7). While bank provisioning increased between 1989 and 1994, however, bank capitalization de- creased by about 30 percent during the same period, and foreign ex- change exposure rose by 275 percent (figures 5.4 and 5.6). As in Thailand and other countries, banks in Indonesia increased their expo- sure to sectoral risk during its inflow episode by raising the share of loans to the service sector by more than 7 percentage points during 1989-94. More worrisome, however, is the fact that the increase in the share of loans to the service sector, typically considered to be a nontradable sec- tor, had a negative effect on lending to the trade sector, a result that could make the overall banking system more vulnerable to variations in the value of the nominal and real exchange rate (figure 5.8). Conclusions Our analysis has shown that while some countries have used surges in capital flows and lending booms to strengthen thieir banking systems, other countries, or even the same countries in different periods, have not. By using the increased profits that often accompany a lending surge to improve banks' health and shock resistance, some countries not only mitigated the microeconomic vulnerabilities that normally surface in a fast-growing industry but, more important, may also have prevented painful banking crises. In other words, while fortifying banks' ability to react to shocks, these countries rnay have managed to limit the size of the lending boom. Conversely, countries that did not improve the conditions in their banking sectors during inflow periods, but rather allowed bank lending to increase without addressing micro- financial vulnerabilities, generally saw a banking c risis later on. This conclusion is clearly illustrated by comparing Chile's two at- tempts during the past two decades to become more financially inte- grated with the rest of the world. The first attempt, in 1979, ended in 254 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANlCtIAN Figure 5.3 Change in Nonperforming Loans Figure 5.4 Change in Foreign Currency Exposure (as a percentage of total loans) Increasing asset quality Decreasing exposure Chdie U First year of inflow 9Chile First year of inflow 1.990S ~ ~ ~ ~ ~ *Last year of inflo Thailand 3Last year of inflow s|*Ltyaof wlO | L980s Index 100 equals Index, :00 equals the value in the Decresing sset uality the value in the frtya fteifo Colombia first year of the inflow ure Chile 4 1~~~~~~~~~~~~~~~980s Philippines ±980s Mexico Brazil Sweden Brazis late ±9805 Mexico Indonesia 1990s ISSOs Thailand Venezuela L99Os, 1 990s 0 100 200 300 400 0 50 100 ±50 200 250 300 350 Figure 5.5 Change in Net Interest Margin Figure 5.6 Change in Capital-Asset Ratios (profitability index.: net interest margin) Decreasing capitalization Decreasing profitability C 9hie Chile SwFirst year of inflow .990s * First year of inflow late S80 Last year of nflow *Last year of7info Thailand 9 s { * Last year of InflOW | 1980s Index, 100 equals Malaysia Increasing profitability Index :100 equals Indones: the value in the 1990s the value in .9905 first year of the inflow Argentina________________the first Brazil Argentina year of the 1990s Vene= |inflow Increasig capitalization :L99Os Colombia 1990s Thailand __________________ ________________ 990S Brazl MalaYsia 1 990s 5 Os Colombia I Malasia ± 990s ±9-:L 90s __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Thailand 0 20 40 60 80 100 120 140 160 ±80 200 0 40 80 120 L60 Figure 5.7 Change in Loan Loss Provisions Figure 5.8 Exposure to Sectoral Risks (as a percentage of total loans) (percentage of loans to real estate and construction) Venezuela ±990s * First year of inflow 35 Argentina * Last year of inflow Malaysia, Ar990tna Index: 100 equals the 30 value in the first year of the inflow. 25 Increasing provisioning 20 Indonesia, Mexico c'90 ±L990s is Th_ln Thailand, rL990s Thailand ±0_ 1990s _ Colombia 5 1990s 0 ± 2 3 4 5 6 7 Indonesia Year 0 equals the first year of the inflow 1990S 1 Note In Indonesia sectoral risk is measured as the percentage of loans 0 150 300 450 to the services sector Source World Bank staff estimates using data from IBCA Ltd, IMF, International Financial Statistics data base, Salomon Brothers, and reports from various national central banks 255 vKE APITAL FLOWS TO DEVELOPING COUNTRIES a major economic and financial crisis after three years, while the sec- ond, which started exactly 10 years later, in 1989, has been extremely successful. In both cases Chile began the inflow period with a sound macroeconomic environment, with the only important difference be- tween the two episodes being the exchange rate policy, a factor that re- sulted in banks increasing their foreign exchange exposure during the first inflow period but not during the second.1'1 Nevertheless, on the microeconomic front the existence of a better regulatory and supervi- sory framework in Chile during the 1990s can explain the different pattern of indicators such as bank capitalization and provisioning. Be- lated attempts were made to strengthen these indicators in the earlier episode, but by then a crisis had become inevitable.12 Another conclusion suggested by the analysis is that initial condi- tions matter. This can be seen from the experience of those countries that have had two capital inflow episodes, and also by comparing coun- tries in which external financial liberalization ended in a banking crisis with those that avoided a crisis. For example, the capitalization rate of banks in Malaysia at the beginning of its second capital inflow period (1987-88) was about twice that at the start of its first inflow period (1980), while banks' capital-asset ratios in the Philippines were also higher at the beginning of the second inflow period (1988) than at the start of the first (1979). Similarly, the average maturity of deposits was shorter at the beginning of the first inflow periocd than at the start of the second in the cases of Chile, the Philippines, and Thailand. Finally, when comparing two groups of countries, it appears that those in which external financial liberalization led to a banking crisis started with lower capitalization rates than those in which a crisis was avoided-the initial bank capital-asset ratios in che former group of countries was in the range of 2 to 5 percent, while in the latter it was 6 percent or higher. Banks in countries where a crisis occurred also started with lower liquidity than those in which financial liberalization has been successful. Although these conclusions must be interpreted cautiously, since differences in accounting procecLures and other rules can invalidate cross-country comparisons, it is clear that countries that have been able to strengthen their banking industries during a surge in inflows-or in a period between two episodes of inflows-have been able to avoid the difficulties that in 1995 affected Argentina, Brazil, and Mexico, and that during the 1980s affected Chile, Malaysia, and the Philippines. 256 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCE1A Financial Sector Reforms in the Context of Integration: Policy Lessons T O CONSOLIDATE THE LESSONS FROM THE PREVIOUS TWO sections, we have constructed financial and macroeconomic vulnerability indexes that summarize the developments which followed the surge in capital flows in a number of countries. These indexes are illustrated graphically in figure 5.9. A positive number in the chart means a deteriorating situation, and a negative number means an improving one.13 Our analysis found that financial sector conditions and macroeco- nomic conditions generally deteriorated in tandem, that banking crises occurred in countries where macroeconomic vulnerabilities increased or remained constant, and that no banking crises occurred during capital inflow episodes in which macroeconomic vulnerabilities were contained. These conclusions are consistent with the view that developments in the financial sector are, in general, a mirror image of developments in the real economy, and that the banking sector amplifies shocks that have already occurred in the real economy. 14 Indeed, none of the country episodes in- cluded in figure 5.9 show an increase in macroeconomic vulnerabilities along with an improvement in the financial sector, and none of the bank- ing crisis cases show a worsening in banking sector vulnerabilities with an improvement in macroeconomic conditions. The evidence therefore suggests that maintaining a strong macro- economic stance may be critical to avoiding a banking crisis. In partic- ular, the cases of Indonesia (1990-94), the Philippines (1988-94), and Thailand (198 8-94) suggest that having strong macroeconomic funda- mentals may enable an economy to escape or at least temporarily dodge a crisis even when financial conditions are deteriorating (box 5.5). Specifically, the need to curb the lending boom and to apply counter- balancing macroeconomic policies to prevent overheating when receiv- ing large capital inflows appears as a very clear message. Because of the long time-usually several years-that it takes to put in place a sound and reliable regulatory and supervisory framework for banks, countries can help make financial integration less destabilizing by maintaining prudent macroeconomic performance. This does not, however, obviate the need for strong macroeconomic and financial sector conditions to avoid a banking crisis in the long run. 257 100-111-ov AL FLOWS TO DEVELOPING COUNTRIES Figure 5.9 Mllacroeconomic and Financial Sector Vulnerabilities during Capital Inflow Episodes, Selected Countries and Years Change In vulnerability index 3~~~~~~~~ N011CflSIS courfltri( ,~ 0 2~~~~~~cii countries |Nncji5cunret -2 sWJ~~~~ ~ CD . , - I I* ,> !2[ ~ W ~ ~O Ea di UMacroeconomi lneabcitiaes F st u ali 1, ~ ~~~~~ _1__ I I0 a I) M r __ | =~~~~~~~~~~~~~~~~. CD Io In I Y0- o FE1 Macroeconomic vulnerabilities 11LFinancial sector vulnerabilities| Source World Bank staff estimates. Banking Crises generally occur in Next, we examine policy options that emerge from the discussion countries where macro- and micro- above, as well as from previous studies, on the irnplementation of fi- economic vulnierabilities increase during capital inflow episodes. nancial sector reforms in different countries. Our analysis adds to pre- vious studies by introducing the issue of integration. In considering various policy options, it is important to understand how financial integration may affect the banking sector in countries re- ceiving large capital flows. Financial integration may result in: * increased competition in the form of, for example, new entrants and financial disintermediation 258 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANN;Cg * new risks in the form of new investment opportunities and in- struments, and risks created by a potentially more volatile envi- ronment, including the risk that flows will be suddenly reversed * access to a larger pool of foreign and domestic capital (savings), which will enable banks to increase lending. These three effects, in conjunction with conditions in the financial sector and the macroeconomy, will determine the outcome of increased financial integration. Although financial integration may be beneficial for developing countries in the long run (chapter 3), potential risks and losses are greater if the process is poorly managed.15 This is so because the size and speed of market reactions increase in an integrated econ- omy, as is shown by the Mexican experience in early 1995 (Calvo 1996, Griffith-Jones 1995). Financial Integration: Guiding Principles for Financial Sector Reform As oudined above, financial integration makes it easier for banks to rapidly increase the riskiness of their portfolios and consequendy incur sizable losses.16 Moreover, standard supervision and monitoring tools be- come less effective with integration because of the increased speed and magnitude of market reaction. Thus, financial integration puts a pre- mium on the need of developing countries to reform the institutional and regulatory framework governing their banking industries. Given the im- portance of macroeconomic stability for sustained reforms, and the possi- bility of increased macroeconomic swings in an open environment, these reforms need to be underpinned by supportive macroeconomic policies. Past attempts by both developing and industrial countries to liberal- ize their domestic financial sectors have led to an evolving view of best practice in financial sector reform. In this section we review how finan- cial integration affects the domestic financial sector reform agenda con- cerning both macroeconomic and microeconomic policies. The policy recommendations discussed below consist of a minimum set of condi- tions that, in an ideal world, should be satisfied before financial integra- tion begins. For countries that have already embarked on the process of integration, improvements in these areas are critical to averting financial distress or banking crises. Table 5.2 summarizes the major findings. Creating a macroeconomic environment to sustain financial sector reform. In a world of increasing financial integration policymakers need to pay 259 _T- 11111*0' TAt FLOWS TO DEVELOPING COUNTRIES Table 5.2 Domestic and External Financial Sector Refonn: Policy Recommendations Macroeconomic environment Microeconomic environment Timi7ng of Context Overall Real reforms within Financial Incentive Regulatory of reform ssability interest rates economic cyce infrastructure strcture framework Domestic Implement Avoid sharp Implement Put in place Put in place a Impose high financialsector reforns when increases in reforms adequate good system capital-asset refisr the econorny interest rates during a institutions to screen ratios and is stable to that will recovery and upgrade potential proceed avoid a reduce period to human capital bankers and slowly with shortening of borrowers' facilitate the in both banks limit deposit reforms to maturities in repayment adjustment of and insurance avoid a sharp banks' capacity and banks to the supervisory coverage to fall in bank liabilities, banks' net new agencies small profits. worth. environment. before depositors Impose Control liberalization only ceilings on interest rates if occurs. self-lending they become within too high. conglomerates. Increased The same Accelerate Pursuing The same policy recommendations apply. finacial recommen- reforms if reformns after However, because potential losses are much greater integrasioa dation applies: interest rates an economic and markets react faster under integration, the implementing are expected slowdown need to move forward on these fronts is even reforms in a to decline. may greater. Increased financial integration puts a volatile However, exacerbate the premium on reforming the regulatory and environment controlling lending boom incentive structures in banking. In addition, the will shorten interest rates and use of guarantees for foreign eachange should be the maturity will be more overheating of limited. of flows and difficult and the economy. induce more costdy. However, speculative other flows. counteracting Emphasize policies can be achieving used to anacro- minimize this economic problem. stability and eliminating debt overhang. attention to at least three aspects of the macroeconomic environment to ensure successful banking sector reform: ovei all macroeconomic stability, the level of real interest rates, and the phase of the business cycle at the time reforms are implemented. 260 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAU Macroeconomic stability. One of the central lessons that has emerged from the experience with domestic financial sector reform is the im- portance of macroeconomic stability as a prerequisite for sustained re- form. Because of the undesirable effects of sudden reversals of capital flows, developing countries that liberalize their financial sectors while becoming financially integrated should therefore move quickly to achieve a stable macroeconomic environment, one in which large macroeconomic swings are avoided and where the volatility in some key prices, such as the exchange and interest rates, has been reduced to internationally comparable levels. In working toward stability policy- makers should aim to build macroeconomic cushions such as holding an increased stock of international reserves and a reduction of the pub- lic debt (Goldstein and Turner 1996). Countries with a public debt overhang problem are more vulnerable than others to changes in mar- ket sentiment and changes in variables such as interest rates or aggre- gate output (chapter 4). Real interest rates. The macroeconomic environment chiefly affects the outcome of financial sector reforms by changing borrowers' and banks' net worth through variations in the level of real interest rates (annex 5.2). For example, several authors have argued that the high real interest rates that prevailed in Chile in the late 1970s are partly respon- sible for the financial crisis of the early 1980s in that country (Hernan- dez 1991). Accordingly, it has been argued that policymakers should carefully monitor interest rates and prevent them from reaching too high a level-for example, by imposing maximum rates (Stiglitz 1994; Hellmann, Murdock, and Stiglitz 1996). The desirability of pursuing or delaying financial sector reforms in an integrated environment also depends on expectations about interest rates that can affect borrowers' and banks' net worth.17 It is likely that real interest rates will be low under integration because of the increased supply of loanable funds and increased competition in banking, as a re- sult of which, spreads will fall. These low rates will have a positive im- pact on borrowers' repayment capacity and through this on banks' net worth. Therefore delaying reforms because of interest rate considera- tions does not seem appropriate. Moreover, the ability of policymakers to control interest rates in a more integrated environment is greatly re- duced, and the process becomes increasingly costly. Phase of the business cycle. Experience with domestic financial sector reforms suggests that reforms should be accelerated in periods of fast 261 F4ATAL FLOWS TO DEVELOPING COUNTRIES economic growth and delayed when the economy slows down (Caprio, Atiyas, and Hanson 1994). The reason for this is that the banking sec- tor-like every other economic sector-can adjust more easily to the structural changes brought by reform in periods of high profitability and growth. In the case of increased financial integration, however, this argument changes. In particular, the surge in flows that normally fol- lows the implementation of reforms triggers a Larger lending boom than in a less integrated economy. Moreover the possibility that the economy might start growing rapidly after reform; are introduced can reinforce market participants' high expectations about growth. This chain of events, as explained in previous sections, may induce addi- tional lending and trigger overheating, which coald exacerbate both macroeconomic and financial sector vulnerabilities. It might therefore seem that in an integrating ervironment financial sector reforms should be pursued during an econornic slowdown rather than during a recovery period, because overheating and overlending are likely to be less severe if the financial liberalizatiorL process starts in an economic peak rather than a trough. On the other hand, pursuing fi- nancial liberalization during a recession may lead to capital outflows rather than inflows. Therefore domestic financial ieform should be ac- celerated in periods of fast economic growth and delayed when the economy slows down, and the problems of overheating and overlend- ing should be addressed through other policies. Creating a microeconomic environment to sustain financial sector reform. Three microfinancial areas need to be given special attention when liberaliz- ing the financial sector in an integrated environment: financial infra- structure and institution building, incentive struciure, and the regula- tory and supervisory framework. Progress in all three areas-and in the macroeconomic stance as well-is necessary to ensure that the reform process leads to a safer and sounder banking industry.18 Financial infrastructure and institution building. One lesson that emerges from past episodes of financial sector reform is the need to es- tablish an appropriate institutional framework, providing clear stan- dards and rules for accounting, auditing, and legal procedures, all of which are necessary for adequate functioning of a market economy.'9 In addition, experience shows that bankers' management and risk as- sessment skills often become obsolete after reforms are introduced and need to be strengthened when the system is liberalized (Brock 1996; Caprio, Atiyas, and Hanson 1994; Caprio and Vitcas 1995; and Caprio 262 THE EFFECTS OF INTEGRATION ON DOMESTIC FINAN 1996). For example, the lack of staff with modern banking skills con- tributed to the banking crises in the Baltic republics of Estonia, Latvia and Lithuania (Fleming, Chu, and Bakker 1996). This is particularly important under integration because lack of knowledge in managing new instruments and assessing their risks increases the potential for losses caused by fraud or wrongdoing. Moreover, lack of experience and knowledge in supervising banks in the riskier new environment in- creases the likelihood that such problems will remain undetected. This has been a problem even in more advanced economies, for example, Finland, Norway, and Sweden, where the lack of supervision was a key factor in banking crises. The incentive structure. One important reason that banks run into fi- nancial difficulties is the presence of distorted external incentives- policies that can be modified by domestic economic authorities and that induce bank managers to pursue unsound banking practices (box 5.6). An example of this, one particularly important in the context of increased financial integration, is the existence of explicit guarantees for foreign exchange. As shown by the Chilean experience of the early 1980s, and more recently by that of Mexico, these guarantee mecha- nisms induce banks to increase their exposure to foreign exchange risk (annex 5.2). Another possible distortion is a deposit insurance scheme that is not properly priced according to each bank's risk level, which may lead depositors to put their money into institutions paying the highest interest and holding the riskiest portfolios.20 The solution is to correct the incentive structure and tighten bank supervision. Increased financial integration may aggravate this problem by giving domestic banks access to a larger supply of funds from abroad, espe- cially if foreign deposits are covered by implicit or explicit insurance. A simple way to minimize the risk that increased financial integration will exacerbate this problem is to limit deposit insurance, thus forcing large depositors to scrutinize banks carefully before deciding where to put their money; this should impose some degree of market discipline on banks. More important, by excluding foreign (or foreign-currency- denominated) deposits from the insurance mechanism, authorities will force foreign lenders (usually large foreign banks or institutional in- vestors) to look very carefully at banks, exerting even more control over them. In this way market discipline complements the imperfect moni- toring of banks by domestic supervisory agencies, which may be ill equipped to effectively control risks and monitor such areas as prof- 263 11ITAL FLOWS TO DEVELOPING COUNTRIES Box 5.46 Good Bankers Make a Safe and Sound Banking Industry DISTORrED EXTERNAL INCENTIVES ALONE ARE The problem of screening prospective bankers not suff cient to create a financial crisis. Indeed, it makes it important to establish a cLear and transpar- has been argued that no matter how distorted the ent application process for bank licenses, in which all incentives, cautious bankers rarely bring their in- high-level managers and owners of banks are subject stitutioris to the brink of bankruptcy (De Juan to careful screening. One important lesson concern- 1987). Banking practices depend on both external ing this issue emerges from the experience with bank incentives and internal factors, which are usually privatization in Chile in the early 1980s and Mexico difficult to observe and not under the control of in the early 1990s. In both episod&s banks were pri- domestic authorities, such as management's repu- vatized without attention being paid to the qualifica- tation a nid moral track record. The main concern tions of the new owners or how they were financing is that risk-taking and fraud-prone individuals the purchase. In some cases banks were bought with could eniter the banking industry after liberaliza- debt instead of equity, so the tnew bankers had tion occurs. In the presence of distorted incentives greater incennves to make risky investments with de- these individuals-as opposed to more conserva- positors' money. Furthermore, reforms were intro- tive and cautious bankers-may pursue unsound duced shortly after privatization occurred, giving the lending practices and thereby aggravate financial new management teams little time 1:o learn how to do fragility in the banking sector. This is particularly business in the new environment. These develop- true in ithe case of increased financial integration, ments were partly responsible for r he banking crises which offers greater opportunities for fraud and experienced by Chile in 1981 and by Mexico in wrongdoing. Thus, the problem consists of both 1995. Poor screening of new entrants also played an carefully screening prospective bankers and cor- important role in the difficulties faced by the bank- recting the external incentives facing bank ing sector in Hong Kong in the late 1970s, and more managers. recently in Poland and Russia. itability, capitalization, and interlocking relationships between banks and the corporate sector. These issues are discussed below. Regulatory and supervisory framework. When implementing financial sector reforms, policymakers should pay particular attention to banks' profitability and capitalization, cross-ownership between banks and corporations, and bank supervision. Banks' profitability and capitalization. Experience shows that poorly capitalized and money-losing banks are more inclined to undertake ex- cess risks than those that are well capitalized and profitable, because the owners of safe, healthy banks have more to lose if their risky invest- ments turn out to be unsuccessful. Therefore, to limit excess risk taking by banks when implementing financial sector reforms, particularly in an integrated environment, authorities should raise capitalization re- quirements to levels consistent with international standards, and devel- 264 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANC IAL _t oping countries have already started doing so.21 Such actions should be undertaken within a consistent and well-articulated incentive and reg- ulatory framework. Moreover, some analysts have recently argued in favor of imposing higher capitalization rates for banks in developing countries because these banks operate in an especially volatile eco- nomic environment (Gavin and Hausmann 1996, Caprio 1996). If this proposal is followed, then developing countries will need to keep strengthening their banking sectors by raising capital requirements in banks, since these banks do not appear to have risk-based capital ratios significantly higher than those in larger industrial countries (IMF 1996, Goldstein and Turner 1996). Financial reform in an integrated environment can trigger a lending boom and temporarily increase banks profits (and risks), but it can also reduce profits in the medium and long term. It does so by increasing market competition and giving nonbanks new business opportunities, a situation that can be aggravated if it precludes domestic banks from entering these new markets. In this context, a financial deregulation process that is not carefully managed may therefore reduce banks' net worth in the medium and long term and induce unsound bank prac- tices. Policymakers in liberalizing economies should proceed slowly to avoid a sharp decrease in banks profits while allowing all market par- ticipants to compete on a level field (Stiglitz 1994). As argued earlier, the temporary rise in profits resulting from lending booms should be used to increase capitalization and provisions. Cross-ownership between banks and the corporate sector. Experience shows that close relationships between the corporate sector and banks tend to distort the incentive for banks to protect depositors' money. More specifically, an interlocking relationship between a bank and a group of firms tends to bias the incentives of bank managers toward protecting the interests of the group of firms rather than those of de- positors. In so doing, banks usually end up investing too much in the firms' assets-a phenomenon called connected lending or self-lend- ing-and therefore do not diversify their portfolios adequately. This has contributed to banking problems in many countries such as Ar- gentina, Bangladesh, Brazil, Chile, Indonesia, Malaysia, Spain, and Thailand. In Chile, for example, it has been estimated that self-lending for the most troubled banks at the peak of the crisis ranked between 20 and 45 percent of total loans (de la Cuadra and Valdes 1992). More- over, an interlocking relationship between a bank and a group of firms 265 A TAL FLOWS TO DEVELOPING COUNTRIES renders standard regulatory and monitoring proczdures less effective, particularly in cases of financial distress, because the ways a bank can channel money to firms-and hide losses-in a conglomerate multiply with the size of the conglomerate. This makes prudential supervision more difficult. These problems become even more complex when both banks and corporations gain access to offshore markets as a result of in- tegration. To solve these problems, policymakers must set limits on both bank holdings of stocks and on lending to individual borrowers, the latter calculated on a consolidated basis.22 More important, be- cause of the difficulties in enforcing these regulations (especially during periods of financial distress), bank managers should be subject to severe penalties if found guilty of violating them. Bank supervision. Any effort to limit risk taking by banks and promote sound banking will be ineffective if supervision and enforcement of reg- ulations are weak. Supervision becomes more difficult in an integrated environment because of the larger pool of assets and risks involved and the fast-changing market conditions. Therefore, the' need to upgrade the skill mix and the effectiveness of the supervisory agency in an integrating environment becomes even more urgent than in an isolated economy. Two elements can make supervision more effective in an integrated environment: the development of prudential standards and cooperation in cross-border supervision. Although progress in developing prudential standards has been made in recent years, the standards adopted pertain mostly to capital requirements and were designed primarily for interna- tional banks in industrial countries. The rash of sex-ere crises in develop- ing countries over the past 15 years has, however, recently rekindled interest in developing more comprehensive guidelines. Indeed, supervi- sory and regulatory authorities are expected to issue such guidelines in the near future. As regards cooperation among supervisors, authorities in the home country bear the main responsibility for cross-border supervi- son, but authorities in host countries can help by monitoring specific as- pects, such as liquidity, and, most important, by removing impediments to the exchange of relevant information between supervisory authorities. The Pragmatic Approach: Containing the Lending 13oom while Strengthening the Banking Sector Although it would be ideal for the policy recomrmendations discussed above to be implemented before integration begins, most developing 266 THE EFFECTS OF INTEGRATION ON DOMESTIC FINA_NA_ countries that are already financially integrated do not have all these pre- requisites in place. This means that they need to move quickly to fix their institutional framework and put in place all the missing elements. Since doing so can take years, developing countries can implement other policies in the meantime to prevent vulnerabilities from being made worse. Curbing the lending boom is one remedy that is especially important in countries where the banking sector is weak and where pri- vate sector spending is biased towards consumption rather than invest- ment. This subsection discusses some options for reducing the degree to which a lending boom can aggravate macroeconomic and banking sec- tor vulnerability. A summary of the discussion is presented in table 5.3. Macroeconomic policies. Fiscal, monetary, and exchange rate policies can all help to mitigate the lending boom. Tight fiscal policy can be highly effective in counterbalancing the overheating caused by the lending boom (box 5.5); at the same time it can prevent an increase in capital inflows by keeping domestic interest rates low (a desirable side effect). By contrast, a tight monetary policy, while containing the increase in aggregate spending, may induce additional capital inflows by raising domestic interest rates (Corbo and Hernandez 1996). A larger volume of capital inflows will exert additional pressure on mon- etary balances and increase the quasi-fiscal deficit of the central bank when it attempts to sterilize the flows. Moreover the positive effect of sterilization on domestic interest rates may significantly worsen the fiscal balance for countries having a large stock of outstanding public debt (for example, India or Mexico). In the end, unless compensatory fiscal policies are implemented, the deteriorating financial position of the central bank-and the public sector in general if the outstanding stock of government debt is large-will cause a worsening of macro- economic vulnerabilities. Furthermore, when banks have weak loan portfolios, an increase in domestic interest rates may aggravate their problems by increasing the stock of nonperforming assets. A recent example of this is Mexico during 1993-94. Exchange rate policy, when directed toward maintaining competi- tiveness and supported by a tight fiscal policy, has proved to bias aggre- gate spending away from consumption and toward investment, particularly in the tradables sector (chapter 4). The use of the exchange rate as a nominal anchor, however, has often been associated with con- sumption booms, as in Argentina and Mexico. Not using the exchange rate as an anchor can also help to reduce the risk that banks will over- 267 CAPITAL FLOWS TO DEVELOPING COUNTRIES Table 5.3 Containing Lending Booms: Policy Options Macroeconomki policies Imposing ceilings on commercial bank Foreign excbange Monetary lending and external system Fiscal policy policy borrowing Main advAntages A semifloating Reduces overheating Helps contain the Litnits the lending exchange rate system - and helps keep inrterest lending boom and boom and the (band) increases the rates low. -reduces overheating, overheating. Most market risk faced by effective if directed banks and othet toward specific uses of domestic borrowers, credit such as thereby redueing over- consumption loans, all external borrowing. credit cards, and mortgages. Main diswavantages Exacerbates capital Creates a micro- inflows by keeping economic distortion domestic interest rates and gives nonbank high. Also, in the financial institutions medium term, it an advantage over impairs the financial banks. situation of the central bank-and of the public sector if the stock ofoutstanding debt is large-and damages macroeco- nomic fundamentals. lend. Indeed, the adoption of a semifloating eschange rate system (band), by increasing market participants' exposure to foreign exchange risk, induces a more cautious approach toward external borrowing, re- sulting in smaller capital inflows and less bank lending and overheat- ing. This was in fact the experience of Chile during the 1990s, as compared with that in the early 1980s. Overall, the experience of countries that have avoided overheating during the most recent surge in capital inflows-such as Chile, Indone- sia, and Thailand-suggests that sound macroeconomic management requires using all three policies within a consistent framework. A tight fiscal policy is crucial for the policy mix to be sustainable, and sterilized intervention accompanied by exchange rate flexibility-mn some cases 268 ITHE EFFECTS OF INTEGRATION ON DOMESTIC FINAAN Microeconomic policies Increasing banks' Imposing indirect Tightening other bank Increasing banks' reserve risk-adjusted capital-asset (economy-wide) capital regulations (provisioning requirements ratios controls for nonperforming loans) Restricts credit growth and Makes banks more resilient Capital controls appear to Increasing provisions for minimizes the risk of to shocks, induces sound change the composition of future loan losses makes overlending. banking practices, and may flows toward longer banks more resilient to reduce the growth in maturities. This may have shocks and reduces the lending. the positive effect of biasing availability of loanable expenditures (and bank funds. lending) toward investment rather than consumption. If not remunerated, the An excessively high capital- It seems that capital controls Provisions against future increase in reserves may asset ratio will reduce will quickly become loan losses are difficult to induce banks to invest in creditors' efforts to monitor ineffective because of the monitor when bank riskier projects. bank managers and can important arbitrage supervision is weak. This erode banks' profitability. opportunities they create. problem becomes more Capital requirements are Most important, these acute in periods of financial dlifficult to monitor if bank profits usually benefit other distress, when banks start supervision is weak, economic agents (nonbanks) rolling over bad debts and especially in periods of and in the long term create capitalizing past-due interest. financial distress. distortions and cause economic inefficiencies. strengthened with the use of capital controls-can play an important role in avoiding overheating and reducing vulnerability, especially so in the early stages of capital inflows (chapter 4 and Montiel 1996). Microeconomic policies. Microeconomic policies directed at containing the lending boom include increasing banks' capitalization and provi- sioning against future losses, raising reserve (liquidity) requirements, and imposing capital controls and ceilings on credit growth and exter- nal borrowing. Increasing capitalization requirements. Unless weighted by the risk of banks' assets, increasing capitalization requirements does not guarantee that banks will pursue a more sound investment and lending strategy- in fact, it could have the opposite effect and induce bankers to increase 269 L FLOWS TO DEVELOPING COUNTRIES the risk in their portfolios. Therefore countries should introduce risk-ad- justed capitalization requirements such as those recommended by the Bank for International Settlements (annex 5.3). Introducing risk-ad- justed capitalization rates, as Argentina and Chile recently did, would- along with the implementation of other reforms discussed in this section-help induce banks to shift their portfolios towards safer assets, thus making the entire banking industry more resilient.23 Although this would not necessarily reduce overall bank lending, ii would negatively af- fect lending to the riskiest activities, such as home mortgages, commer- cial real estate, and consumption credit.24 Excessively high capital requirements, however, could reduce banks' profitability and create other economic inefficiencies, such as inducing financial disintermediation.25 Raising the minimum provisioning againstfuture loan losses. This will improve banks' ability to absorb shocks. In addit[on, it may induce a portfolio shift toward safer assets-those with lower provisioning re- quirements-and thereby indirectly reduce overheating. More impor- tant, by reducing banks' net profits and capikal, this policy will constrain the growth in bank lending, with a direct negative effect on credit growth. This policy has been recently applied, for example, in the Czech Republic, where in September of 1994 tighter norms were introduced that caused an increase in the recorded volume of bad loans in that year from 23 to 38 percent of outstanding credit. One problem with relying on provisioning and capital requirements to contain a lending boom, however, is that especially during periods of financial distress bank managers tend to hide their problems by under- estimating the stock of nonperforming assets. It has often been ob- served that during periods of financial distress, banks will roll over nonperforming loans and capitalize past-due interest. Both practices were standard in Chile and Colombia in the early 1980s, and more re- cently in Mexico. In the latter case these practices were facilitated by rules permitting banks to declare as nonperforming only a portion of a loan if the borrower was partially meeting his payments. A more cau- tious approach would have been to declare the entire loan nonper- forming. As in Mexico, in a number of Asia-Pacific Economic Cooperation Council developing countries, loans are classified as non- performing only after the loan has been in arrears for at least six months, and in some cases bank management itself-rather than bank supervisors-set the classification criteria (IMF 1995). This problem has been particularly difficult to monitor in situations where lending 270 THE EFFECTS OF INTEGRATION ON DOMESTIC FINA)NNC goes to related parties (self-lending), and it is especially acute in finan- cially integrated economies where banks and corporations have access to offshore markets. Therefore, for provisioning and capital require- ments to be effective, it is critical to have a strong bank supervisory agency with a highly qualified staff that is capable of uncovering a bank's actual financial situation. Reserve requirements. Although normally understood as a device to prevent liquidity crises, reserve requirements directly affect the amount of funds banks have available to extend credit. Therefore, increasing minimum reserve requirements will help to dampen the lending boom. If, however, these requirements are unremunerated, they can act as a tax on banks (Brock 1996) and may induce them to pursue a riskier in- vestment strategy. Reducing reserve requirements, on the other hand, may exacerbate the credit boom. This is illustrated by the contrasting experiences of Malaysia and Mexico. In the case of Malaysia, the au- thorities increased reserve requirements on banks and other financial intermediaries by one percentage point in 1991-from 6.5 to 7.5 per- cent-to dampen the monetary impact of capital flows. Overall, the cumulative increase in the statutory reserve requirement during the capital inflow period was 6 percentage points-from 5.5 percent in 1990 to 11.5 percent in 1994. In contrast, in the case of Mexico in early 1993 the authorities reduced reserve requirements de facto by al- lowing banks to meet this obligation through the sale of securities to the central bank under a repurchase agreement. A direct consequence of this policy was increased market liquidity and a rise in lending by small, and riskier, banks that exacerbated further both macroeconomic and financial vulnerabilities in Mexico. Other countries that have in- creased bank reserve requirements recently are Colombia, the Czech Republic, Indonesia, and Sri Lanka (IMF 1995). Ceilings on commercial bank lending. Imposing ceilings on commer- cial bank lending (or credit growth) has a direct negative effect on ag- gregate spending and therefore helps to reduce overheating. This type of policy is most effective in protecting against macroeconomic and fi- nancial sector vulnerabilities if directed toward particular types of credit, such as consumption loans, credit cards, and mortgages. In re- cent years countries such as Malaysia and Thailand have used these re- strictions (Corbo and Hernandez 1994). Countries that have such restrictions should consider leaving them in place while receiving large capital inflows, in order not to aggravate the overheating. 271 C,APITAL FLOWS TO DEVELOPING COUNTRIES Restrictions on commercial banks' external borrowing. A similar ratio- nale applies to imposing restrictions on commercial banks' external borrowing, such as those implemented in Indonesia in the early 1 990s. The use of direct capital controls on banks can be jiastified also because of the existence of implicit or explicit deposit insurance (or a foreign exchange guarantee), which country authorities do not want to extend to foreign depositors (Dooley 1996). The increasing coverage of rising deposits may put in Jeopardy the solvency of the deposit insurance scheme or, what is the same, the viability of the government budgetary program. The latter, in turn, can lead to a change in market sentiment toward the recipient country, trigger a reversal of flows, and cause a banking crisis (Brock 1996). Although the last two types of policies-ceilings on commercial bank lending and foreign borrowing-can be justified by the fact that commercial banks are the largest intermediaries in most countries, it is important to note that they may also cause significant economic ineffi- ciencies. By discriminating against commercial banks, these policies may induce financial disintermediation and thereby erode banks' prof- itability and increase their risk-the former because of the rise of other nonbank financial institutions and the latter because banks' prime clients start tapping into international capital markets for loans. Indirect capital controls. Controls that apply to different types of for- eign flows regardless of the agent intermediating the funds have been used in several countries, such as Brazil, Chile, anc[ Colombia (chapter 4). They put a tax on the use of foreign funds and impose a wedge be- tween domestic and foreign interest rates. Moreover they create impor- tant arbitrage opportunities that induce agents i o look for ways to bypass the restrictions. Indirect capital controls may reduce overheat- ing and vulnerabilities in the banking system to the extent that they af- fect the overall volume of intermediation, but this is a doubtful result (chapter 4). A more plausible result is that banks will increase their long-term lending, leading to an increase in investment rather than consumption. Nevertheless, because evasion is usually easier for non- banks than for banks, in the medium and long term this type of re- striction erodes banks' profitability and induces disintermediation. In sum, a variety of macroeconomic and microeconomic policies can be used to contain the lending boom and miniimize the impact of private capital inflows on macroeconomic and financial sector vulnera- bilities (box 5.7). Nevertheless, discretion is advised because microeco- 272 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL nomic policies could be distortionary and have an adverse effect on banking in the long run. Crisis preparedness and management As noted above, financial integration usually gets under way before countries possess all the elements necessary for their financial markets to function smoothly. During this transition period developing coun- tries may experience strains in domestic financial markets that could, if not properly managed, lead to systemic crises. Because financial crises often cause the loss of several percentage points of GDP (box 5.1) and may cause the derailment of painfully achieved macroeconomic (fiscal) stability, it is very important for a country to be able to contain them. Accordingly, crisis preparedness and management form an important aspect of bank supervision and regulation.26 This subsection reviews the main lessons that emerge from the analysis of actual banking crises and highlights their relevance in the context of increased financial inte- gration. The main conclusion is that the need to act promptly is even more urgent under integration. Table 5.4 summarizes the discussion. Among the objectives of banking crisis management are to quickly restore public confidence in the banking system in order to avoid severe reductions in liquidity, bank runs, and contagion effects, and to con- tain the extent and cost of a potential crisis, which is usually borne by taxpayers. Experience shows that to achieve these two objectives, poli- cymakers need to act promptly and decisively and avoid policy reversals that increase market uncertainty. To do so, they need to identify ahead of time the areas of greatest vulnerability in the banking system and have a clear plan to deal with a crisis when it occurs. Crisis preparedness. One lesson that emerges from past banking crises in developing countries is that policymakers, regulators, and bank supervisors need to have adequate information to realistically assess the nature and magnitude of the crisis. Since incomplete or mislead- ing information may induce wrong policy actions, they need to emphasize data collection and use these data to identify the most criti- cal areas of weaknesses. Common areas of concern are banks' poor risk assessment and management, self-lending, low capitalization and insufficient provisioning for future losses, poor portfolio diversifica- tion, foreign exchange exposure, and maturity mismatches.27 Proper information in these areas will help in preparation of a crisis contin- 273 1= ITAL FLOWS TO DEVELOPING COUNTRIES Box 5.ir Containing the Lending Boom: Taxation ONE ARGUMENT GENERALLY OVERLOOKED IN Chilean tax system provided a disincentive for indi- the literature is that tax incentives may increase bank viduals to invest in equity and therefore hindered lending. Although taxes should not be intended to firms from financing their operations through eq- contain a lending boom, at the margin they can ex- uity. This disincentive to equity investment meant acerbate (or ameliorate) its effects. Tax structures that individuals taxed at a 10 percent rate (marginal) that artif cially favor indebtedness in the corporate or would receive as much as 67 percent more dispos- household sectors-or that do not discourage it- able income-net of taxes-when investing in a may lead to increases in borrowing by either sector bank deposit than when buying a stock rendering during f nancial integration, when credit becomes the same payoffs. This tax incentive for debt financ- more easily available. In the presence of poor risk as- ing was eliminated in the tax reforrm implemented in sessment, supervision, or management in the bank- 1984. Nevertheless, its presence during the first ing sector, such borrowing helps to increase financial episode offinancial liberalization and opening of the and macroeconomic vulnerability. The experiences capital account, when credit becane more easily of Chile during the late 1 970s and early 198ts, and available, facilitated the rapid grow:h of the banking those of Finland, Norway, and Sweden during the sector and helped to aggravate both macroeconomic mid-198)s and early 1990s, each case ending in a and financial fragility in the Chilean economy banking ,crisis, clearly illustrate this point. (HernAndez and Walker 1993). The tax system in effect in Chile during 1975-84 Similarly, the tax structure in effect in Finland, (after a major tax reform in 1974) provided a signif- Norway, and Sweden when their domestic banking icant tax advantage to debt financing-as opposed systems were liberalized, in the second half of the to equity financing-for firms.I By taxing dividends 198Os, providled incentives for household indebted- twice, first at the corporate level, when companies ness at a time when credit became nore easily avail- paid taxcs on profits, and then at the personal level, able because of the reforms. In particular, in all three when individuals paid taxes on personal income, the countries high marginal tax rates-above 60 percent gency plan, which in turn can help reduce decisionmaking time, trial- and-error policies, market uncertainty, and the overall cost of a crisis. Chile's new banking law, which increased disclosure requirements on banks, represents a recent attempt to overcome this problem. Because financial integration implies a higher speed of market reac- tion to any unusual developments, and wider swings in prices and quantities, the need to collect consolidated information on banks, as- sess their main weaknesses, and plan for possible contingencies be- comes more urgent than it would be if the econorny were isolated. In particular, bank supervisors need to carefully monitor off-balance and offshore operations and the extent to which these are used as a device to hide nonperforming assets and bypass domestic regulations. The practice of lending to third parties abroad-which then relend the 274 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCj4 in Finland-and full tax deductibility of interest ductible from taxable income at the same flat rate, payments led households to increase their borrow- excluding interest paid on consumer loans other ing for both mortgages and consumption loans. The than those related to the purchase of a permanent increase in bank lending and household expendi- residence. Other earned income, such as wages, is ture, in turn, reinforced the lending boom by raising still taxed at an increasing marginal rate that goes real estate prices and collateral values and financing above 60 percent. The new tax structure, however, a construction boom. In the end, banks' portfolios eliminates tax arbitrage by incurring greater debt. became more exposed to cyclical sectors such as con- In sum, it is clear from these examples that at the struction, real estate, and services, and to exchange margin tax incentives can exacerbate or ameliorate rate risk-because of the increase in foreign-cur- the effects of a lending boom, even though they rency-denominated loans (Drees and Pazarbasioglu should not be used as a main policy tool to contain 1995). a boom. Therefore such incentives can be thought of Since their banking crises in the second part of as a complement to enhance the effect of other, the 1980s and early 1990s, all three Nordic coun- more permanent policies (such as fiscal policy and tries launched reforms that reduced the tax rates on strong bank supervision) aimed at containing a capital income and therefore the incentives for quick expansion in bank lending. households to incur high indebtedness-Sweden in 1991, Norway in 1992, and Finland in 1993. In- 1. The tax incentive for debt financing existed prior to deed, "all three countries currently impose a fairly the 1974 tax reform. However, the financial reptession and low flat rate on capital income, and one can speak of credit rationing that existed in the Chilean economy until the mid-1I 970s reduce its importance for explaining the in- a Nordic model of taxation of capital income" debtedness in the corporate sector. For a more detailed (Tikka 1993, p. 348). In the case of Finland, for ex- analysis of these issues see Hern5ndez and Walker (1993). ample, starting in 1993 capital income was taxed at a flat 25 percent rate, and interest expenses are de- same funds to a related domestic party-has been used in some instances to bypass domestic regulations against self-lending. This occurred in Chile during the early 1980s and Venezuela during the 1990s. In the latter case, however, the Venezuelan authorities had no legal powers to supervise off-balance and offshore bank operations until 1994. Moreover, the recent Mexican and Venezuelan banking crises illustrate how off-balance transactions-involving financial de- rivatives-and offshore operations can become a very important source of losses (Garber 1996). In the case of Venezuela's Banco Latino, for ex- ample, it is estimated that one-third of the $3 billion in losses were reg- istered off-balance (Celarier 1994). In contrast, Indonesia recently tightened the regulations and monitoring of offshore activities and borrowing by banks (1991) and nonbank financial institutions (1995). 275 MC-A-P-TAL FLOWS TO DEVELOPING COUNTRIES Table 5.4 Banking Crisis: Policy Options Palicy option Major advantaWges NYo intervenion No cost to taxpayers. Strengthens market discipline. Financial assistance and recove'y Avoids runs and disruptions in the payments system. Preserves the economic value of bank f ranchising (intangible assets). Takeover and liquidation Strengthens market disciphme. Limits the cost of rescue operations to only insured deposits. Imposes higher losses on bank owners and managers (in the latter case only if it is possible to remove them.) Takeover and assisted sale offailing institution Avoids runs and disruptions in the payments system. Preserves the economic value of bank 1ranchising (intangible assets). Imposes higher losses on bank owners and managers (in the latter ease only if it is possible to rerno ve them). 276 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL, Major disadvantages Main effect of increasedfinancial integration Not feasible or credible, especially when failing institutions are large Induces depositors to put their money into offshore or constitute a large segment of the market. banks and, depending on the exchange rate regime, increases volatility either in capital flows and interest May cause important economic losses if it leads to disruptions in the rates or in exchange rates. credit and payments system. Promotes unfair competition because it induces a de facto "too big to fail" doctrine. Promotes unfair competition (healthy banks have to compete for Increases the potential cost to taxpayers because of funds with failing institutions) and protects managers and the greater possibilities for risk taking and increasing stockholders of risk-taking institutions. difficulties in monitoring. Gives wrong incentives and signals that in the long run will exacerbate moral hazard problems. If left unconstrained, banks can continue investing in risky activities. May induce runs and disruptions in credit and payments if May induce holders of unprotected deposits to go to information leaks, a likely outcome in the case of large institutions. offshore banks and, depending on the exchange rate regime, may increase volatility either in capital flows Destroys the franchise value of banks. and interest rates or in exchange rates. Likely to lead to a "too big to fald" type of behavior; in the end it is In an integrated environment, it may become easier strictly enforced only when the failure involves a small banking to liquidate bank assets. institution, since supervisors are afraid to liquidate large institutions for political reasons as well as to avoid panics and runs. More difficult to impose large losses on bank owners and managers-requires early intervention. Usually leads to protecting all deposits because significant time is Increases the potential cost to taxpayers because of needed to assess the value of assets and work out a sale, a time during the greater possibilities for risk taking and increased which unprotected depositors can withdraw their funds. monitoring difficulties. If no cleanup of banks occurs before privatization, then a repetition It may become easier to implement in an integrated of events leading to insolvency (and future intervention) is likely to environment where foreigners are allowed to buy occur. bank assets and failing institutions. A delayed sale can reduce the value of the failing institution. More difficult to impose large losses on hank owners and managers-requires early intervention. Difficult to implement in developing countries because of shallow markers flr bank assets and institutions, especially in the case of large banks. More expensive to implement when accounting standards are poor and bank liabilities are not accurately reported. 277 AL FLOWS TO DEVELOPING COUNTRIES Crisis management. Crisis management involves allocation of losses and management of failing institutions-whether and how to inter- vene in ailing banks. In an integrated environment, policymakers need to intervene promptly and decisively in ailing institutions in order to contain the crisis and impose losses on the parties responsi- ble-the bank owners and managers (see below). In this regard a set of rigid policy rules, similar to those in the structured early interven- tion and resolution system proposed by Benston and Kaufman (1988) and incorporated into recent U.S. banking legislation (Federal Deposit Insurance Corporation Improvement Act), can be valuable in helping bank regulators to resist the opposition of interest groups affected by the intervention process. However, in many instances a set of rigid rules can have undesirable effects such as forcing the interven- tion and liquidation of viable institutions (Goldstein and Turner 1996). Thus, good judgment and discretion are needed in managing a banking crisis, and in many instances policymakers may choose to fol- low a more heterodox approach rather than a strict application of rules and pure orthodoxy (box 5.8). Intervention strategies. Policymakers have four broad options for managing a banking crisis: (a) not intervening at all, (b) providing fi- nancial assistance to ailing institutions without restraining their activi- ties, (c) taking over and liquidating ailing financial institutions, and (d) taking over and privatizing ailing banks. The first two policy options will, in most cases, aggravate the crisis and exacerbate moral hazard, a problem that becomes even more acute in financially integrated economies and that restricts their applicability even further (box 5.9). The second two policy options are more feasible and are discussed below. * The strategy of taking over insolvent institutions and removing or restraining their management is intended to stop the practice of throwing good money after bad-that is, to pi-event bankers from continuing unsound banking practices with depositors' money that, in turn, is insured by the government. TIhis policy has been implemented in many instances in the past, for example, in Chile and Norway during the 1980s. In doing this ii: is important to act promptly after distress is detected, since delays tend to induce withdrawals of noninsured deposits, as well as additional risk tak- ing by bank managers. One example where prompt action helped 278 THE EFFECTS OF INTEGRATION ON DOMESTIC FINC-A Box 5.8 Orthodox versus Heterodox Bank Regulations PRUDENTIAL BANK REGULATION IS GENERALLY profits in the deregulated environment wiLl allow the structured around three orthodox principles regarding bank to make provisions for the bad loans. But bank capital, supervision, and recapitalization or exit: equally important, most regulatory authorities . Bank capitalustbepsitivendsuffiientto adopt a policy of forbearance because intervention withstand most shocks to a bank's assets. would involve an immediate large financial cost. wtsAnd most shocks to a bank's assets. Low or negative net worth gives banks the incen- • A bank supervisory agency must have the ca- p a ban sUpes a ency must have th banks tve to undertake risky loans and investments, while pabily to collect good inormaton on banKs liberalized access to funds gives them the means to assets and liabilities. ' Bank supervisors must have the power and fi overborrow. During the credit boom that often fol- nancial resources to either recapitalize bank- lows financial liberalization, authorities almost al- rupt banks or to force their liquidation. ways pay lip service to orthodox regulatory policies, even while staffs of bank inspectors are kept small, Underlying this orthodox approach to bank regu- poorly trained, and powerless to alter banks' lending lation is the idea that deposit insurance-which most or provisioning practices. In addition, if financial countries have put in place to safeguard the payments liberalization is accompanied by a trade reform, reg- system, protect smaU savers, and prevent bank runs- ulators will be reluctant to discourage high-risk places government financial resources at risk. Regula- loans to export-oriented activities for fear of derail- tion of bank capital, supervision of bank balance ing the government's reform program. During such sheets, and the recapitalization or forced exit of bank- a period the lack of orthodoxy is primnarily due to ne- rupt banks enable the government to control its expo- glect by the monetary authorities rather than an al- sure to financial risk related to such deposit insurance. ternative set of policies. These orthodox principles of bank regulation When the gamble to let banks grow out of their have often been disregarded in economies that suffer problems fails, the formal commitment to orthodox from financial repression. In a financially repressed banking regulations loses its credibility, and the in- economy interest rates are controLled (often at nega- ability to resolve systemic solvency and liquidity tive real interest rates), portfolio guidelines are rigidly problems with standard measures forces monetary set, and much financial intermediation bypasses the authorities to try to contain the crisis with ad hoc banking system, often in the form of an unregulated policies. Unlfike orthodox policies, these heterodox curb market. The liberalization of a financial system policies vary widely from crisis to crisis, and their marks a movement away from centralized control of only common element is their attempt to distribute financial resources and toward decentralized deci- the costs of the crisis across different groups in the sionmaking by agents in the private sector. economy. Some of the poLicies are aimed at deposi- The move to a liberalized financial system, how- tors: the imposition of capital controls prevents de- ever, often takes place before orthodox banking reg- positors from turning their assets into foreign assets ulations are implemented. Instead, the regulatory (U.S. dollars). Other policies are aimed at debtors: framework is often severely bent to accommodate preferential dollar exchange rates help debtors who the special circumstances surrounding the liberaliza- borrowed in dollars, debt write-downs help firms tion. A common problem during this time is a sig- and mortgage holders, and the swap of bad bank nificant portfolio of bad loans, which can threaten loans for central bank bonds helps banks improve banks' regulatory capital. Regulators are often toler- their balance sheets. All heterodox policies to help ant of such problems because they hope that bank (Box continues on thefollowingpage.) 279 C-APITAL FLOWS TO DEVELOPING COUNTRIES Box 5.8 (continued) debtors and banks involve either a direct cost to the These considerations suggest that policies should central bank or treasury or a postponement of the be specified well in advance of anr crisis. Much of debt payment (or bailout). Chile's 1986 banking law, for example, was moti- The imposition of such policies is essentially an vated by the recognition that a special bankruptcy ad hoc bankruptcy process for much of the private code for the banking system was needed to prevent a sector. lUnlike a formal bankruptcy process, these are recurrence of the disorder that surrounded the 1983 often arbitrarily imposed and implemented with financial crisis. Its intention was to deter risky be- long lags after the crisis begins. The long lags are the havior by specifying the policies that will be pursued result of poor information regarding the extent of in the event of a future crisis. In addition to im- portfolio problems, institutional inability to handle proved supervision and bank capital requirements, large-scaLe crises, and pressure by politically power- that law specifies a number of mechanisms by which ful groups to delay actions that would hurt members bank solvency problems can be resolved without a of the groups. The lags are often associated with government bailout. These mechantisms include two high real interest rates that transfer wealth, in an ac- different voluntary recapitalization plans, as well as a counting sense, from borrowers to lenders. This plan that permits depositors to choose one of two transfer complicates the implementation of policies ways to convert deposits into bank- equity. In addi- to resolve the crisis, since the deadweight costs asso- tion, the law makes detailed provisions for the dis- ciated with the collection of taxes to pay for the posal of mortgages and other bank assets. The law transfer iaise the social cost of the transfer above the was successfully tested in small ways when regulators narrow fiscal cost. The greatest difficulty with ad intervened in the affairs of a meditum-sized bank in hoc policies is that the lack of rules can easily give the late 1980s. rise to expectations that policies will be revised in the fiuture itn the direction of greater debt forgiveness. Prncipal Heterodox Policies Used dnnug Two Such expectations may encourage a mass debtor de- Finacial Crises: Chile and Mexico fault, even by debtors who are in a position to repay. In both Chile and Mexico, for example, programs to Chile aid debtors proved insufficient to prevent the deteri- Debt rescheduling, 1983-88. Twc' separate across- oration cf debtors' net worth and later had to be re- the-board debt rescheduling programs for firms and vised with more generous terms (see below). one across-the-board rescheduling lor mortgage and to reduce the cost of crisis is Malaysia in the mid-1 980s. There the authorities devised rescue packages for the ailing financial institu- tions early in the crisis, in 1985, reducing total losses to only 2.4 percent of total deposits, and allowing the economy to start recov- ering only two years after the crisis started--in 1987.28 In con- trast, analysis of the 1980s savings and loan crisis in the United States, for example, suggests that delaying the closure of troubled institutions, by up to 38 months on average, doubled the cost of resolving the crisis (Dellas, Diba, and Garber 1 996). In the case of 280 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIA1`- consumer loans. The April 1983 programn proved in- into UDI (unit of investment, or CPI-indexed) loans. adequate to solve the problem and was expanded in The UDI loans involve an interest subsidy to banks June 1984. Some loans were rescheduled again on a and can be carried at book value even though they case-by-case basis by individual banks in 1986. might be sold at a loss. Some mortgage loans were partially forgiven (up to Direct subsidies to borrowers, 1995-96. Under 25 percent) in 1988 through a special refinancing fa- the September 1995 ADE program, created to cility set up by the central bank, neutralize the growing political importance of an Purchase of risky loans by the central bank, organized debtors' movement, credit card and 1982-87. The central bank purchased bad loans at mortgage obligations receive a direct interest rate par for up to 150 percent of capital and reserves of subsidy from the government for a year. The each bank. In exchange, banks were given central original ADE program failed to resolve growing bank bonds indexed to the consumer price index arrears and was consequently expanded in April that paid 7 percent in real terms. Banks were re- 1996. quired to repurchase these loans out of earnings Dollar loans, 1995. Following the December until 1996, when the central bank began negotia- 1994 devaluation of the peso, commercial banks in tions that are expected to forgive about $1.5 billion Mexico had great difficulty rolling over dollar- of the outstanding debt. denominated certificates of deposit. In response, the Dollar loans, 1982-85. The central bank estab- Mexican government established a special discount lished a preferential exchange rate for the repayment window for dollar loans to banks at penalty interest of dollar-denominated loans. This was a costly res- rates of25 percent. cue program for the central bank. Recapitalization, 1995-96. Following an initial Recapitalization, 1985-87. Through a program temporary recapitalization program, PROCAPTE, called Capitalismo Popular, the government subsi- implemented in February 1995, FOBAPROA (the dized the purchase by small investors of new equity Bank Liability Protection Agency) instituted a loan capital in banks that had undergone intervention. purchase/capitalization program for the eight banks that had undergone intervention. In the pro- Mexico gram FOBAPROA purchases two dollars of nonper- Debt restructuring, 1995-96. Commercial, mort- forming loans for each dollar of new capital gage, and consumer loans have been rescheduled infusion. developing countries, where resolution costs appear to be higher (box 5.1), it becomes even more urgent to act promptly to contain the crisis, especially in an integrated environment, where bank managers have a number of channels through which they can un- dertake risky investments or commit fraud. Indeed, this occurred in the case of eight financial institutions that failed in Venezuela in the 1994 banking crisis, after the collapse of Banco Latino (box 5.9). In addition, in an integrated environment financial instabil- ity may lead to capital flight, which can increase macroeconomic 281 -C-APITA1L FLOWS TO DEVELOPING COUNTRIES Box 5.9 Infeasible Policies in Banking Crisis Management A POLICY OF NO INTERVENTION GENERALLY banking crisis in Venezuela (after tle collapse of lacks credibility-and is therefore ineffective-un- Banco Latino), which involved eight financial insti- less thecrisis is not systemic and involves only small tutions, is instructive. Because the crisis occurred financial institutions whose failure does not threaten only weeks before a new administration took office, the stability of the financial system. Moreover, inter- the authorities provided financial assistance to the vention may be inevitable-even if it affects only ailing institutions but left any further action to the small institutions-if there is in place a deposit in- new administration. The delayed response in con- surance mechanism that allows banks to undertake trolling bank actions exacerbated fraud and caused risky investments with depositors' money and poli- larger losses. cymakers aim to minimize the cost of crisis to the Thus, from an efficiency point of view the latter taxpayers. In a financially integrated environment a policy can be justified only if the ailing banks are policy of nonintervention will lead to capital flight econotnically viable-have a positive net worth-so whenever depositors suspect that even a small-scale that bankers have the incentive to invest the funds in financial crisis may occur. This, in turn, will lead to a sound manner. For this to occur, intervention higher interest or exchange rate volatility-depend- must take place at an early stage, before banks' net ing on the exchange rate system-which will exacer- worth evaporates or becomes negative. On the other bate macroeconomic vulnerability. Moreover, a hand, from an equity point of view this policy can policy of nonintervention is self-defeating, since it only be justified if bank problem,s are due to sys- will lead in the medium term to the establishment of temic risk and not to poor management; otherwise large banks because markets tend to act as if the too- this policy will exacerbate moral hazard and discrim- big-to-fail doctrine prevails. inate against sound and well-managed banks. Finan- A policy of providing financial assistance-for cial integration increases the potential losses that example, through emergency (soft) loans-to banks may result from providing cheap financial assistance in distress without directly intervening in their oper- to ailing banks without intervention. This highlights ations, gives a competitive advantage to risky banks the need for good information so that bank supervi- by subsidizing them and protecting the interests of sors can quickly discriminate between solvent and their owners and managers. In addition, if no con- insolvent institutions and provide financial assis- trol mechanisms are established for the use of the tance to only the former group. Nonviable financial emergency funds, this policy may increase risky institutions, on the other hand, reed intervention, lending and the associated losses. In this regard the and their lending and risk-taking activities must be experience during the second stage of the 1994 restricted. vulnerability. In taking over insolvent institutions, bank regulators should seek to remove managers as a way oF imposing on them part of the cost of the crisis (see below). Nevertheless, regulators may be unable to do so if manpower is scarce and there are no al- ternative teams capable of managing the ailing institutions (De Juan 1987). In this case, bank managers should be severely regu- lated and closely monitored, while perquisites and dividends should be drastically curtailed. 282 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL * The decision to liquidate or rehabilitate and sell an ailing finan- cial institution should be based on the costs and benefits of each course of action. In general, the main benefit of rehabilitation and posterior sale-privatization-is that it preserves goodwill, while liquidation implies losing the economic value of the insti- tution's intangible assets. Rehabilitation, therefore, may be more valuable in the case of large financial institutions with a large net- work of branches and some degree of specialization and know- how. On the other hand, the main advantage of liquidation is that it limits the cost of deposit insurance to only the amount of deposits covered by the insurance mechanism. Because rehabilita- tion and posterior sale is a lengthy process-especially if it in- volves a large and specialized financial institution-it usually leads to the protection of all deposits to prevent a withdrawal of funds and a disruption in credit and payments. In addition, reha- bilitarion and sale may be more difficult in developing countries, which lack specialized markets to dispose of financial institutions, while a delayed sale may destroy some of the economic value of the intervened bank. Liquidation, too, can be problematic, espe- cially since in a financially integrated economy it can lead to cap- ital flight and increased volatility in either exchange or interest rates, exacerbating macroeconomic vulnerabilities. On the other hand, the privatization of ailing institutions can be facilitated in an integrated environment if foreign banks can compete in the bidding process. Therefore, increased financial integration may bias the resolution of banking crises toward intervention and pos- terior sale of ailing banks. For this outcome to be preferable to liquidation, however, it is important to minimize the losses that result from risk-taking activities, which calls for early interven- tion and prompt action once problems are detected. Although these conclusions are generally valid, they need to be consid- ered or analyzed with caution. For example, rushing the privatization of banks may create fuiture troubles if the buyers are not adequately screened. The purpose should not be to sell quickly but to sell to good bankers. Therefore, it is worthwhile to take the time to assess the new owners. This is illustrated by the recent experience in Mexico, where in the early 1 990s state-owned banks were sold to inexperienced bankers who in many in- stances incurred large debts to finance their purchases (box 5.6). In an in- 283 -CA2ATAL FLOWS TO DEVELOPING COUNTRIES tegrated environment the issue of carefully screening market participants acquires special relevance because of the potential for greater losses and the larger pool of market participants, including foreign-based institutions, which may enter the banking sector without the appropriate skills or in- centives. The recent establishment of Russian-backed banks in some countries in Eastern Europe and the former Soviet UJnion-such as Hun- gary and Latvia-has been a matter of concern, foi example, because in many cases these are only pocket banks of large state-owned enterprises and lack the expertise to function in a modern integrated economy (Flem- ing and Talley 1996; Fleming, Chu, and Bakker 1996).7 Also, because the restructuring and privatization of banks has proved to be unsuccessful-leading to second rounds of intervention and financial rehabilitation-when the macroeconomy has been in dis- array, a macroeconomic stabilization program should be implemented prior to or at the same time that bank restructuring starts. Given the greater market response and the increased number of channels through which capital outflows can occur in an integrated economy, more em- phasis and higher priority must be given to achieving and maintaining macroeconomic stability when managing a banking crisis and trying to rehabilitate failed banks. In the same vein, because of the potentially greater negative effects of financial distress on the inacroeconomy in an integrated environment, a quicker response is needed to correct the un- derlying financial causes of the crisis. Allocation of losses: guiding principles. The allocation of losses in a banking crisis is a difficult political problem. There are four groups that can be targeted: stockholders and bank managers, depositors, borrow- ers, and taxpayers. * Bank owners and managers should suffer maximum losses-their institutions should be taken over and they should be removed from management. As argued earlier, however, this may be diffi- cult to implement in countries where banking skills are scarce and supervisors lack experience in banking. Moreover, to effectively impose losses on owners and managers, intervention must take place at an early stage-before losses reach several times the insti- tution's capital and the owners (using depositors' money) have moved funds offshore while managers have been overpaid. Im- posing losses on owners and managers in an integrated environ- ment requires early detection and prompt action. 284 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCAL NC * Depositors rarely bear a part of the losses, especially in the case of large banks (the "too-big-to-fail" syndrome), although there are exceptions to this rule. For example, in C6te d'Ivoire in 1991 de- positors lost about 15 percent of their deposits, in Chile during 1982-84 depositors in banks that were liquidated (small banks) lost 70 percent of theirs, and in Estonia in 1993 depositors lost about 40 percent of their deposits.30 It is politically costly to im- pose losses on depositors, although ideally governments would like to impose losses on large depositors to enhance market disci- pline. Trying to tax depositors in an integrated environment, however, will most likely lead to capital flight and exacerbate macroeconomic and financial problems. * Borrowers should be targeted to the extent that seeking repay- ment from them does not distort their incentives to invest and undertake new projects. They should not be left with a debt overhang problem that would deter new investment and cause economic growth to decline (box 5.1). Alternatively, in an inte- grated environment borrowers may seek to start new businesses offshore. * Taxpayers will bear less of the cost of a crisis the sooner the crisis is detected and contained. The cost to taxpayers results from the need to restore confidence in the market by bailing out borrowers and protecting depositors. The difficulties encountered in man- aging a banking crisis-for example estimating its actual cost and resolving incentive and other institutional problems-tend to delay policy responses and increase market uncertainty. There- fore, a prompt policy response will reduce the cost of the crisis shared by taxpayers. To the extent possible, however, the tax rev- enues raised to pay for the bailout should not deter economic growth in the medium and long term-for example, the author- ities should avoid increasing tax collection through inflation. In an integrated environment the allocation of losses will most likely be biased toward the less protected and informed groups-small de- positors, small shareholders, and taxpayers-since large depositors, bank owners, and managers will often be able to move funds offshore. Thus, the need to intervene promptly is even more urgent if the objec- tive is to impose maximum losses on the largest market participants, those who have greater incentives to be well informed and for whom it 285 AŽTAL FLOWS TO DEVELOPING COUNTRIES is cheaper-proportional to their investment-to move funds from onshore to offshore locations. In sum, financial integration puts a premium on the need for good and reliable information, prompt decisionmaking, and swift corrective action by bank supervisors. Given the greater volatility and responsive- ness of financial markets in an integrated environment, and the poten- tial for larger losses, the need to detect and contain a banking crisis at an early stage becomes even more urgent. Special attention must be paid to offshore and off-balance transactions and to the screening of potential bankers, as well as to the need for econoniic authorities to put in place a contingency plan and provision adequate funding while, at the same time, maintaining macroeconomic stabiLity. Indeed, analysis of past episodes shows that lack of appropriate funding has led bank su- pervisors not to intervene but to let distressed financial institutions re- main in business, a course that in the end increases the cost of crisis. A sound macroeconomic stance and other aspects of bank regulation be- come even more important when managing a banking crisis under more stringent conditions, such as in the presence of a currency board. Two recent experiences of banking crises under a currency board, in Ar- gentina and Hong Kong, are briefly discussed in annex 5.4. 286 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANIN Annex 5.1 Capital Inflows, Lending Booms, and Banking Crises: Country Episodes T X rHE ANALYSIS IN THIS CHAPTER IS BASED Several features of the country episodes analyzed on a subsample of the country episodes here are worth discussing. First, in all the episodes- studied in previous chapters. In particular, except Sweden during 1989-93-the country in- we study a set of countries that have experienced a volved received significant private capital inflows surge in private capital flows at some point during (measured in terms of GDP). Indeed, the sample aver- the past 20 years while, at the same time, attempt- age for all the episodes (including Sweden) is almost 5 ing to implement reforms in their financial sectors. percentage points of GDP. Second, more than half of Country episodes were selected on the basis of the the country episodes ended in a banking crisis (14 out importance of the banking sector in determining of 20). This validates the point argued earlier that the the outcome of the financial liberalization and outcome of the liberalization and integration experi- integration experiment. In other words, the sample ment was largely affected by developments in the includes country episodes in which the banking banking sector. Third, economic growth was signifi- sector has played a significant role in the observed cantly reduced in the years following a banking cri- outcome. The sample selection was partly influ- sis-except for the cases of Brazil during the 1 990s and enced by the attention that each country experi- Thailand during the early 1980s.31 Moreover, in most ence has attracted in the relevant literature. The cases economic growth during the years following a sample comprises both industrial and developing banking crisis was either negligible or negative. Fourth, countries, and in the case of some developing in all the cases where a banking crisis occurred, credit countries more than one episode is analyzed. In to the private sector-from either banks or other fi- particular, we include episodes occurring in the nancial institutions-grew in real terms at very high early 1 980s, prior to the debt crisis, and compare rates in the years before the crisis, with rates of growth them with those occurring in more recent years. usually several times that of real output. Furthermore, The list of country episodes included in our sam- these periods of high growth were followed by years of ple is shown in table 5.5, on the next page. very low or even negative credit expansion. 287 CAPITAL FLOWS TO DEVELOPING COUNTRIES Table 5.5 Capital Inflows, Lending Booms, and Banking Crises, Selected Countries and Years Average percentage ofgrowth Bank Nonbank credit Bank credit Nonbank GDP (inflow credit (inflow credit (infloow GDP Inflow period (crisis period (crisis period (crisis as a Banking or and or and or and Inflow percentage crisis precrisis postcrisis precrisis postcrisis precrisis postcrisis Country period of GDP years years) years) years) years) years) years) Argentina 1979-82 1.98 1980-82 14.64 13.19 6.57 -20.38 4 36 -1.79 1992-93 4.03 1994-95 16.94 14.86 45.56 24.91 7.34 1.34 Brazil 1992-94 2.19 1995 52.57 - 17.41 - 3.00 3.35 Chile 1978-81 12.68 1981-83 43.26 10.29 71.55 -33.22 8.09 -3.45 1989-94 5.48 no crisis 7.46 - 21.79 - 6.96 - Colombia 1992-94 3.84 no crisis 15.23 - 15.15 - 5.03 - Finland 1987-94 4.18 1991-93 11.99 -7.67 - - 3.65 -1.94 Indonesia 1990-94 1.22 no crisis 18.59 - - - 6.92 - Malaysia 1980-86 6.66 1985-88 16.53 6.32 23.41 7,18 6.87 3.52 1989-94 9 75 no crisis 12.76 - 11.92 - 8.75 - Mexico 1979-81 5.27 1982-83 14.58 -25.19 -1.82 -11.38 8.47 -2.42 1989-94 5 15 1994-95 33.47 - 15.16 - 3.06 -1.79 Norway 1984-89 2.04 1988-89 18.71 2.22 0.02 1.49 4.29 0.19 Philippines 1978-83 4.57 1982-87 9.18 -10.84 12.48 -22.95 4.95 -0.37 1989-94 4.17 no crisis 11.92 - 18.67 - 2.56 - Sweden 1989-93 0.77 1991-93 7.54 -11.98 28.67 -1.96 1.87 -1.77 Thailand 1978-84 4.42 1983-87 7.92 14.31 9.43 7.55 6.34 6.19 1988-94 9.34 no crisis 21.35 - 21.43 - 10.01 - Venezuela 1975-80 7.81 1980 15.13 -2.14 16.99 3.31 4.97 -1.84 1992-93 2 65 1994-95 -7.97 -26.81 -13.45 -31.95 3.13 -0.32 - Not avsilable. Note: Comparisons across countries are not valid because periods differ in length. Source: MF, International Financial Stawstics data base, World Economic Ouldook data base, Caprto and Klingebiel (1996); Kaminsky and Reinhart (1996). 288 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIC1kY Annex 5.2 Government Guarantees, High Real Interest Rates, and Banking Sector Fragility A FINANCIAL SYSTEM INTERMEDIATES gamble on future economic growth. This can occur funds between savers and investors, deter- because a guarantee operates in the same way as a mines the quality of investments undertak- put option on a bank's assets: if the value of the en in the economy, and provides the means of pay- bank's assets falls below the value of its liabilities, the ment for many transactions. When the financial guarantee makes up the difference. As with a put system is working well, these functions increase option, an increase in the riskiness of the asset port- output in the short run, as well as the level and folio raises the value of the guarantee to bank own- productivity of the capital stock in the longer run. ers (and possibly to bank executives), since greater The functions of the financial system are, in fact, risk increases the potential for large up-side profits so important that virtually all governments provide while the government guarantee insures depositors explicit or implicit guarantees to participants in the against downside losses. Without a countervailing system. These guarantees may simply take the form force, therefore, guarantees will be a source of finan- of protection to otherwise solvent banks during a cial instability, since banks will end up attempting run on deposits, without protecting depositors in to maximize the value of the guarantees. the event that banks turn out to be insolvent The incentives for risk taking associated with de- (lender of last resort). Or the guarantees may be posit guarantees can be counteracted by ensuring extended to small savers, on the assumption that that the value of a bank as an ongoing operation, in- they are not sophisticated enough to monitor bank cluding the value of specialized human capital, is owners and managers (limited deposit insurance); positive. When a bank has high franchise value and a or they may cover banks whose failure is judged to healthy balance sheet, it will generally protect itself pose a threat to the financial system. This "too big against risk, since the value of the government's guar- to fail" guarantee usually covers all the liabilities of antee to a banker is relatively small. If, on the con- larger banks while leaving liabilities of smaller trary, the franchise value of the bank decreases, then banks at some risk. Often governments extend-at bankers will pursue risky strategies that depend on least implicitly-blanket guarantees for all banks, the government guarantee. When banks have an in- which shift private risk associated with bank liabili- centive to maximize the value of the guarantee, the ties to the government. result can be a banking crisis that may involve a fis- cal cost of up to 8 percent of GDP, as happened re- Guarantees cently in Mexico. This cost, moreover, can easily be doubled by the deadweight costs of taxation, bank- The advantage of government financial guaran- ruptcies, and the interruption of new lending. Given tees is that they protect the payment system, prevent the large potential cost associated with government bank runs, and protect depositors against losses. guarantees, therefore, it is important for policymak- The dark side of these guarantees is that they may ers to keep track of variables that may signal an envi- promote risk taking and give overambitious bank ronment in which banks have the incentive and owners access to financial resources with which to opportunity to maximize the value of the guarantee. 289 IT~ITAL FLOWS TO DEVELOPING COUNTRIES Interest Rates thermore, if the real deposit rate (the nominal rate adjusted by actual inflation) appears low, while the A substantial rise in real interest rates almost always deposit rate spread appears high, there has usually results in a cleterioration of bank balance sheets be- been an appreciation of the real exchange rate com- cause of the term transformation done by banks- bined with expectations of a future discrete depre- deposits are short term, while bank assets are longer ciation, the so-called peso problem. The peso term-and because higher real loan rates cause an problem is important because expectations of fu- increase in nonperforming loans. If the high real ture exchange rate depreciation will drive up inter- rates are onlv transitory, then healthy banks will not est rates even when the government has no be unduly affected. But if the high rates persist, intention of making a sharp exchange rate adjust- banks' net w7orth may be so compromised that the ment. Such high rates, if mairtained over a long banks try to grow out of their problems with risky period, will erode the net worth of both banks and loans and investments. To understand why interest companies, pushing bankers to behave according to rates can be high in a country, it is useful to look at the incentives created by the dark side of the gov- two primary, interest rate spreads: the deposit rate ernment deposit guarantees. For example, some an- spread (the cdifference between the nominal domes- alysts have argued that fixing the exchange rate tic deposit rate and the sum of the relevant nominal signals an implicit government commitment to international rate-usually the three-month U.S. promote and guarantee the safety of long-term in- treasury bill rate-and the actual change in the ex- ternational borrowing at that exchange rate. This change rate), and the loan rate spread (the difference implicit guarantee makes the open capital account between the loan rate and the deposit rate). The look inviting to foreign lenders and may encourage table below separates the deposit and loan rate overborrowing in dollars when domestic interest spreads into their respective parts and shows the rates are high. economic reasons for these spreads. Remaining factors that raise the deposit rate Typically, a high deposit rate spread is accounted spread, grouped under the heading of deviations for by foreco sting errors resulting from a prediction from interest rate parity, can be caused by exchange of domestic inflation that turns out to be too high rate risk, country risk, or imperfect capital mobility. or expectations of exchange rate depreciation. Fur- In addition, the risk that the government may not Deposit rate spread id - it- Ae Macroeconomicfactors Expected depreciation Lack of monetary/exchange rate credibility Deviations from interest rate parity Exchange rate risk, imperfect capital mobility, de- fault risk associated with government debt, guaran- tor risk associated with banking system Loan rate spread I -id Microeconomicfactors Taxes on intermediation Reserve requirements, directed credit programs Net spread Default risk, industry structure Note 'd = deposit rate; i*= international rate; Ase = change in the exchange rate, zs= loan rate. 290 THE EFFECTS OF INTEGRATION ON DOMESTIC FINAN-C stand by its implicit, or even explicit, guarantees to An Illustration of Spread Analysis the banking system will increase the differential be- tween the domestic deposit rate and the interna- The analysis of interest rate spreads is most useful tional rate of interest. For example, if there is a when dollar deposit rates exist. In that case expected perception that the government is likely to render depreciation can be approximated by the difference dollar-denominated deposits inconvertible, the in- between the domestic currency interest rate and the terest rate spread on dollar-denominated deposits dollar deposit rate. Deviations from interest rate will rise to reflect that probability. parity will then be picked up by the difference be- The loan rate spread can be decomposed into tween the dollar deposit rate and the three-month two components, the first due to taxes on financial treasury bill rate. intermediation, and the second, the net spread, due The behavior of interest rates in Chile during its to the banking system itself-its structure, its costs, transition to democracy in 1989-90 can be used to and its information and incentive structure. The net illustrate the analysis of interest rate spreads. In spread reflects the costs of banking. When banks Chile many of the events of 1989 revolved around have been sheltered from international competition the upcoming December election, in which a candi- for many years, net spreads are generally high. date approved by the military was opposed by an al- Opening the capital account will tend to lower net liance of various parties. The alliance won and took spreads but may also cause formerly profitable power in March 1990. Between March 1989 and banks to become unprofitable and, therefore, to un- February 1990 the annualized nominal deposit rate dertake risky activities, as noted above. rose from 12.5 percent to 45 percent, and the nom- To summarize, sustained high real deposit rates inal loan rate rose from 19.5 to 57 percent. This and loan rates are a strong signal of future trouble rapid rise in interest rates appears to have reflected because such rates are usually accompanied by a rise concerns about inflation as a result of expansionary in nonperforming loans. In an integrated environ- macroeconomic policies during 1989, as well as un- ment, when high real rates and weak bank balance certainty about the economic policies of the new sheets are matched with government guarantees government. Interest rates remained high during that permit banks to borrow internationally, then most of 1990 and did not come down until the be- the stage is set for a credit boom in which rapid ginning of 1991. The deposit rate spread became credit expansion both masks underlying portfolio positive in July 1989 and stayed positive through problems and creates the opportunity for high-risk January 1991, reaching a sustained level of around loan strategies. During such a boom-as occurred 20 percent during most of 1990. in Chile in the late 1970s and early 1980s, the U.S. During this period, the central bank was at- savings and loan industry in the 1980s, and Mexico tempting an uneasy tradeoff between fighting infla- in the early 1990s-it is difficult not to get caught tion and preventing an appreciation of the real up in the euphoria associated with the economic exchange rate. This tradeoff was especially marked growth financed by bank credit. But if the boom re- at the beginning of 1990, when the central bank flects the dark side of government banking guaran- raised interest rates on its long-term indexed debt tees, then there will be a fiscal reckoning at the from 6.9 to 9.7 percent in order to dampen aggre- boom's end. gate expenditure. This high interest rate policy pro- 291 `CA LTAL FLOWS TO DEVELOPING COUNTRIES Figure 5.10 Exchange and Interest Rates in Chile, 19B8-91 A. Expected and Actual Percentage Changes In the Exchange Rate Percent (annual) 70 - - 6 |----- Actual change in the exchange rate 60 l____ Expected change In the exchange rate 50--,1\ 40~~~~~~~~~~~~~ 30 -10 -20 1988 1989 1990 1991 S. Dollar Interest Rates and the T-bill R'ate Percent (annual) 14 12 8 4 _ - Dollar loan rate 2 - ' Dollar deposit rate - T-bill rate 0 1988 1989 1990 1991 Source Brock (1996). 292 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANC IAU$ duced an immediate appreciation of the currency Figure 5.1 OB shows the position of dollar interest (to the bottom of the central bank's preestablished rates in Chile relative to the T-bill rate. It is easy to exchange rate band) as well as short-term capital in- detect the upward shift in the dollar loan rate rela- flows which the central bank sterilized. tive to the T-bill rate that took place at the end of In terms of the spreads, figure 5.10A shows the 1989. This upward shift was initially sustained by difference between the expected change in the ex- the central bank's sterilization policy and then later change rate (calculated from the peso and dollar de- by the reserve requirement on capital inflows. In posit rates) and the actual annualized change in the terms of the spread analysis, this was a deviation exchange rate. The figure shows that from July 1989 from interest rate parity caused by imperfect capital to January 1991 the expected change in the ex- mobility. Note, however, that the size of this com- change rate exceeded the actual change, often by 30 ponent of the deposit rate spread was very small percentage points on an annualized basis. During compared with the difference between the expected this period market participants were betting that the and actual exchange rate shown in figure 5.1 OA. central bank would ultimately reverse its inflation- Chile 's high interest rate policy also affected the dampening policy and actively depreciate the ex- loan rate spread. In 1988 the loan rate spread aver- change rate in order to prevent a real appreciation. aged 6.06 percent, while it rose to 8.19 percent in The expectation proved correct; the exchange rate 1989 and 8.56 percent in 1990. This increase, had was allowed to depreciate at the end of 1990. After it continued, could have increased the fragility of this episode, the monetary authorities revised their the banking system, but following the exchange rate policy in two ways: first, the nominal exchange rate depreciation at the end of 1990 and the change in was allowed to appreciate in several discrete steps; stabilization policies by the central bank, the loan second, a 20 percent unremunerated reserve require- rate spread fell to 6.23 percent in 1991, or approxi- ment was imposed on capital inflows in June 1991. mately its level in 1988. Annex 5.3 Risk-Adjusted Capital-Asset Ratios: The BIS Classification of Risky Assets I N DECEMBER 1987 THE BASLE COMMITTEE OF obliged to impose a minimum risk-adjusted capital- bank supervisors, operating under the auspices of asset ratio of 8 percent on all banks operating under the Bank for International Settlements (BIS), pub- their jurisdiction by the end of 1992. Supervisory lished proposed guidelines for the measurement and authorities in each country were given significant dis- assessment of the capital adequacy of banks operating cretion in the interpretation and implementation of the internationally. The guidelines were approved in July new rules. 1988 by bank supervisors of 12 industrial countnes The main purpose of the new capitalization rules comprising the Group of Ten (G-10) plus was to put internationally operating banks under Luxembourg and Switzerland.32 Under the July 1988 common regulatory conditions to prevent unfair agreement, bank supervisors of these nations were competition by banks with lighter regulatory bur- 293 _rrAL FLOWS TO DEVELOPING COUNTRIES dens. The new capital-asset ratio guidelines ad- m They initiated a procedure to equate off-bal- dressed only one aspect of banking regulation, how- ance-sheet items (standby letters of credit, ever: the capital requirements needed to protect swaps, options, futures contracts, and other depositors against credit risk. The guidelines ac- contingent assets or liabilities) to asset equiva- cordingly dealt with the identity of banks' debtors lents based on the type of item and the initial but did not address other sources of risk, such as in- contract terms. terest rate risk, exchange rate risk, equities risk, or banks' overall portfolio risk-all of which implies Risk-based capital requirements under the Basle taking into account the correlation among different Accords are generally expressecL as: ACB > 0.08 types of bank- assets and liabilities. Furthermore, the TOWRA, where ACB and TOIVRA stand for ad- new capitalization rules generally did not advance justed capital base and total weighted risk assets, re- the use of market prices to evaluate bank assets and spectively. ACB is the sum of allowable components liabilities.33 of primary (tier 1) and secondary (tier 2) capital, Although the Basle Accords concerned only in- subject to prescribed limits and deducting certain ternational banks and had other limitations, the items. For example, goodwill is excluded from pri- capitalization guidelines have been voluntarily mary capital, and term-subordinated debt included adopted by an increasing number of countries in secondary capital cannot exceed 50 percent of around the world, and have been extended to the primary capital. TOWRA, on the other hand, is the regulation of other financial institutions. For exam- weighted sum of on-balance and off-balance items ple, the European Union extended the guidelines to as shown in the following expression: aLl credit institutions in 1989, and the United States S S w I I has extended them to bank holding companies and TOWIRA (A, W, ) + , (B,, Xk W,) their subsidiaries. The major innovations concerning capital re- where A is the value of the ilth asset with risk weight quirements introduced in the accords are: W and B1Jk is the notional principal amount of off- balance-sheet activity i with risk weight W and con- * The accords assigned percentages for weighting version factor X*. Table 5.6 surnmarizes the main each asset category according to its credit risk. risk categories for on-balance sheet items as recom- * They redefined the composition of a bank's mended in the Basle Accords and the conversion fac- primary capital by, for example, excluding tors for off-balance sheet transactions as applied by general loan loss reserves. Japan, the United Kingdom, and the United States. 294 THE EFFECTS OF INTEGRATION ON DOMESTIC FINAIA Table 5.6 BIS Risk Categories of Bank Assets BIS-recemmended rik weit (paernt) On-balnce asset item 0 Cash, gold, or loans to or fully guaranteed by OECD central governments and central banks; daims fully collateralized by cash; loans to or fiully guaranteed by non-oECD central governments or central banks when denominated and funded in local currency. Less than 20 Holdings of fixed interest securities issued or guaranteed by OECD central governments and float- ing rate or index-linked OECD central government securities; daims fully collateralized by OECD central government fixed interest securities and similar floating rate securities; holdings of non- OECD central government securities when denominated and funded in local currency. 20 Claims on multilateral development banks (or claims fully guaranteed by or fully collateralized by the securities issued by these institutions), daims on credit institutions incorporated in the OECD countries and claims guaranteed or endorsed by oEcD-incorporated credit institutions, claims on or guaranteed by non-OECD incorporated credit institutions when they have a residual maturity of up to one year, claims on or guaranteed by OECD public sector entities (exduding the central gov- ernment), cash items in the process of collection. 50 Loans fully secured by mortgage on residential property owned or rented out by the borrower. 100 Claims on the nonbank private sector, claims on credit institutions incorporated outside the oEcn with residual maturity of more than one year, claims on or guaranteed by non-OECD gov- ernments and central banks that are not denominated in local currency and funded locally, claims on public sector enterprises, fixed assets, capital instruments issued by other banks, real estate and wa4e investment and all other nonspecified assets. Cenemien factor appled in Japan, the United Kndm, and the Untd States (pemnt) Off-balance-seet transaction 100 Direct credit substitutes, including general guarantees of indebtedness, standby letters of credit, acceptances, and endorsements; sale and repurchase agreements and asset sales with recourse where credit risk remains with the bank; forward agreements to purchase assets, including finan- cial facilities and commitments with certain drawdown. 50 Certain transaction-related contingent items not having the character of direct credit substitutes (performance bonds, bid bonds, warranties, and standby letters of credit related to particular transactions); note issuance facilities and revolving underwriting facilities; other commitments (such as credit lines) with original maturities of more than one year. 20 Short-term, self-liquidating, trade-related contingent items (for example, documentary credits collateralized by the underlying shipments). 0 Similar commitments with original maturity of up to one year or commitments that can be unconditionally canceled at any time, endorsements of bills that have previously been accepted by a bank. S&nr. Hall (1994). 295 WIPITAL FLOWS TO DEVELOPING COUNTRIES Annex 5.4 Crisis Management in a Constrained Setting: The Case of Currency Boards A CURRENCY BOARD IS A MONETARY REGIME if a run occurs, tend to withdraw their money as in which the domestic authority is pre- soon as they suspect that a large number of other de- clu ed by law from issuing liabilities- positors may do the same. In this context, an other- high-power money-unless the Issue is backed by wise solvent bank may become insolvent if it is an equivalent amount of international reserves. unable to contain a run on its deposits and is forced Under this system the monetary authority basically to sell assets at a large discouni34 Thus, banks are relinquishes its right to grant credit to the public or intrinsically unstable institutiorns subject to self-ful- private sector, and therefore, fiscal imbalances can- filling confidence crises that may lead them into in- not be financed through an inflation tax. Although solvency (Diamond and Dybvig 1983). a currency board works as a fixed exchange rate sys- To overcome the problem of self-fulfilling confi- tem, it is more credible because a fixed exchange dence crises in banking, countries rely on a lender of rate system can always be abandoned and is subject last resort to provide emergency loans to banks and to discrete dlevaluations of the currency, so there is guarantee full convertibility of c[eposits into domes- no guarantee that the value of the domestic curren- tic currency at face value; they thereby prevent some cy in terms of foreign currency will remain con- of the economic costs associated with bank crises stant through time. The main goal of a currency that can affect otherwise solvent institutions. Al- board, by contrast, is precisely to guarantee the full though the lender of last resort can be any public convertibility of domestic into foreign currency at entity capable of lending to troubled financial insti- a constant rate. tutions, in practice this function in most countries The main purpose of a lender of last resort, on is performed by the domestic monetary authority, the other hand, is to guarantee the full convertibil- which does it by creating money. Therefore, unless ity of a subset of bank liabilities-usually those with a foreign government, central bank, or (multilat- the shortest maturities-into domestic currency at eral) credit institution is willing to provide a credit their face va] ue. Having such a mechanism in a frac- line in foreign currency, a curl ency board will be tional reserve banking system prevents bank runs limited in the amount that it can lend to domestic and liquidity crises. Such crises could reduce the banks during a liquidity crisis. [n particular, it will value of bank assets and lead to insolvency, causing be able to extend credit to banks only if it holds re- disruptions in the payment and credit systems and serves in excess of the amount required to guarantee great econorriic losses. Bank runs and liquidity crises the value of the domestic currency at the established lead to insolvency because bank assets are of longer exchange rate. In sum, the adoption of a currency maturity than bank liabilities, and because of asym- board implies that the monetary authority relin- metries of information in banking that make bank quishes its role as a lender of last resort, and unless assets illiquid (Diamond 1984). Therefore, in a fire an alternative entity assumes that role and provides sale bank assets are usually heavily discounted, caus- the same type of insurance to bank depositors, the ing important losses for the selling bank. Bank de- banking system will become more vulnerable to liq- positors, whn understand the risk of suffering a loss uidity crises. 296 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAt1 The experiences of Hong Kong during the mid- The banking crisis in Argentina started in late 1980s, and more recently of Argentina during 1994, when depositors began withdrawing funds 1994-95, provide interesting cases in which a bank- from small and provincial banks and making de- ing crisis occurred under a currency board (a quasi- posits in larger banks (which were perceived as less currency board in the latter case). In both cases the vulnerable or "too big to fail"), a shift that led to the local monetary authority was limited in the amount bankruptcy of several small financial entities and of credit it was permitted to extend to banks having the crisis in the provincial banks. Initially the shift financial difficulties, and in the end had to resort to affected mainly peso-denominated deposits, reflect- alternative mechanisms to contain the crisis. ing a loss of confidence in the currency, but as the In Hong Kong several financial institutions faced crisis deepened it also affected dollar-denominated difficulties starting in 1982, because of the slow- deposits, reflecting lack of confidence in the bank- down in economic activity and the dive in stock and ing system. The crisis resulted in total withdrawals property prices from their peaks in 1980-81 (prop- of about US$8 billion between December 1994 and erty prices fell between 60 and 90 percent and stock April 1995, equal to about 16 percent of total bank prices by 50 percent). In addition, the uncertainty deposits, and the central bank's losing about US$4 surrounding the Sino-British talks on the political billion in international reserves. In addition, the future of Hong Kong, and the currency crisis that number of financial institutions decreased from 201 unfolded in 1983, led to the reintroduction of the in December 1994 to 157 a year later, mainly currency board system, with the Hong Kong dollar through mergers and acquisitions. At the same time pegged to the U.S. dollar, in October of that year.35 there was a shift toward U.S.-dollar-denominated However, increasing financial difficulties in several deposits in large banks. Thus, between November banks and deposit-taking companies forced the 1994 and June 1995, the 10 largest banks increased Hong Kong government to intervene to protect de- their share in total private deposits from 49 to 57 positors. Interventions occurred during 1982-86 percent, while of the increase in total deposits dur- and consisted of the government's taking over three ing the second half of 1995, about 80 percent was in medium-size commercial banks, arranging financial U.S.-dollar-denominated accounts. support packages for other financial institutions, Argentine authorities responded to the crisis by and providing guarantees against additional irrecov- significantly reducing bank reserve requirements to erable debts when the troubled banks were sold to permit an increase in liquidity in the system, negoti- private entities. Although the Hong Kong Mone- ating a financial assistance package with the multi- tary Authority could not act as a lender of last re- lateral institutions and creating a privately financed sort, the government responded by using reserves limited deposit insurance mechanism. Reserve re- held in excess of those in the Exchange Fund, a fund quirements were reduced so that the 25 largest banks whose main purpose is to stabilize and protect the were able to buy assets from the smallest ones. Thus, value of the Hong Kong dollar. Indeed, in April out of the approximately US$8 billion in deposit 1986, after several government-led banking rescues losses occurring between the end of December 1994 had occurred, the total reserves were estimated to be and mid-May 1995, about US$3.4 billion (41 per- at least HK$35 billion, HK$1.4 billion above the cent) was compensated with a fall in reserve require- minimum required in the Exchange Fund.36 ments, and only US$1.1 billion (13 percent) with a 297 W-11P CAPITAL FLOWS TO DEVELOPING COUNTRIES credit cut (other sources of liquidity were external form 40 percent reserve requirement, and the stock credit lines, repos, and central bank loans to banks). of liquid resources held at the central bank at the Three features of the Argentine banking system time of the crisis was about US$9.4 billion, approx- that helped reduce the impact of the crisis are worth imately 20 percent of total deposits. noting. First, banks were highly capitalized. Indeed, In sum, in order to achieve stability in the value although an 11.5 percent risk-adjusted capital-asset of the domestic currency and in the value of a sub- ratio is requLired for Argentine banks, when the con- set of bank liabilities, authorities in financially inte- fidence shock hit the economy, banks had a capital- grated economies must rely on policy instruments asset ratio of 13.4 percent nominal and 18.2 percent that can complement the role played by interna- when adjusted by risk (Fernindez and Schumacher tional reserves and can help ensure more resilient 1996). Second, the increased dollarization of the banks. High reserve and capitalization require- economy had created a bimonetary system. Thus, ments, although effective in increasing depositors' by the end of 1994 about half of total deposits were confidence, may reduce banks' profitability. Liquid- dollar denominated (compared with only 21 per- ity requirements seem to be a raore efficient instru- cent at the end of 1989), and dollar-denominated ment, as illustrated by the recent change of policy in loans represented about 57 percent of the total. Argentina. Although holding excess international Third, banks had high reserve requirements-al- reserves also seems effective in containing a banking though after the crisis these were changed for matu- crisis, it is important to note that such a policy is rity-related liquidity requirements ranging from 5 equivalent to the government's having a sound fiscal to 15 percen t of bank liabilities. Indeed, since 1993 stance (a surplus) that allows it to have permanent demand and savings deposits were subject to a uni- access to the international credit market. Notes 1. Increased competition occurs because of new-for- 1984-89, the Philippines 1978-83 and 1989-94, Swe- eign-banks and other nonbank financial intermediaries, den 1989-93, Thailand 1978-84 and 1988-94, and while new risks appear in the form of new instruments Venezuela 1975-80 and 1992-93 and investment opportunities. The new sources of risk comprise not only credit risk but also currency, settlement 3. It is important to distinguish conceptually the term and paymen ts, interest rate, and country risks. Also of an investment from its liquidity. A long-term invest- among the new risks are external shocks, such as a crisis in ment (long-term bonds or stocks) cm be liquid if investors a neighboring country, that can affect market sentiment can sell their holdings without forcing the dismantling of negatively, triggering a capital outflow and causing a the factory that was financed with the securities. squeeze in lic[uidity. 4. This school of thought highlights the importance 2. The country episodes analyzed here include Ar- of bank credit as a channel of monetary transmission. gentina 1979-82 and 1992-93, Brazil 1992-94, Chile This theory assigns a more relevant role to banks than 1978-81 and 1989-94, Colombia 1992-94, Finland standard macroeconomic (IS-LM) models do. See 1987-94, Indonesia 1990-94, Malaysia 1980-86 and Bernanke 1983, Bernanke and Blinder 1988, Bernanke 1989-94, Mexico 1979-81 and 1989-94, Norway and Gertler 1989. 298 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANC1A1&' 5. For a more detailed explanation see Kindleberger 12. For a detailed discussion of the Chilean experi- 1978, Calomiris and Gorton 1991, Hubbard 1991, ence see Valdes-Prieto 1994, Larrain 1989, de la Feldstein 1991, Calomiris 1995, and Minsky 1995. Cuadra and Valdes 1992. 6. In the case of inflows that enter the recipient 13. Both indices can take positive and negative values. country not through banks but through a different The macroeconomic vulnerability index is constructed by channel, such as portfolio investment, the first effect adding the scores of countries (in each episode), which are will be felt through an increase in asset prices, which based on their performance in terms of the current account will, in turn, increase the value of collateral and the fi deficit, consumption, and investment growth. For exam- nancial wealth of the private sector. Later on, as the in- ple, a country receives a score of -I (or 0) in terms of con- flows increase the monetary base, the volume of funds sumpuon growth if consumption growth in that episode being intermediated by the domestic banking system was smaller than (or equal to) that predicted by the size of will grow. the country's capital inflow. A country that scores a 3 on the macroeconomic front is one in which all macroeco- 7. The country episodes showing the greatest appre- nomic variables indicated an increase in vulnerability. A ciation of the real exchange rate are Brazil (1992-94), country that scores a negative number (or 0) is one in which Chile (1978-81), and Mexico (1979-81, 1988-94). macroeconomic conditions improved (or stayed constant) over the period of inflows (relative to the size of the coun- 8. It is important to note that lending booms are try's capital inflow). The financial index is constructed fol- not the primary cause of banking crises. However, lowing a similar methodology. Countries scores on the large lending booms, in the presence of weak macro- basis of the financial behavior are added. Countnes that lreoo lendin bom,inacl sethe presencenof weakb macro score a positive number are those in which financial condi- important factor leading to situations of distress (and tions deteriorated over the period of inflow. The opposite is impotantfacor ladin tositution ofdistess and tine in countries where the value of the financial index is in the extreme, crisis) in the banking sector. negative. Although the indices are on a numeric scale, they 9. The boom-bust in asset prices has been an out- should be interpreted as qualitative indicators more than as come of financial liberalization-and has had a signifi- qrantitative ones. In other words, the sign of both indices cant effect on the macroeconomy-in many country episodes other than those in our sample: for example, 14. Using different samples, other authors observe sit- Japan, the United Kingdom, and the United States uations of financial distress and bank insolvency that during the 1980s. (Schinasi and Hargraves 1993.) persist for some time, without a significant deterioration in the macroeconomy, or situations of financial distress 10. The banking crisis in Argentina, triggered by accompanied by balance-of-payments crises (Caprio and the devaluation of the Mexican peso, affected mainly Klingebiel 1996, 1997; Lindgren, Garcia, and Saal small provincial banks and was compounded by the 1996, Kaminsky and Reinhart 1996). However, in our lack of a deposit insurance scheme and a lender of last sample we have intentionally focused on those country resort. Its effects were, however, contained by a strong episodes where a surge in capital flows, and associated macroeconomic response adopted by the authorities intermediation by the banking system, played an impor- and by the takeover of weak institutions by the tant role. stronger ones. 15. One clear example of increased potential risk is the 11. During the first surge in capital inflows (1979- opening of the banking sector to foreign competition Al- 82), the authorities used the exchange rate as a nominal though increased competition leads to greater efficiency anchor to reduce domestic inflation, while during the in the long term, an abrupt change of regime that erodes most recent episode (1989-95), they used a crawling the franchise value of banking will aggravate moral hazard band of increasing width. and lead to more risk taking by domestic banks. Hence, 299 Q-APITAL FLOWS TO DEVELOPING COUNTRIES there may be problems in increasing foreign competition the 8 percent ratio). For details on the Basle Accord too quickly (Stiglitz 1994). guidelines see annex 5.3. 16. This does not contradict the view that in the 22. Most countries impose rules of maximum expo- long term financial integration is beneficial. In particu- sure to a single borrower as a percentage of capital and lar, it increases competition and enhances efficiency in other reserves of the lending bank. For each of the fol- banking, and permits greater diversification of bank lowing examples, the percentage is shown in parentheses: portfolios. Argentina (15-25), Brazil (30), Chile (5-30), Colombia (10), Germany (25), Hong Kong (25), India (25), In- 17. The level of interest rates also depends on expec- donesia (10-20), Israel (15), Japan (20-40), Korea (15), tations of cLevaluation. A highly appreciated real ex- Malaysia (30), Singapore (25), Thailand (25), the United change rate can lead to sustained high real interest rates States (15), and Venezuela (10). From Goldstein and (ex post) that undermine the health of the banking sys- Turner (1996). tem (chapter 1 and annex 5.2). 23. The Argentine approach differs from others-in 18. Regarding the incentive structure, for example, particular, the guidelines of the Basle Accord-in that it the incentives of all market participants must be taken uses the interest rate applied to eoch credit as a measure into account: bank owners and managers, depositors, of the risk associated with that operation. and bank regulators. Appropriate tools to do this include capital-asset ratios and extended (or unlimited) liability 24 According to current rislt-based capital regula- for the first group-prior to the mid-1930s double liabil- tions in the United States, family' residential mortgages ity was routinely imposed on U.S. bank shareholders and are in the 50 percent risk category while other real estate managers-limited deposit insurance for the second, and loans and loans to individuals are in the 100 percent risk appropriate funding, greater accountability, and political category (Grenadier and Hall 1995). See also annex 5.3. independen e for the last group. 25. For a detailed analysis of the pros and cons of cap- 19. Without these basic institutions the financial sys- ital requirements see Berger, Herring, and Szego (1995). tem as a whole, and the banking sector in particular, will not perform the function of intermediating funds prop- 26. This does not mean that banks should not be al- erly and will be more prone to fraud. lowed to fail or that the likelihood of a banking crisis should be reduced to zero. The need for preparation 20. Properly pricing a deposit insurance scheme is not arises because banks are bound to frequently encounter an easy task, mainly because the composition of a bank's financial problems and, for reasons related to market dis- portfolio cart change much faster than a regulator can re- cipline, shouldencounter such problems. The authorities alistically assess its risk level. Also, bank regulators should should try to keep these problems from becoming sys- continually update their pricing mechanisms to incorpo- temic crises. Also, it can be argued, based on a cost-ben- rate market developments and financial innovations. Al- efit analysis, that the likelihood of a banking crisis should though imperfect, however, a variable (adjusted by risk) not be reduced to zero. deposit insu rance premium can help to induce safer banking practices if other mechanisms, such as risk-ad- 27. Two risk areas that tend to he overlooked are the re- justed capital-asset ratios, limited deposit insurance, un- payment capacity of public enterprises and payment and limited (or extended) liabilities for managers, limited settlement risk. Concerning the former, the implementa- foreign exchange risk insurance, are put into place. tion of macroeconomic stabilizatiorn packages aimed at re- ducing fiscal deficits have led public enterprises (for 21. According to a recent survey of the Basle Com- example, in Venezuela) to default on their debt, thereby mittee, 92 percent of countries apply a Basle-like risk- compounding bank difficulties. Regarding the latter, in- weighted capital approach (although not necessarily at dustrial countries have recently begun to establish mecha- 300 THE EFFECTS OF INTEGRATION ON DOMESTIC FINANCIAL- nisms aimed at addressing this problem (IMF 1996), but 33. An amendment to the 1988 Capital Accords, ap- developing countries still have a long way to go. proved by the Basle Committee in 1995, adjusts banks' capital requirements to incorporate market risk as a vari- 28. By contrast, the losses in other nonbank deposit- able in its calculation. The amendment takes effect in taking institutions, which were lightly supervised, 1997 (IMF 1996). reached 40 percent of total deposits. 34 The number of withdrawals needs to be large 29. One way to minimize this risk is to permit entry to enough for the bank to run out of reserves and be forced only a few well-known international firms and banks with to liquidate assets at a discount. good reputations. However, this may be difficult to do in Eastern Europe and the former Soviet Union, where eco- 35. Hong Kong used a currency board, pegging the nomic activity is partly motivated by historical links. Hong Kong dollar to the British pound, up to 1972 30. Sources: Fleming, Chu, and Bakker (1996); Baer 36. The fund is required to maintain assets equal to at and Klingebiel (1995). least 105 percent of the stock of debt certificates and Hong Kong dollar notes issued. The total spent in help- 31. The case of Brazil can be partially explained by the ing Hong Kong's troubled banks was estimated to be less recovery of the economy after economic reforms were in- than 10 percent of the fund's total assets as of April 1986 troduced following several years of stagnation. (Far Eastern Economic Review, April 10, 1986, pp. 78-79). 32. The G-10 comprises Belgium, Canada, France, Germany, Italy, Japan, the Netherlands, Sweden, the United Kingdom, and the United States. 301 TER 6 Preparing Capital Markets for Financial Integration NE OF THE CHARACTERISTICS OF THE CURRENT phase of international financial integration, as noted in chapter 1, is that-given the changing investor base-a growing proportion of flows to developing countries is being channeled through = ~~their capital markets in the form of portfolio equity capital. This trend will continue and even intensify over the medium term.1 These investments represent an important opportuni- ty for developing countries and have been accompanied by a spectacu- lar increase in activity in the equity markets in these countries. Although the improvements in many emerging markets have been remarkable, most of these markets are still in the early stages of devel- opment and need to close the gap between themselves and the more advanced capital markets to be able to compete. Emerging markets are also increasingly competing among themselves for new issues and investors. To attract additional portfolio flows, therefore, developing countries need to address investor concerns regarding the attributes of their capital markets that raise transaction costs and risks, especially the reliability and efficiency of the infrastructure and trading systems, and transparency and fairness. By focusing on these concerns, policy- makers can also reduce their own fears that financial integration will increase volatility in their capital markets and the risk of a financial crisis. This chapter will discuss the implications of financial integra- tion for the functions and efficiency of domestic capital markets, and how policymakers can reconcile their concerns with the concerns of foreign investors. 303 CtPITAL FLOWS TO DEVELOPING COUNTRIES The Main Issues ¶ TT7HILE CAPITPAL MARKET DEVELOPMENT IS A KEY ELEMENT \ XX/ of the policy agenda for developing countries as financial vYv integration deepens, it is not a prerequisite in the same sense as a strong macroeconomic policy framework and a sound bank- ing system. The robustness of the macroeconomic and banking sector conditions in developing countries, and how these countries deal with overheating pressures and reduce macroeconomic and banking sector vulnerability, will determine whether they enter a virtuous rather than a vicious cycle of financial integration, and reduce the potential for large reversals. Capital market development, however, plays less of a role in minimizing these downside risks but is essential to maximize the upside potential of financial integration. The Benefits of Financial Integration Gmwth in market capitalization and activity. In parallel to the surge in port- folio equity flows, the capitalization of equity markets in many devel- oping countries has been expanding rapidly since the mid-1980s.2 As shown in table 6.1, by the end of 1996 the combined market capital- ization of 18 major developing countries included in the IFC Emerging Market Global Composite Index was 14 times larger than in 1985, rising from US$95 billion to US$1,371 billion. As a ratio to GDP, the average market capitalization of these countries had increased from 7 to 49 percent. This growth has been much more rapid than in developed markets. While market capitalization in emerging markets is, on average, still much smaller than in industrial countries, that difference declined substantially during the past decade. Because of this growth, equity markets in some of the more dynamic emerging markets, particularly in East Asia, have become increasingly important in domestic financial interrmediation relative to banks, despite the continued growth of bank deposits. For example, in Malaysia and Thailand, the share of equity markets in the stock of financial savings increased, respectively, from 49 percent and 9 per- cent at the end of 1985 to 79 percent and 56 percent at the end of 1994.3 In Chile, the equity market share in savings increased from 32 to 80 percent during the same period. 304 PREPARING CAPITAL MARKETS FOR FINANCIAL INTTQ', Table 6.1 Stock Market Growth in Selected IFC Index Countries, 1985-96 Stock market Stock market Trading value capitalization capitalization Trading value percentage of market (billions of doars) (rtagofGP) (bilions of dollars) capitalization) Country I985 1996 1985 1996 1985 1996 1985 1996 Emerging markets Argentina 2.0 44.7 2.3 15.3 0.6 4.4 31.0 9.8 Brazi 42.8 217.0 19.2 30.9 21.5 112.1 50.2 51.7 Chae 2.0 65.9 12.2 92.7 0.1 8.5 2.8 12.8 Colombia 0.4 17.1 1.3 20.4 0.0 1.4 7.2 7.9 Hungary 0.0 5.3 0.0 11.3 0.0 1.6 - 31.1 India 14.4 122.6 6.9 36.2 5.0 26.6 34.5 21.7 Indonesia 0.1 91.0 0.1 41.2 0.0 32.1 2.6 35.3 Korea 7.4 138.8 7.8 28.9 4.2 177.3 56.4 127.7 Malaysia 16.2 307.2 52.0 323.6 2.3 173.6 14.4 56.5 Mexico 3.8 106.5 2.1 37.6 2.4 43.0 61.9 40.4 Pakistan 1.4 10.6 4.4 17.6 0.2 6.1 17.2 56.9 Peru 0.8 13.8 4.4 22.6 0.0 3.8 5.0 27.5 Philippines 0.7 80.6 2.2 97.5 0.1 25.5 16.6 31.6 Poland 0.0 8.4 0.0 6.2 0.0 5.7 - 68.0 SriLanka 0.4 1.8 6.2 13.3 0.0 0.1 0.8 7.2 Thailand 1.9 99.8 4.8 53.1 0.6 44.4 30.6 44.4 Turkey - 30.0 0.0 17.5 0.0 36.8 - 122.7 Venezuela 1.1 10.1 1.8 16.7 0.0 1.3 2.7 12.7 Total 95.3 1,371.4 7.2 40.5 37.0 704.3 38.8 51.4 Developed markets France 79.0 585.9 15.1 37.6 14.7 859.0 18.6 146.6 Germany 183.8 664.8 29.7 27.9 71.6 1,405.4 38.9 211.4 Japan 978.7 3,019.7 72.9 65.1 330.0 933.1 33.7 30.9 United Kingdom 328.0 1,711.2 70.8 151.5 68.4 1,319.4 20.9 77.1 United States 2,324.6 8,478.0 55.6 111.8 997.2 7,265.6 42.9 85.7 Total 3,894.1 14,459.7 54.6 83.6 1,481.8 11,782.4 38.1 81.5 - Not avaiable. Note: End-of-year stock market capitalizarion figures are used. Japan includes only the Tokyo Stock Exchange. Soure. itc Emerging Markers data base; data from International Federation of Stock Exchanges (FrBv). In parallel to the growth in market capitalization, the volume and value of stocks traded have also risen dramatically. The combined an- nual value of shares traded in the IFC Index countries increased from US$37 billion in 1985 to US$792 billion in 1994 but declined to US$704 billion in 1996. Since this simple measure of trading activity is affected by changes in share prices, perhaps a better measure of trad- ing activity is the turnover ratio-that is, the annual value of trades normalized by market capitalization. This ratio indicates that trading 305 >AACAPITAL FLOWS TO DEVELOPING COUNTRIES activity in emerging markets roughly doubled between 1985 and 1994, about the same increase as in industrial countries. T he turnover ratio in emerging markets has since declined to 51 percent in 1996 but still re- mained significantly higher than in 1985. Combinred with the growth in market capitalization, the increase in the volume of shares traded suggests that liquidity has significantly improved in emerging markets. Increase in the foreigin presence in domestic markets. Foreign investors have made a significant contribution to these improvements. While the sources are not fully reliable, the data do show that the share of foreign investors in trading volume and market capitalization was very large in most major emerging markets in 1995 (table 6.2). Since meaningful foreign involvement in these markets is a recent phenom- enon starting in many cases only in the early 1990s, it is clear that the increasing foreign activity played an important role in improving the depth and liquidity in emerging markets. One of the few countries for which time-series data on foreign activity are available is Thailand. In 1986, the year preceding its capital inflow episode, foreign investors accounted for 8 percent of the trading turnover in Thailand. By 1990, Table 6.2 Estimates of Foreign Presence in Emerging Stock Markets, 1995 (percent) Forign share, i:> Foreign share Market market capitalizati.n -Estimate by in trading Estimate by Argentina 35 National securities comnmission B fKrazil -- - - - - 35 Stock exchange India 25 Loca brokers Indonesia - -75- Local brokers 29 Stock exchange 81 Stock exchange Korea 13 Securtifes and-exchange 6 - Securities and commission exchange com-mission Malaysia ~ -- ~ -~ - -~- -~ -=-- ~ ~- -~- ~ .~~- -~- -- -- ~ - - - - -- - - - 50 Local brokers Mexico -25 S.oc-- xchange Paista - - -50- Lotal brokers Peru-- 60 Local brokers Philippines 38 Local brokers 50 Local brokers -Poland - -25- -Stock exchange Thalland 21 Stock exchange 26 -Stock excharge Not available.- NImte: Mostrfigues represent rough estimates, of feign partiip aion,sinIcef6v coluntries documentfows. - -Soaree: IFC data, 306 PREPARING CAPITAL MARKETS FOR FINANCIAL INTEC G they accounted for 15 percent of turnover in the Thai stock market. This share declined during 1991-92, but rose again to 26 percent by 1995 and to about 32 percent during the first quarter of 1996. Increase in access to foreign markets. Increasing financial integration has also enabled developing country corporations and banks to gain direct access to equity and bond markets in industrial countries. Many devel- oping country firms have been able to cross-list in the world's major equity markets through global and American depository receipts (GDRs and ADRs, respectively). The growth in these instruments has been very rapid: at the end of 1995 there were some 228 ADRs and GDRs issued by firms from 26 developing countries with an original market value of about US$34 billion. On the debt side, since 1990, both corporations and banks have been able to issue paper in international markets, as well as in the U.S., Japanese, and some European markets. In addition to large sovereign issues, the amount of capital raised by the private sector from developing countries through international issues of debt during 1989-95 amounted to some US$52.4 billion. The number of countries whose issuers were able to gain access to international mar- kets through these instruments also increased. The Challenges of Integration for Emerging Markets Increasing competition. With the benefits of financial integration have come challenges for the integrating capital markets. Just as domestic issuers gain direct access to foreign capital markets, so must domestic markets compete with industrial country exchanges for listings and new issues. As noted above, the growth in these alternative means of channeling funds to developing countries has been enormous. By the end of 1995, ADRs and GDRs represented 6 percent of the underlying market capitalization of the IFC Emerging Market Investable Index. And as can be seen from table 6.3, this average 6 percent ratio hides a large variation among countries, with some having surprisingly high ratios. In some of these cases, the success of ADRs and GDRs is explained in part by the weaknesses of developing country capital markets. Weaknesses in these markets may cause foreign investors to acquire shares of a developing country firm through depository receipts rather than directly, particularly if transaction costs and delays in the market in question are high or investor protection is poor. For example, foreign participants in the Indian stock markets report that 307 IAA1ITAL FLOWS TO DEVELOPING COUNTRIES inefficiencies in these markets' clearance, settlement, and depository systems are a key reason for their interest in Indian GDRs. As can be seen from table 6.3, the value of outstanding Indian depository receipts relative to the value of shares that foreign investors can acquire (that is, the market capitalization of the IFC Investable Index) is among the highest of the 20 countries included in the table. With regard to the Chinese markets, foreign investors report concerns regarding the quality of disclosure and corporate governance practices of listed firms. Since depository receipts need to mneet the regulatory standards of the industrial country where they are issued, in this case too they are preferrec[ by foreign investors. Depository receipts are also preferred in the case of Argentina, where the domestic market is Table 6.3 Issues of American and Global Depository Receipts, 1989-95 (millions of dollars) Carrected issuesmarker capitalization Total issues - ernt) Country Unicorrected Corree -dGlobal index Investible index Argentina 4,204 3,898 10.3 17.7 Brazil 1,286 1,281 °0.9 2.0 Chite 1,538 1,996 2.7- 17.8 China 1,717 1,466 6.0 77.3 Colombia 329 277 3.3 3.4 Czedh Reptblic 32 32 0.3 0.7 Euigary 362 3-30 41.5 75.9 India -3,85-9 2,6 50 2.1 19.6 Indonesia 2162 2,267 3.4 - 11.5 Korea 3,102 3,175 1.7 16 6- Mexico 10,567 6,698 7.4 -12.1 Pakistan 1,143 771 83- 15.9 Peru 51 56 0.5 0.8 Philippines 1,015 1,295 2.2 7.6 Poland 51 51 1.1 2.6 - South Africa 477 49 0.3- 0.3 Sri Lanka 33- 20 1.0 2.6 Thailand -358 314 -0.2 1.1- Turkey 540 435 2.1 3.2 Venezuela 187 86 2.3 3.6 Total 33013 27,596 2-3 6.1 a. This esd des some $70G millionI issued by bS. firomn six cuntries not included in the iFC indexes. b. FigsteBs are eorreCEed for changTes in stoxck pric:es betwreen the date of issue sand the end of 1 995 ipsing the wc Globa Indexes (Ircc). Refets to market uApitalzaiont as meassred by- the Fcc, and iFc Iivestablo indexes, respertive- Source: Jrc, ergnStgclkMarken Fetbooklg 1f99World Bankdata. 308 PREPARING CAPITAL MARKETS FOR FINANCIAL INT perceived as very illiquid and has been losing volume to industrial country markets. In 1995, for example, the volume traded in New York of ADRs of Yacimientos Petroliferos Fiscales, a recently privatized Argentine oil company, was 25 times higher than the volume traded on the Buenos Aires market. In Korea the domestic market has been more restricted to foreign investors, and the high depository receipts- market capitalization ratio is an indicator of repressed foreign demand for equity. Addressing investor concerns. To compete in an increasingly integrated world, developing countries need to make their markets more attrac- tive to foreign investors. As explained in box 6.1, investors are con- cerned about the unreliability of emerging markets in three main areas: market infrastructure that results in delays in settlement and failed trades; lack of protection for property rights, including those of minority shareholders; and lack of transparency and fairness of mar- kets, because of insufficient disclosure of accurate information that would enable investors to assess the merits of alternative investments and because of insider trading and other abusive practices. In addi- tion, despite recent progress, emerging markets remain significantly less liquid than capital markets in industrial countries. In an illiquid market, investors fear they will not be able to liquidate their interests quickly without incurring a substantial loss. The concerns of developing countries. For their part, policymakers in developing countries also have a number of concerns about growing international integration. Perhaps most important, they fear that finan- cial integration may increase volatility in their capital markets. As dis- cussed in chapter 2, financial integration makes developing countries more susceptible to external shocks. In addition, foreign investors may add to excess volatility in asset prices in emerging markets through herding behavior or contagion effects.4 Policymakers are also con- cerned about the increasing vulnerability to financial crises as foreign investors become more important in both trading activity and market capitalization. In particular, they fear that foreign investors, because of herding, fads, or momentum trading, may increase the likelihood or magnitude of market bubbles, with securities prices rising far above the underlying fundamentals, followed by an inevitable market crash. Policymakers in developing countries are also concerned about the equity implications of increasing foreign ownership of domestic firms. Some officials perceive foreign portfolio investors as fair-weather 309 + A!0A AL FLOWS TO DEVELOPING COUNTRIES Box 681 Investors' Viewpoint: Risks and Transaction Costs in Emerging Markets THE HIGHER RETURNS THAT CAN BE EARNED BY their portfolios without incurring high costs. capital in developing countries are the fimdamental Liquidity, to a large extent, is an endogenous force driving investor interest in these countries. How- variable-that is, the result of capital markets ever, the stylized attributes of emerging markets (shal- having the right attributes to be able to attract low, ifliquid, and opaque markets; high transaction large numbers of buyers and sellers. Liquidity costs; and weak regulatory frameworks) and the un- needs to be nurtured, it camnot be bought or derlying developing economy (larger, more frequent created by decree. macroeconomic shocks and less mature, undercapital- ized firms) mean that risks are also high. In particular, Investors will be concerned about these risks to both domestic and foreign investors in developing lifferent degrees, depending on their risk-return countries are subject to very large risks and costs not preferences and investment strategies. For examnple, direcdy related to the underlying investment, value investors who do not inordinately turn over Nonmuarket risks. These nonmarket risks mainly their portfolios would be concernedl about company regard the lack ofreliability in four basic areas: disclosure practices but would be less concerned with transaction costs and liquidity. • Investors require a reliable system to settle Setemn and operaoal dsks and cost. Capital transactions, with either cash or securities, to market infrastructure in many emerging markets is recLuce principal risk and the opportunity cost stilL in a nascent stage, and investcors are subject to of a delay. high settlement risks, both operational and counter- * Investors demand reliable systems that record party. Risks that a party will default on payment or ownership, ensure safe custody of securities, delivery obligations are large, and failed trades or andi protect property rights. Investors require long delays in settlement are widespread. The long that securities truly represent a claim on a fu- delays reduce liquidity and increase market risk. For ture income stream and that claims be en- example, a Temnpleton fund bought shares in India forceable by law. Related to this is protection that subsequently rose by 150 percent. The fund against abuse by management or majority was not able to make good on the potential profit shxreholders, which could reduce the value of because by the time the original trarnsaction had set- the investment. tded and the fund could sell, the stock price had * Investors also want reliable systems to ensure declined back to its original level (Seeger 1996). that they pay or receive a fair price for a secu- Failed trades expose investors to counterparty risks rity and, most important, that ensure disclo- if the setlerment system does not ensure that shares sure of material information to evaluate are only delivered versus payment (DvP-that is, investment choices. that the finaL delivery of securities takes place if and * Investors desire liquidity to be able to liqui- only if final payment is made). Most emerging mar- date securities or change the composition of ket do not conform to DVP. Failed trades can have friends, interested in benefiting from large short-term capital gains and dividends but contributing little to the long-run health of domestic firms and the development of the economy-and selling at the first hint of trouble. Perhaps more important, there is a perception in many 310 PREPARING CAPITAL MARKETS FOR FINANCIAL I 1N 1 systemic consequences for a securities market, since value of the firm, and practices such as dealings with they may produce a chain reaction of failures. insiders, need to be submitted to a vote by share- Investors also desire reliability in postsettlement holders. There is a fear that in developing country actions, particularly the timely payment of divi- firms such protections may be lacking or not effec- dends to reduce opportunity costs and market risk, tively enforced, especially since many of these firms such as an unfavorable change in the exchange rate. are closely held and managed directly by majority Legal and custodial nsks and costs A key risk in- shareholders. vestors face in emerging markets is that securities lnfnnatul and regutory nsks and costs. Lack purchases may not be recorded in the legal registry, of quality information on firms, combined with providing them no way to prove ownership. If the high asset price volatility, is another prime concern omission is deliberate, recourse to the courts may be for investors. One reason why the information base costly and lengthy, with an uncertain outcome if is lacking in emerging markets is that the accounting property rights are not well defined. Real or per- and auditing systems and standards are weak, and ceived bias by the judiciary against foreign interests, that there are not enough qualified accountants and or the Iack of adequately functioning arbitration and auditors, as shown by numerous World Bank re- legal systems, would further increase uncertainty. views. More to the point, the regulatory systems in These risks are especially prevalent in transition many of these markets are weak in terms of quality, economies where the concept of private property, let quantity, and frequency of information disclosure. alone legal protection for it, is very recent. For ex- Foreign investors and fund managers feel parricu- ample, Dmitry Vasilev, chairmnan of Russia's Federal larly at risk, believingdomestic investors to be better Stock Commission, commented in May 1996: informed. Another concern, which is also related to "Current legislation does not defend shareholders. protection of minority shareholder rights, is insider Confidence in the stock market has been bruised by trading. For example, the International Organiza- secret company meetings at which share registers tion of Securities Commissions (Iosco) reports that have been altered and new issues voted on" (Seeger only about one-half of its developing country mem- 1996, p. 15). Even if securities are registered, it may bers require disclosure of securities transactions be ordy with a lengthy delay with a loss of dividends made by company insiders. Even in the more ad- and other shareholder rights. Operational risks such vanced emerging markets, such as Thailand, insider as loss of shares or counterfeit securities are another trading is perceived to be difficult to control despite concern. There have been instances of counterfeit the strong monitoring and enforcement powers of securities in India, Indonesia, Malaysia, and Turkey. the Thai Securities Commission. A recent scandal involved Russia's Minfin bonds. In addition, investors are widely concerned about Source: Authors interviews with fund managers and ad- their rights as minority shareholders. In developed visers, Gray (1996), josco (1992a), Mobius(1995), Seeger markets, decisions that would significantly affect the (1996). developing countries that foreign and domestic investors are not com- peting on a level playing field, that foreign investors, because of their wealth, will be able to buy a large share of the equity of domestic firms, to the detriment of long-term national income. 311 IM CAPITAL FLOWS TO DEVELOPING COUNTRIES Finally, integration and globalization are raising new issues for capi- tal market regulators in developing (and industrial) countries. For ex- ample, integration increases systemic risk, since the failure of a financial intermediary overseas could have an impact on domestic markets. In addition, globalization may reduce the effectiveness of monitoring and supervision of financial intermediaries because of difficulties in assess- ing the financial status of firms that are active in many markets, and be- cause of potential gaps in the responsibilities of regulators in different countries. And two parallel global trends, the development of deriva- tive products and financial conglomerates, are making financial mar- kets even more opaque. The Barings affair of 1995 is a good example of these new problems: a securities subsidiary of a U.K. bank located in Singapore was taking positions in the Nikkei futures index. Who was responsible for monitoring what and where? The Policy Agenda In summary, developing countries will be able to be nefit from increased investment, and from the deepening and improved liquidity of their capital markets, only if they put in place the institutional and policy prerequisites to attract capital inflows, and reduce the risks of instabil- ity. To this end, developing countries face three main tasks: * Capital markets-especially equity markets, given the expected composition of capital flows-must be made more attractive to foreign investors. While investors are attracted by the potential for rapid growth and high returns, they are d scouraged by oper- ating inefficiencies and lack of reliability of market institutions and infrastructure, and by regulatory frameworks that increase transaction costs and reduce transparency. * Developing countries need to implement policy reforms and strengthen institutions to reduce the risks of instability. As ex- plained below, improvements that increase the attractiveness of emerging markets for foreign investors also serve to reduce volatil- ity and risks. * Authorities in developing countries also need to deal with the new regulatory concerns resulting from globalization. These con- cerns are shared by industrial countries, and their resolution will be greatly facilitated by international initiatives. 312 PREPARING CAPITAL MARKETS FOR FINANCIAL JNT,kg These institutional and policy reforms are not only prerequisites for successful financial integration but are also essential to develop capital markets in a more closed economy. Domestic and foreign investors generally share the same concerns, and hence both would welcome the same institutional and policy improvements. Similarly, measures that reduce volatility and risks that originate from foreign shocks would generally also be beneficial for domestic sources of volatility. Financial integration increases the urgency of these reforms: developing countries need to act quickly to take full advantage of the substantial opportuni- ties offered by financial integration. Organization of the chapter. These issues are developed in the remain- der of the chapter. The next section discusses whether financial inte- gration facilitates the role of capital markets in investment. It also reviews how integration might exacerbate inefficiencies in domestic capital markets as a result of information asymmetries and price volatility. The third and fourth sections review the reforms and improvements required for emerging capital markets to operate and develop in an increasingly integrated world-that is, measures that enhance the attractiveness of domestic capital markets to foreign investors and reduce potential volatility and address regulatory con- cerns regarding globalization.5 The third section will focus on market infrastructure (that is, microstructures, clearance, settlement, and depository systems), and the fourth on the regulatory and legal frameworks. Each of these sections will discuss alternative institu- tional and policy options in their respective areas, and establish, to the extent possible, best practice, as well as describe the constraints typically encountered in implementing these reforms.6 Each of these two sections will also describe the progress of the more dynamic emerging markets in achieving the desired institutional and regulato- ry attributes. Finally, the last section will be a summary that offers conclusions. Capital Markets and Financial Integration Capital markets in a market economy fulfill three functions: * First, capital markets serve as a source of long-term capital for fi- nancing investment. 313 I CA1TAL FLOWS TO DEVELOPING COUNTRIES * Second, capital markets expand the menu of financial instru- ments available to domestic savers, allowing risk diversification and encouraging resource mobilization. * Finally, capital markets-especially equity markets-continu- ously monitor the corporate sector, serving both as a signaling de- vice for the allocation of capital and as a means of corporate control, thereby promoting managerial and organizational effi- ciency. However, informational asymmetries, including princi- pal-agent problems in the area of corporate control, as well as price volatility, may impede equity markets from correctly fulfill- ing their monitoring and signaling functions.7 The basic questions we will address in this section are whether and how financial integration facilitates the first and third functions and ex- acerbates volatility. The impact of financial integration on risk diversi- fication and resource mobilization was discussed in chapter 3. How Financial Integration Helps Foster Investment Properly functioning capital markets help increase investment by af- fecting both the supply of and demand for capital in several ways. First, they are a cost-efficient way to attract savings from a large group of small savers, thereby reducing the cost of capital for firms through economies of scale. Second, capital markets have two risk-sharing and diversification properties that promote the finaicing of riskier but higher-return investments: they reduce the vulnerability of firms to in- terest rate and demand shocks by facilitating the process of raising eq- uity by firms, and they reduce risks faced by investors by easing portfolio diversification. Finally, capital markets, like banks, perform a term transformation function. Many investments require a long time span to generate returns, while investors generaLly wish to commit funds for a shorter period. With liquid and active secondary capital markets, both requirements can be met simultaneously, since investors feel assured that they will have access to their funds quickly and with- out paying an excessive price. The benefits of financial integration. Financial integration enhances this role of capital markets in several ways. Most cirectly, integration expands the supply of investment resources by tapping foreign sources, increasing the demand for domestic securities. The increased 314 PREPARING CAPITAL MARKETS FOR FINANCIAL iNT demand will drive up the price of domestic securities, raising the price-earnings ratio and reducing the cost of capital. Less directly, internationally integrated stock markets allow wider risk diversifica- tion and thereby facilitate the implementation of higher-return but riskier projects (for example, see Levine and Zervos 1996b). Finally, as noted above, increased foreign activity improves the depth and liquid- ity of domestic capital markets, key ingredients for these markets to perform their term transformation function. The fact that a growing share of foreign investment is accounted for by institutional investors could magnify the positive impact on liquidity, since institutional investors are very active traders.8 With improved liquidity in domestic markets, investors will lower their demands for higher yields, reflect- ing their ability to sell securities at declining costs, and the cost of capital will decline. These favorable effects should lead to changes in the behavior of domestic agents. The declining cost of capital and the enhanced risk diversification should induce the corporate sector to issue initial public offerings (wos) and additional shares, including offerings and shares in emerging sectors, such as private infrastructure projects. In addition, as liquidity in domestic capital markets improves, new domestic investors will be attracted to these markets. The data support the hypothesis that financial integration enhances the role of capital markets as a source of investment finance. As dis- cussed in chapter 3, the growth in both stock market capitalization and turnover in the major emerging markets is correlated with the level of foreign activity, as measured by the magnitude of portfolio equity in- flows. And most important, the analysis in chapter 3 also suggests that foreign activity has had significant positive spillovers to domestic activ- ity, including the level and growth of domestic trading activity and the number of new listings. All of this rise in activity in emerging markets has real and positive implications for investment and growth prospects in developing coun- tries. While conventional wisdom based on industrial country data sug- gests that capital markets are not a large source of investment financing (see Mayer 1989), there is increasing evidence that they are much more important for this purpose in developing countries. For example, ana- lyzing data on the 100 largest corporations listed on the stock market in 10 developing countries during the 1980s, Singh (1994) finds that these corporations relied strongly on external sources for financing, and in particular on equity markets (41 percent of total financing, on average).9 315 CAPITAL FLOWS TO DEVELOPING COUNTRIES In addition, there is growing empirical and theoretical support for the idea that development of the stock market has positive implications for economic growth. Levine and Zervos (1996b), for example, find a strong positive long-run empirical association between stock market de- velopment and the increase in per capita GDP.'0 Financial Integration and Improvements in Corporaite Governance As noted above, one means through which equity markets increase in- vestment efficiency is by serving as a mechanism for corporate control. In essence, shareholders can exercise their right to change the manage- ment of the firm if they perceive that it is not acting in their best inter- est. In addition, investors can react to weak management performance by selling (or by refraining from buying) shares, acd ions that can lead to a decline in share prices. In turn, low or declining share prices can in- fluence owners and managers to change their behavior and improve corporate performance. Indeed, underperforming firms will have share prices that are low relative to their underlying value and hence will be more vulnerable to takeovers.'1 However, the academic literature is di- vided on how well the market for corporate governance works. For ex- ample, shareholders may not be able to monitor management without incurring high costs because of information asymmetries. The incen- tives for individual shareholders (unless their holdings are large) to incur these monitoring costs may be perverse because of the free-rider problem.)2 Information vendors and analysts play an important role in reducing these information asymmetries and enhancing the effective- ness of the market as a means of corporate control. Corporate governance and institutional investors. Another issue widely dis- cussed in the literature is whether the principal-agent problem improves or worsens as institutional investors come to account for a larger share of a firm's market capitalization.13 Some authors believe that institutional investors, with their strong professional background, will be able to monitor corporate performance effectively. On the other hand, their incentives to monitor management behavior may not be strong because they turn over their portfolios quickly. In addition, the free-rider problem mentioned above can also affect institutional investors, given prudential regulations that limit t-he concentration of the portfolios of institutional investors in an individual firm. According to Samuel (1996), the evidence, based on U.S. data from 316 PREPARING CAPITAL MARKETS FOR FINANCIAL INTE- the 1980s, is that there is no discernible effect of institutional owner- ship on corporate performance but that the monitoring activities of institutional investors may be functioning as a substitute for the disci- plinary role traditionally played by the providers of debt financing.14 However, these empirical analyses do not capture the full impact of the recent sharp increase in shareholder activism by institutional investors, which in some instances has forced key firms in the United States and some European countries to improve their structures for corporate gov- ernance. Some of these experiences have been pathbreaking, in particu- lar the influence of the California Public Employees Retirement System during the board shakeup at General Motors in 1995. Foreign investment and corporate governance. This discussion suggests that increasing foreign participation in domestic stock markets in developing countries would have both negative and positive implica- tions for corporate governance. On the negative side, overseas investors, because of information asymmetries, would seem to lack the familiarity with local conditions needed to be effective monitors of management. In addition, foreign, especially institutional, investors may not have strong incentives to participate actively in corporate governance functions; they may be more interested in liquidity (if unhappy with performance, they sell) than in control. On the other hand, as discussed in chapter 2, the nature and objectives of foreign investment change as emerging markets mature. During the period in which foreign investors follow an index-based approach, they take lit- tle interest in the underlying companies. But as they become more selective and pick stocks more carefully, they may take a more active role in corporate governance. In addition, to attract increased foreign funds, and as foreign practices are adopted by domestic shareholders, financial integration may lead domestic companies to improve corpo- rate governance. This demonstration effect may prove to be quite important in the medium term, as suggested by the experience of the large Indian development-investment bank described in box 6.2. Preventing an Increase in Volatility If capital markets, in particular equity markets, are to function as a sig- naling device for the allocation of capital, information markets need to work efficiently so that asset prices reflect all material information. Ob- vious impediments to efficiency are lack of information, delays in its 317 R"LTAL FLOWS TO DEVELOPING COUNTRIES Box 6.2 Corporate Governance in India FOREIGN PORTFOLIO INVESTMENT HAS LED TO Based on the Cadbury Commirtee's recommen- radical changes in corporate governance at the In- dations on corporate governance in the United dustrial Credit and Investment Corporation of India Kingdom, the main reforms implemented at iciCI (IcicI), one of India's largest development-invest- regard the role and composition of the board of di- ment ban-ks. While foreign portfolio investors in de- rectors, including having a distinct chairperson who veloping countries are generally not considered is separate from the chief executive officer. The re- demanding, and tend to vote with their feet rather forms have created a more balanced, responsible, than at board meetings, ICICl's experience probably and independent board, with greater participation presages the future. Indeed, the changes that are tak- by the independent external directcors. To further in- ing place at ICiCI parallel those of a growing number crease the sense of responsibility toward sharehold- of corporations in developed countries. ers, directors now have a fiduciary responsibility. ICICI's ongoing experiment with corporate gover- Complementing these reforms at the board level, the nance started with the issue of several GDRs in the risk management and internal audit departments early 1 990s, which led to a change in the ownership have been strengthened and made more indepen- structure of the company. Today, ICICI still has more dent, and a new key performance indicator for mid- than half a million shareholders, but some 34 per- dle management is the impact of their work on cent of the shares are held by foreigners, especially shareholder value. large institutional investors, and 41 percent by large Some developing countries are uneasy about the domestic institutions, including the central govern- presence of foreign investors on the boards of their ment. Foreign investors were critical of the com- largest firms. But if developing country govern- pany's activities in several areas, including ments wish to encourage long-term investors with substandard IPOs, diversion of funds into nonproject buy-and-hold strategies, they must be prepared to areas, poor accountability and transparency, prefer- accept their demands for increased transparency and ential allotment of shares to promoters, and, more control to ensure the value of their investment. generally, little concern for managing the company Judging from the Indian experience, foreign repre- to increase the value of shares. Foreign investors sentation on the board has had significant and posi- joined large domestic investors in voicing these con- tive impact on corporate governance. Indeed, this cerns at board meetings and were instrumental in may prove to be one of the mote important and having nmanagement accept sea changes in corporate long-lasting contributions of foreign investment to governanice. In turn, ICICI has pushed for similar emerging markets. changes in its many client companies. Recognizing the value of these changes, the government has been a passive but approving spectator. Source. Kamath (1996). dissemination, information asymmetries among market participants, and weak analytical capacity of market participants. Some authors (for example, Kyle 1984) have argued that the potential to make a profit in the stock market on the basis of new information promotes research by market participants. But this self-correcting mechanism seems to work best in active and liquid markets, where the potential for profit is 318 PREPARING CAPITAL MARKETS FOR FINANCIAL 1L higher. Another impediment to efficiency is volatility in asset prices, which makes it difficult for market participants to distinguish whether changes in equity prices are due to noise or to new material informa- tion on fundamentals such as dividends or interest rates.15 Another source of volatility and inefficiency comes in the form of speculative bubbles, which Stiglitz (1990) defines as asset prices departing from values justified by fundamentals because of expectations about future additional price increases. Volatility and lack of information may feed on each other Without information, some investors and market mak- ers are less likely to make bets against the market, a fact that could ex- acerbate price movements. The impact of financial integration on market efficiency and volatility. In theo- ry, financial integration has both positive and negative implications for the price discovery process in domestic capital markets. On the positive side, foreign investment increases depth and liquidity in domestic capital markets, thereby reducing volatility. (Shallow mar- kets are more prone to volatility since even small trades in these mar- kets have a disproportionate effect on prices.) In addition, increasing foreign participation in domestic capital markets may induce improvements in accounting, information, and reporting systems, as well as increase the analytical sophistication of the domestic securities industry. There is, in fact, strong anecdotal evidence that this spillover effect of financial integration has been quite important in some devel- oping countries. These two benefits should interact and reinforce each other: improved liquidity and profit-making opportunities should lead to increased research and better information systems, which in turn should provoke additional investor interest and activity. On the negative side, however, other factors suggest that financial integration may lead to an increase in the volatility of domestic asset prices and returns. This is because, with financial openness, domestic capital markets are exposed to new external financial shocks (or these shocks may be transmitted more quickly across borders), such as changes in global interest rates, spillover effects from foreign stock mar- kets, and investor herding. As discussed in chapter 2, some of these ex- ternal shocks, particularly changes in global interest rates and certain stock market spillover effects, make asset prices and returns more volatile by affecting the fundamentals of an emerging market. But other shocks, such as investor herding and pure contagion effects, may change investment in a country even though its fundamentals are un- 319 1T A ITAL FLOWS TO DEVELOPING COUNTRIES affected. These shocks are often the result of foreign portfolio investors having little access to information, worsening irformation asymme- tries. Perversely, the improvements in liquidity noted earlier in this chapter may make emerging markets more susceptible to external fi- nancial shocks, since better liquidity reduces transaction costs and makes it easier for foreign investors to open and liquidate positions. Given the high share of foreign investors in the major emerging mar- kets, these potential external sources of volatility are important. Information asymmetries may also increase volatility through inter- action effects between domestic and foreign investors. For example, a defensive reaction by local investors to the sale of domestic securities by foreign investors, who in turn are responding to events overseas, may magnify the impact of foreign stock market spillover effects on the do- mestic market. Since local investors generally do not know why foreign investors are changing their holdings of domestic securities, they may react to such changes even though the fundamentals of the domestic market have not changed. Similarly, information asymmetries could re- sult in foreign investors magnifying the impact of the behavior of do- mestic agents. What does the empirical evidence say? Most recent empirical studies have concluded that asset price volatility in emerging markets is gener- ally higher than in developed countries but that volatility did not in- crease during the current inflow period."6 For example, Richards (1996) found no evidence to support the hypothesis that volatility in emerging markets increased in recent years concurrent with the boom in portfolio inflows. Indeed, his results suggest a decline in absolute volatility. IMF (1995) also found that absolute voladility of stock market returns did not increase during periods of high and volatile portfolio inflows in Korea, Mexico, and Thailand. Bekaert and Harvey (1995), as well, observe that the volatility of returns remained unchanged or de- clined in 13 out of their sample of 17 countries after liberalization of their capital markets. How can these reassuring empirical results be reconciled with the theoretical predictions that volatility may increase: A possible explana- tion is as follows. As shown in figure 6. 1, volatility may originate from both domestic and international sources, as well as result from changes in country fundamentals or market inefficiencies. Although emerging markets became more susceptible to external financial shocks during the 1990s as they opened their economies, they were also undertaking 320 PREPARING CAPITAL MARKETS FOR FINANCIAL INT, Figure 6.1 Factors Affecting Volatility of Asset Prices in Emerging Markets, 1990s Factors increasing volatility Factors reducing volatility Increased susceptibility Policy reforms Improve to fluctuations in global stability and perfor- Fundamental Interest rates and for- mance of the domestic volatility eign stock markets. economy. Fundamental volatility Increased susceptibility Vltlt a tonhereaseding susehviby t remained con- Diverslflcation of to herding behavior by stant or domestic economy intecrnatiounal y i onvetr declinedcn os reduces vulnerability to and crs ntarion - emerging mar real external shocks. tagion. ~~~~~~kets. Excess volatility Interaction effects Improvements In capi- between foreign and tal markets reduce domestic investors instances of Incom- because of incomplete plete or asymmetric Excess or asymmetric informa- Information and volatility tion that magnifles fluc- ipoelqiIy tuations. impfove liquidity. policy reforms aimed at improving domestic fundamentals and stabi- Despite the increased susceptibilit lizing their economic policies. These economic reform programs also to external sources of volatiliy, improvements in macroeconomic led to the diversification of their economies, which reduced vulnerabil- management and capital market ity to traditional external shocks such as changes in terms of trade. In attrbutes may lead to a decline in addition, many developing countries during this period also improved overall volatility. their capital markets, reducing excess volatility arising from informa- tion asymmetries and other market imperfections, including foreign investor herding and pure contagion that as noted above may be caused by incomplete or asymmetric information. All of these effects have re- duced fundamental volatility from traditional sources of shocks and may have moderated the potential for volatility arising from new sources. Indeed, chapter 2 concluded that although there is evidence of herding behavior by foreign investors, it is not robust enough to war- rant the commonly held view that foreign investors destabilize emerg- ing markets. Chapter 2 also concluded that the contagion effects do 321 ioAlTAL FLOWS TO DEVELOPING COUNTRIES exist but that they appear to be relatively short-lived. In fact, the most significant corrections in developing country stock markets-for ex- ample, Mexico in 1994 and Thailand in 1996-have resulted not from external shocks but from shifts in the perceptions of domestic and for- eign investors regarding domestic fundamentals. As shown in figure 6.2, the improvements in capital markets have been a significant factor in reducing volatility. Figure 6.2, which re- lates excess volatility with an index of market development, suggests that market development is strongly associated with lower volatility, although the direction of causation can run both ways. Later in this chapter, we will explain in detail how the index was constructed. For the moment, it suffices to say that the index takes into account the depth and liquidity of a market, the efficiency of its infrastructure, and some additional variables that measure by proxy institutional development. Figure 6.2 Excess Volatility and Market Development Log (excess volatility) Correlation coefficient = -0.60 2.5 Korea India 2 \ I ~~~~Indonesia, * Pakistan 1.5 ?>xico Argentina Brazil Malaya' Chile 0.5 Thailand Improvements in capital mariets can reduce excess volatility. 0 3 4 5 6 7 8 9 Emerging market development index Source World Bank staff estimates 322 PREPARING CAPITAL MARKETS FOR FINANCIAL INTRj< Conclusions Two basic conclusions can be drawn from this discussion: * First, liberalizing foreign access to domestic capital markets can bring substantial benefits to developing countries. Liberalization enables countries to tap into large overseas pools of capital, bring- ing in foreign portfolio investment that increases price-earning ra- tios and the depth and liquidity of the domestic capital market. This, in turn, reduces the cost of capital for domestic firms. More- over, foreign participation may have important spillover effects on emerging markets in the form of improved accounting and disclo- sure practices and human capital. To realize these benefits, how- ever, developing countries need to reduce transaction costs and take other actions to increase the attractiveness of their markets. * Second, asset prices in developing countries are more volatile than in developed markets, and financial integration may in- crease volatility even further Volatility and lack of information interact with each other and together constitute a major impedi- ment to the price discovery process in emerging markets. Infor- mation asymmetries will, in addition, increase the agency cost for foreign investors. Volatility tends to decline, however, as emerg- ing markets become less prone to fundamental shocks through improved economic policies and diversification. But, excess volatility resulting from information asymmetries and deficien- cies will have to be tackled through reforms and improvements in the attributes of the capital markets themselves. Improving Market Infrastructure W H HAT IS MARKET INFRASTRUCTURE? MARKET INFRA- structure comprises the systems and institutions that facili- tate the trade and custody of securities. These functions can be subdivided into matching buyers and sellers, determining price, exchanging securities for good funds, registering securities to the new owners, and collecting dividends and other custody functions.17 A good way to understand the main components of market infra- structure is to follow the main steps in a market transaction. A transac- 323 API1TAL FLOWS TO DEVELOPING COUNTRIES tion begins with the instructions of a buyer and a seller to their respec- tive brokers to buy or sell a security. Market infrastructure kicks in when the market microstructures (that is, the trading system) match buyers to sellers and facilitate the price discovery process. The details of the trade are confirmed with the two counterparties by the trade comparison sys- tem before any further processing takes place. Then, the dearance sys- tem determines what the counterparties are to deliver and receive at settlement. Then comes settlement, when the transaction is completed and the securities and funds change hands.18 The depository provides for safekeeping of securities but, more important, can facilitate the set- tlement process and other aspects of custody such aos reregistration and payment of interest or dividends. Box 6.3 illustrates the main compo- nents of market infrastructure using the example of Thailand. Why does it matter? During the 1990s, the exchanges, the govern- ments, and the securities industry in emerging markets have all empha- sized the need to develop capital market infrastrucriure, in response to long-standing concerns of foreign investors. When the infrastructure systems work well, as they usually do in developed markets, trading flow and custody seem painless. But in many developing countries, inefficient infrastructures have often been unable tc handle the rapidly expanding volume of cross-border portfolio flows, in particular equity flows. The unreliability of market infrastructure has led to delays and failed settlements, the loss of principal because shares were not regis- tered to the new foreign owners, and difficulties in exercising entide- ments such as subscription rights and collecting dividends. Since these problems were discouraging many investors from entering their mar- kets, the authorities in developing countries made improving infrastructure one of their top priorities. In developed markets, the importance of these functions became most recently apparent with the settlement problems that arose during the 1987 stock market crash. Improving market infrastructure is a means not only to attract po- tential investors and issuers, but also to reduce systemic risk. The major sources of systemic risk in capital markets are volume and volatility surges, the default of a major market player, and operational break- downs. Because of weaknesses among financial intermediaries, emerg- ing markets are already prone to defaults by market players, to which integration adds increasing volume and creates the potential for in- creased volatility. A well-designed infrastructure will not only reduce the risk of operational breakdowns but can also make a significant con- 324 PREPARING CAPITAL MARKETS FOR FINANCIAL INIV Box 6.3 Clearance, Settlement, and Depository Functions in Thailand THAILAND'S CLEARANCE AND SETTLEMENT simplifies C&S, since the estimate of what is (C&S) system is a good example of the basic mech- owed by each counterparty and the transfer of anisms often used to clear and settle transactions in shares from the seller to the buyer can be done securities markets. The concepts used in this box are within the same institution. explained in detail in the main text. * With regard to the depository function, since The C&S of transactions on the Stock Exchange of September 1992, the securities market in Thailand (SET) are handled by a subsidiary of SET-the Thailand has been moving toward a demateri- Thailand Securities Depository Co. (TsD). TSD also alized environment, with stocks transferred functions as the central depository for equities listed on directly between accounts of the members of SET. The chart describes the main steps in the process. the depository, instead of physical delivery of share certificates. Investors can also ask for * Trade instructions are received and executed physical delivery of shares if desired. In the by the brokers at time T. case of physical settlement, the selling dient * On T+2 days, matching is completed via tele- must deliver the share certificates no later than phone among market participants. 1,400 hours on T+3. * C&S takes place on the afternoon of T+3. Al- a There are no fixed C&S practices for though the system is presented as DVP, pay- bonds and other securities not traded on ments are made by check and these take one the SET. day for clearing. * C&S in Thailand is the responsibility of the Source: ISSA (1996), author's analysis of data from central deposirory-TSD. This arrangement Citibank and Bank of New York. Box Figure 63 Activities Time__ _ _ _ _ _ _ _ _ _ _ Buying bokew Execution~ of the transaction ~ bOI T X T+2 9 t4noptele___ T43 PhysicalS | ScriptessC&SthroughTSD| Share certficates Cash t brokerbmkerol 325 4AITAL FLOWS TO DEVELOPING COUNTRIES tribution in reducing the potential systemic impaci of the other sources of risk. International norms. Given increasing cross-border flows and the efforts by many developing countries to improve their infrastructures, a central issue during the late 1980s was to establish best practice and harmonize systems across countries. The landmark effort in this area was the G-30 initiative in 1989 that established what was to become best practice in trading and settlement for some six years.19 Many developing and industrial countries, as well as international industry associations and organizations such as the International Federation of Stock Exchanges (FIBV in French) and iosco, endorsed the G-30 rec- ommendations.20 In 1995, the International Society of Securities Administrators (ISSA) organized a series of workshops to update the G-30 recommendations, taking into account changing technological capabilities and evolving views in the industry.21 Annex 6.1 lists the original recommendations and the suggested ISSA revisions, to which we will turn when best practice is described below. This section will discuss four aspects of market infrastructure that have been the focus of foreign investor concerns in emerging markets: the trade, comparison, clearance and settlement (C&S), and depository systems. The design of these systems, especially the last three, is also critical for reducing systemic risk. Why Emerging Markets Improve Their Trading Systems Functions and characteristics of trading systems. A good trading system should reduce transaction costs, ease the price ciscovery process by facilitating the incorporation of information into the price of securi- ties, reduce volatility, and increase liquidity. Thete is no best system; rather, each system seems to work best in attaining one or two of these four market attributes, sometimes at the cost of the others.22 Annex 6.2 describes the main typology of trading systems and the basic tradeoffs among them. Whether a particular trading system should be adopted depends on the circumstances of the market and the objectives. Because different systems have different advantages, equity markets are increasingly adopting two or more trading systems, depending on the characteristics of the stock being traded or the investor. Indeed, there seems to be a growing consensus on the desir- ability of having several trading systems in a market. However, the 326 PREPARING CAPITAL MARKETS FOR FINANCIAL INT'g4 market and regulatory institutions need to be capable of handling the potentially complex interrelationships between the different systems and ensure transparency and fairness for all investors. The impact of financial integration. Financial integration does not change these basic tradeoffs and affects the choice of the system only indirectly, in four ways. First, the trading system should be able to cope with the surge in activity that accompanies financial integration. Second, foreign investors may add their (powerful) voices to those of domestic market participants who are concerned about transparency of the market, most likely small investors and the regulatory authori- ties. This concern for transparency reflects a perception on the part of foreign investors that they are less informed than domestic partici- pants and thus are more likely to be hurt by abusive practices such as front-running and insider trading. Third, foreign investors are likely to undertake large trades relative to the size of many emerging mar- kets and prefer systems that add most to liquidity and immediacy so they are able to conduct transactions rapidly and at a lower cost. Given these preferences, market authorities face a dilemma, since the systems that maximize transparency are not the same ones that most enhance liquidity and immediacy. And fourth, given the premium placed by foreign investors on efficient market infrastructure, market microstructures should facilitate other aspects of market infrastruc- ture, in particular trade comparison and C&S. The Need for Improvement in Comparison Systems A comparison system facilitates the comparison of trade details between the counterparties, and if the details do not match, the trade is not al- lowed to settle. The comparison system, by facilitating C&S as well as by being a prime source of information about market transactions, is a basic building block of a well-functioning capital market.23 The key match criteria are trade date, security traded, face value or number of shares, and price and currency.24 To accelerate the settlement flow, a good matching system should work quickly. The G-30 recommenda- tions indicated that trade matches between direct market participants should be accomplished by trade date plus one (T+ 1). ISSA has revised the norm to T+O, that is, the same day of the trade. A good system, how- ever, also needs to be accurate to avoid failed trades at settlement. For some emerging markets, it may be more cost-effective to compare trades 327 T-C'AP1TAL FLOWS TO DEVELOPING COUNTRIES at T+1 or even later to ensure accuracy. The ratio of compared to un- matched trades and the speed with which unmatched trades are resolved are good indicators of the quality of the matching system. The other key norms and practices for trade comparison are: * A centralized system by which the two counterparties submit in- formation on the match criteria to a central agency that does a two-sided comparison. The system is usually operated by the ex- change or a central clearing agency that integrates matching with the C&S systems, enhancing efficiency through standardization and computerization. Computers are cost-efficient for all but the smallest markets, allowing quick and accurate matching of many trades and criteria. * Locked-in trades (trades compared at execution) are the norm in markets using a computerized system foi trade execution. Locked-in trade systems are especially efficiernt in markets with many retail investors. * An efficient one-sided system for large (institutional) investors that are not members of the two-side system--that is, of an ex- change or clearing agency. In these cases, the broker or custodian acts as intermediary and requests confirmation of trade details from the investor,, However, large investors are increasingly using the central system or an electronic confirmation system. Both ISSA and the G-30 recommend that all indirect market partici- pants be members of a trade comparison system that achieves positive affirmation of trade details by T+ 1. The Need for Improvement in Clearance and Settlemient Systems Desirable characteristics of C&S-the investors' perspectihe. From the point of view of market participants, these systems shouLld facilitate settle- ment by: * Reducing counterparty risks by ensuring DMT, preferably with same-day funds (that is, funds that are available the same day they are deposited).25 A central depository (see below) linked with a payments system would facilitate achieving DvP. This link can be achieved by locating all systems within the sarne institution and making simultaneous transfers or by linking the systems. In the 328 PREPARING CAPITAL MARKETS FOR FINANCIAL INT latter case, assets would be provisionally deposited in the buyer's account, pending receipt of funds through the payments system. ISSA has recently defined more precisely the desirable characteris- tics of a DVP system.26 * Reducing operational risk and problems (lost transactions, bad record-keeping, computer and power problems) that disrupt the C&S process. The settlement time frame adopted by a market will affect operational and market risks. Some markets use an ac- count day cycle by which all trades for a given period are sched- uled to settle on a given day. Other markets use rolling settlement, by which trades are scheduled for settlement a certain number of days after execution. Rolling settlement has the ad- vantage of effectively limiting the number of outstanding clear- ances and reducing the time between trade and settlement dates. Both G-30 and ISSA recommend rolling settlement at T+3. In ad- dition, the clearing system should have backup systems capable of completing the daily processing the day after a failure and good data security. * Minimizing the movement of money and securities, thereby re- ducing transaction costs. This is usually achieved by instituting netting between market participants for both securities and cash, rather than relying on a trade-for-trade system that indepen- dently clears and settles each individual trade. Netting can be daily bilateral or multilateral, or continuous net settlement, in which open positions at the end of a day are offset against the next day's trade until settlement. A multilateral netting system re- quires a central C&S entity that acts as a guarantor of all trades, taking on the postnetting counterparty risk. In other words, a member's net obligations to deliver securities or funds are to the central agency, not to the original counterparty. A key point is that these systems require a legal framework that accepts and en- forces net obligations transferred to a third party. In building a C&S system there may be a difficult tradeoff between reducing transaction costs through netting and reducing counterparty risks. In a trade-for-trade gross settlement system, exposure to each party can be tracked, but the system can be prone to breakdown as vol- ume increases, and cash requirements and cost are large. 'While the re- cent development of communication and computer technology has 329 YE CAPITAL FLOWS TO DEVELOPING COUNTRIES made gross settlement feasible in the more sophisticated markets, net- ting could bring significant savings for markets with growing or already high volume, and with currently high settlement costs and trade failure rates.27 According to Stehm (1996), multilateral netting reduces the vol- ume and value of setdements by 70 to 98 percent, depending on trad- ing patterns. However, multilateral netting increases uncertainty, since the final counterparty is unknown, a fact that has a chilling effect on the market. Even with trade-for-trade settlement, the market may suffer from similar uncertainties if market participants are unable to evaluate each other's financial position because of lack of information. A central clearing agency that acts as a guarantor may reduce the uncertainty faced by individual participants but may increase systemic risk. Risk management techniques. From the point of view of the authorities responsible for market development and regulation, and indeed for the market as a whole, a key design issue for a central clearing agency is risk reduction, since the purpose of the agency is 1 o concentrate setde- ment risks. In addition, if it acts as a guarantor ol all transactions, the incentives for market participants to evaluate prudently counterparty risk are weak, since the central agency is guaranteeing the transaction. Clearing agencies have developed several methods to reduce settle- ment risk. One form of risk reduction is to set stiingent standards for membership in the central clearing agency, although there is a potential tradeoff between achieving economies of scale and promoting broad participation and competition within the market. [n addition, in many emerging markets, given inadequate information and weak disclosure standards, the clearing agency may have only a limited ability to de- velop a comprehensive picture of its members. Risk reduction can also be achieved through net debit caps that limit the net exposure of the clearing agency to each participant and require collateral for some types of exposure. For either procedure to be effective in containing risk, however, the clearing agency must develop a risk control system that marks to market and monitors, in real time, the exposure of its mem- bers.28 More important, the clearing agency should design its proce- dures so that participants themselves have more of an incentive to manage and contain the risks they bear. For example, loss-sharing rules among participants in the event of a default coulc[ be established with a view to increasing the incentives for participants to assess prudently counterparty risk.29 In addition, loss-sharing rules should also clearly define the potential liabilities of the central agenqc and participants to 330 PREPARING CAPITAL MARKETS FOR FINANCIAL INTE T reduce uncertainty during periods of market stress. Finally, as a last de- fense, the central clearing agency should have quick access to sufficient liquidity and be adequately capitalized. Reflecting the increasing concern with systemic risk, international norms in this area have been modified since the G-30 issued its recom- mendations in 1989. While the G-30 encouraged the adoption of net- ting in a high-volume market to enhance market efficiencies and to reduce counterparty risk, ISSA stresses reducing overall settlement risk. ISSA recommends real-time gross settlement or a netting system that fully meets the Lamfalussy Recommendations for multilateral netting systems.30 Among the most important of these recommendations is that the netting system should be capable of completing daily settle- ment in the event of the default of the largest single net-debit position. Its procedures should also encourage participants to monitor and con- tain their own credit exposure. The Need for Improvement in Depository Systems Central depository systems have four main functions: * They maintain facilities for the deposit, withdrawal, and safe- keeping of securities. A registry maintains a record of who owns the shares whereas the depository maintains records of ownership for market participants. Good coordination between the two is essential, in particular in maintaining records regarding the secu- rities held in the nominee name of the depository. * They facilitate achieving settlement on a DVP basis by easing the de- livery or recording the book entry transfer of securities against si- multaneous payment between members (in accordance with members' instructions or resulting from a netting system), and maintaining an institutional delivery system that allows brokers and dealers to deliver securities to indirect market participants. * They simplify postsettlement actions such as registration, pay- ment of interest and dividends, and other corporate actions. * They maintain links with depositories in other markets. How- ever, these linkages are too complex and costly and have not really developed (Stehm 1996). Physical transfer versus book entry. One key distinguishing characteris- tic among depository systems is whether or not settlement involves 331 ,AI'TAL FLOWS TO DEVELOPING COUNTRIES the physical transfer of securities. Book entry is t-he process of settle- ment without the physical movement of securiiLies. It requires that physical documents or certificates be either immobilized (stored and lodged at the depository) or dematerialized (when no physical securi- ties are issued at all), with transfer of property taking place as a book (or computer) entry only.3" Immobilization can result in large savings for the entire industry, especially when securities are fungible and the depository can act as a nominee for the beneficial owners. Dematerial- ization can further reduce costs by eliminating the costs of custody of physical securities and of maintaining records of ownership on a seg- regated certificate basis. Legal constraints and adverse investor reac- tion can be a barrier to immobilization and dematerialization, as will large numbers of certificates from many issuers. Direct versus indirect holding systems. Another important distinguishing characteristic is whether the beneficial owners (that is, the investors) hold their securities directly at the registry or through the accounts of financial intermediaries.32 In an indirect system, the most common, an investor's securities are held at the depository i a the nominee name of the investor's custodian. In turn, the depository's holdings are regis- tered in its nominee name. In a direct system, the holdings of investors are registered directly at the registry or the depository, which can be merged into one institution. Both systems have disadvantages: most important in an emerging market context, while direct holding systems may discourage competition among custodians and brokers to provide quality services to investors, indirect systems may be plagued with investor protection issues that arise when financial intermediaries hold shares on behalf of their clients.33 The two systems, however, can be combined or tailored to attenuate these problems. In the short term, what is crucial to attract foreign, in parnicular institutional, investors, is that there be financial institutions of high quality that can provide good and safe custodial services. Successful emerging markets have welcomed the establishment of global cusitodians within their borders or encouraged strong links between these custodians and domestic financial institutions. The Experience in Emerging Markets There are four key conclusions that can be drawn from the experience in emerging markets in the area of market infrastructure. First, market 332 PREPARING CAPITAL MARKETS FOR FINANCIAL INTt infrastructure in most emerging markets still compares unfavorably with that in mature markets. Second, infrastructure standards in the more dynamic emerging markets in East Asia and Latin America are, however, approaching international norms, and a number of these mar- kets meet many of the G-30/ISSA standards. As described in box 6.4, East Asia has the made the most progress in market infrastructure in re- cent years and is on its way to becoming the region with the best mar- ket infrastructure. Efforts to upgrade infrastructure are, however, ongoing in almost all markets, and the situation is rapidly evolving. Third, progress has been weakest in adopting those standards designed to reduce counterparty and systemic risk, especially in achieving DVP. Fourth, many emerging markets, in particular in Asia, have adopted computerized automatic matching trading systems to enhance market transparency, despite the fact that these systems are not the best suited to increasing liquidity for large orders. The gap between emerging and mature markets. There is still a gap between emerging markets and mature markets such as the United States. Figure 6.3 compares, for a selected group of developing and industrial countries in 1995, the reliability of settlement and the effi- ciency in collecting dividends and reregistering securities-two key custody functions. In all three areas, industrial countries are shown to be more efficient-by a factor of 10 to 1 regarding the average num- ber of days it takes to collect dividends, 20 to 1 regarding the average number of days to register, and between 2 and 3 to 1 regarding trades that settle with a delay. There is also a wide range of performance among emerging markets. For some countries, such as India, the gap with industrial markets is very striking-the proportion of trades that settle after the contractual settlement date is about 75 percent, com- pared with 10 percent or less in most industrial country markets. Other markets, such as Korea and Mexico, have a performance close to or even better than that of industrial markets. Meeting the G-30 recommendations. Emerging markets now meet many of the G-30 standards, as shown in table 6.4. Developing countries, in particular in Asia, as described in box 6.4, are upgrading their infra- structures very quickly, trying to keep up with rapidly growing trad- ing volumes, by leapfrogging to state-of-the-art systems. For example, 12 out of the 16 markets included in the table have central deposito- ries, and the remaining 4-India, Indonesia, Pakistan, and the Philippines-are all scheduled to have central depositories by mid- 333 _vI APTAL FLOWS TO DEVELOPING COUNTRIES Box 6A Developing Capital Market Infrastructure in Asia and Latin America ASIA HAS MADE REMARKABLE PROGRESS OVER and a large settlement fund, with rolling settlement the last four years in improving capital market infra- expected to be launched within weeks. structure. For example, as shown above, the effi- The figure suggests that in contrast to those of ciency of Malaysia's infrastructure (based on a Asia, the more advanced capital markets in Latin measure that takes into account the ease and relia- America started the decade with relatively well func- bility of settlement and postsettlement actions) just tioning infrastructures and have maintained this about doubled between 1992 and the first quarter of performance throughout the decade. In some cases, 1996. Indonesia's performance was almost as good. such as Argentina and Chile, however, the surge in India's successful effort to set up from scratch the trading volume has been smaller than in East Asia, National Stock Exchange (NSE) in about two years is so market infrastructure has not been under the another good example of this progress. The early same stress. As a result, there has been less urgency focus of the NSE was on improving the transparency to upgrade these systems and adopt G-30 bench- of trading, developing liquidity, and shortening set- marks. Indeed, although Chile's C&S system is ca- tlement cycles and risk. To do so, the NSE instituted pable of implementing netting, this has not been a compu,terized trading system and a clearing corpo- done so that the authorities can monitor compliance ration, while a central depository is expected to be with capital controls (Bank of New York 1996). operationial in India during the next six months. The This is not the case for Brazil antd Mexico, which NSE has also introduced a weekly settlement cycle have had larger increases in volumes. 1997. These improvements in market infrastructure are remarkable. Some five years ago, most developing markets were just beginning to improve their market infrastructure, and many market analysts thought that central depositories were only pipe dreams. Table 6.4 suggests that shortening the settlement cycle to the G-30 benchmark of T+3 through more efficient comparison systems, net- ting, and central depositories has been a key focus of all this effort. Many of the countries that do not meet the T+3 ob ective are now close to it, some settling at T+ 4 (Indonesia and the I'hilippines) or T+5 (Malaysia and Sri Lanka). But this emphasis on speed may not be fully justified. Foreign investors do not necessarily demand G-30 speed stan- dards, but rather reliability: they prefer a system that settles in five days but in a predictable process with no undue backlogs and failed trades to one that is faster but less reliable. Malaysia, one of the more successful emerging markets, makes this point very well: although only settling at T+5, as figure 6.3 suggests, it has been able to achieve one of the more reliable settlement systems in the developing world. 334 PREPARING CAPITAL MARKETS FOR FINANCIAL INT1c Box Figure 6A Market Infastructure in Asia and Latin Amneuica ~~~~~~~~~~~~~100- ___ 0-_ r InOnesi _ Un"it Sa ts U Mayia - ntdStae 100 - 1.00-- . _._-1 so ---- ----------- so o 0.~~~~~~~~~~~~ 1992 :M1992 1993 1993 1995 1996 U *Avgntban _ United States U _f2 Unteld States Note: The operationlal benchmark plotted above is a acore Oult of a maximum of 100. A higher score indicates higher settlement and poarsettlement efficiency and a lower overall operational risk associated with the market. Efficiency is measured takting into account the numnber and operational costa of failed and delayed transactions, the cost of administrative effort, the complexity of the market, an d compliance with the recommendations of the Group of Thirty. The date foir 1996 refers to the first quarter. Source: Global Securities Consulting Service, Ltd. 335 AITAL FLOWS TO DEVELOPING COUNTRIES Figure 6.3 Settlement and Postsettlement Efficiency, Selectedl Emerging and Developed Maikets, 1995 A. Average Number of Days Taken to Collect Dividends B. Average Number of Days 1Taken for Reregistration' Philippines _----- *_-____India_ __ Turkey -PhilIppines * Perau _ _____ __=Venezuela Venezuela Indonesia Indonesia Argentina - -- Thailand Malaysia Ta_ a Brazil Malaysia Thailand* Turkey Mexico, u - t - 4 Korea Argentina Singapore - ---------- Brazil United Kingdomn ital) -- -Korea Japan -- ' -- - - Peru United States France 5Mexico Germany _____ ____ __ __ 0 10 20 30 40 50 60 70 80 90 100 110 0 5 10 15 20 25 30 35 Days Days C. Settlement Performance India - j- Venezuela _ _ _ Philippines _ Indonesia _ _ _ _ ___ Argentina Share of transactions that __________________ __________________settle on: Thailand S-day (contractual date Malaysia of settlement] Greece Mexico * S-day + to S-day + 5 Xorea y Sday + 6 to S-day + 15 Turkey _ _ ______- Brazil * S-day + 15 Japan United Kingdom Germany ._ ._ United States France Capital market infrastructure in 0 20 40 60 80 100 emerging markets is less reli- Share of transactions (percent) able and ifficient than that of more mature markets. a In most deN eloped markets, reregistration is immediate Source. Global Securities Consulting Service, Review of Emerging Markets and Review of Major Markets 336 PREPARING CAPITAL MARKETS FOR FINANCIAL 1 T N Table 6.4 Conformity with the Group of 30 Recommendations Trade comparison Same- system CT +1) Deliveoy day Tt3 Direct Indirect central Trade verss finsk rofing &Scuriie Country participanis paricipant deposities netting payment pament stement enng Argentina Yes Yes Yes Yes Yes Brazil Yes Yes Yes Yes Yes Yes Yes Chile Yes Yes Yes China Yes Yes Yes Yes Yes Yes India Yes Indonesia Yes Yes Korea Yes Yes Yes Yes Yes Yes Yes Malaysia Yes Yes Yes Yes Yes Mexico Yes Yes Yesa Yesa Yes Yes Pakistan Yes Yes Philippines Poland Yes Yes Yes Russia Yes SriLanka Yes Yes Yes Thailand Yes Yes Yes Yes Yes Turkey Yes Yes Yes Yes Yes a. See box 6.5. Soserce: Data from Bank of New York and Citibank, iSsA (1995). Least progress has been accomplished in achieving DVP in the settlement process. With central depositories, the delivery side of the equation is gen- erally working well. It is on the payment side that DVP seems to fail, perhaps because of weaknesses in the domestic banking and payment systems. Emerging markets are computerzing. As described in annex 6.2, many emerging markets have adopted computerized order-driven automatic matching systems. Computers can handle large volumes in a cost-effi- cient manner and can trade around the clock. They also enhance trans- parency by generating a detailed audit trail as well as voluminous real- time information regarding market transactions. However, computer- ized systems are not good at handling large orders in a market with lit- tle liquidity. The fact that many developing countries have adopted such computer-based order-driven systems suggests that this concern is outweighed by its advantages, including its facilitation of other aspects of market infrastructure such as trade comparison and C&S. 337 Rt-APTAL FLOWS TO DEVELOPING COUNTRIES Best Practice and Lessons from Country Experience The G-30 recommendations and the revisions suggested by ISSA are a good indication of what constitutes best practice for infrastructure in se- curities markets, for both stocks and fixed income securities. However, the discussion above also indicates that benchmarks need to be tailored to the needs and possibilities of each emerging market. In particular, poli- cymakers and market participants need to carefully consider the tradeoffs that arise from adopting particular institutional arrangements, taking into account local circumstances when deciding, for example, whether to set up a central clearing agency. With this important caveat, the main conclusions on best practice for improving market infrastructure are: * Matching the matching system should be integrated with the C&S system, and all trades by direct and indirect (institutional) participants should be matched at most by T+ 1. * Clearance and settlement. settlement should be accomplished by a DVP system with same-day funds; there should be real-time gross set- dement or a netting system that meets stringent risk control stan- dards, depending on the characteristics of the market; and the rolling settlement system should have final settdement occur by T+3. * Depository there should be one independent central depository managed for the benefit of the industry, broadly defined; there should be an independent registry or registries; and immobiliza- tion and dematerialization should be encouraged, and the legal framework revised, if necessary, to permit them.34 Mexico is a good example of how countries are adapting these stan- dards to local circumstances. One peculiarity of the Mexican infrastruc- ture is that while securities are settled on a gross basis, the central bank (Banco de Mexico) functions as the central clearing agency for the money side of market transactions. Banco de Mexico basically functions as the guarantor of the netting system, although it has recently introduced sev- eral measures to reduce systemic risk and change the incentives for market participants. Box 6.5 describes the Mexican experierLce in more detail. The G-30/IssA benchmarks do not say very muchI about the dynamics of constructing well-functioning capital markets. The problem with for- mulating specific recommendations on how to facilitate this process is that market and country circumstances are critical for deciding policy and institutional reform priorities. For example, in India the authorities de- 338 PREPARING CAPITAL MARKETS FOR FINANCIAL INT-- cided that the most cost-effective manner of improving market infrastruc- ture would be to construct from scratch a modern equities exchange, with the expectation that competition would promote improvements in the other exchanges. But this approach may not be advisable in markets where existing exchanges are able to overcome vested interests and adopt modern systems. Notwithstanding this caveat, basing our conclusions on both country experience and the work undertaken to develop international benchmarks, we can make the following recommendations: * The G-30/IssA benchmarks are objectives to be attained over time. While it is possible and in some cases advisable to leapfrog to modern systems, efficiency and reliability rather than speed of settlement should be the primary objectives. The example of Malaysia shows that investors and issuers will not be discouraged if the market does not fully conform to these benchmarks. * Reducing systemic risk should, from the start, be an important objective in the design of market infrastructure. The stylized at- tributes of emerging markets suggest that they are more prone to systemic problems than are developed markets.35 This suggests early attention to the interface with the banking and payments systems to facilitate attaining DVP. More controversial is the rec- ommendation that, despite the possible negative consequences for market development, it may be advisable to have strict mem- bership criteria for exchanges and clearinghouses, at least during the transition period, until other risk reduction systems are in place and domestic investment firms become financially stronger. Opening up domestic markets to well-capitalized foreign inter- mediaries may be a way of increasing competition in the market without increasing systemic risk. * Country experience suggests that setting up a central depository should be encouraged early. While it is possible to have a well- functioning C&S system without one, a depository brings large benefits by reducing transaction costs and operational risks and facilitating DVP. That all the countries in table 6.3 have or soon will have central depositories reveals a strong preference for cen- tral depositories by the most dynamic emerging markets. How- ever, an essential prerequisite for a central depository is the legal basis for nominee ownership.36 In addition, there must be suffi- cient volume for the central depository to be cost-effective. 339 *LTAL FLOWS TO DEVELOPING COUNTRIES Box 6-5 The Anatomy of Bolsa de Mexico's Clearance, Settlement, and Depository System MEXICO HAS BEEN ABLE TO GRADUALLY IMPROVE settlement operations involving securities are inter- its clearance, settlement, and depository systems, nal to Indeval, reducing operational risks. moving toward international norms while taking Indeval has also been gradually improving its ef- into account local conditions. This process is a good ficiency as a depository. It was founded in 1978 as a example of how to apply in practice some of the state-owned institution but was privatized in 1987. principles of a good infrastructure. Today, it is owned by financial institutions, includ- First, Mexico has rationalized and consolidated ing banks, brokerage firms, and insurance compa- the institutional structures of the clearance, settle- nies, on a nonprofit basis. Since 1984, all Mexican ment, and depository systems. One institution, In- securities have had to be in registered form, and deval, is the sole depository for equities, private these securities are transferred by Indeval either by bonds, oammercial paper, and bank acceptances, as book entry or by physical delivery, The transaction well as the clearinghouse for these instrutments. It is registered in the name of the current stockholder, shares the latter responsibility with the Central Bank and, if physical delivery is made, the stockholder's (Banco de Mexico) and the Mexican Stock Ex- name must appear on the stock certificate. Indeval change (Bolsa de Mexico), which are responsible for then advises the corporation about the changes to be the money side of transactions. As the sole deposi- made in its stockholders' register. The long-term ob- tory, Indeval has succeeded in minimizing the con- jective of Indeval is to dematerializu all securities. To tinual movement of documents, thereby reducing this end, it is promoting the use of jumbo or global the risk of loss, destruction, or forgery. As the sole certificates. Indeval also provides extensive postset- clearinghouse, Indeval processes all transactions, dement services for securities owrters and custodi- whether executed on Bolsa de Mexico or outside it, ans, as well as issuers. In an elfort to simplify and settles them operation by operation. Hence, all cross-border transactions, it has established a link Country experience, in particular that of Asia, also brings good news to nascent markets. A key lesson that can be drawn from this ex- perience is that while these markets need to develop appropriate in- frastructure to prepare for international financial integration, with a sustained and well-organized effort they can achieve rapid progress. The Regulatory Challenges of Financiial Integration ARE THE OBJECTIVES OF THE LEGAL AND REGULA- \.^/ tory framework? The legal framework is the basis on t vV which capital (and other) markets function. Essentially, it should provide for the effective enforcement of private contracts 340 PREPARING CAPITAL MARKETS FOR FINANCIAL CI AL- with Cedel and the International Securities Clearing throughout the day. Banco de Mexico will make Corporation in the United States, and it is also one net debit or credit to each financial institution's studying options with similar institutions in other account at the Banco de Mexico at the end of the countries. processing cycle. With regard to incentives, if a par- In regard to the settlement system, in April ticipant's account at Banco de Mexico is overdrawn 1995, Banco de Mexico implemented structural in terms of cash and collateral for three business changes to reduce settlement risk. The objectives of days, Banco de Mexico wil debit the accounts of all these changes were to attain a DVP environment and market members who gave settlement lines to the reduce systemic risk by instituting better risk con- participant under stress. The debit will be propor- trol procedures and increasing the incentives of tional to the settlement lines. Industry sources, market participants to monitor their counterpart however, report that the trading system does not exposure. To this end, Banco de Mexico imple- allow identifying easily who counterparties are. mented a new payment system that monitors the While Mexico's clearance and settlement system ability of a participant to effect payments, as a func- is effective, it does not fully conform with the G-30 tion of the participant's cash, collateral, and line of recommendations in two respects. The system does credit at Banco de Mexico. These risk control mea- not permit trade comparison for indirect market sures should prevent participants from effecting participants, and the practice of securities lending payments in excess of their credit lines. They are and borrowing is not operational, although it re- also linked to Indeval's clearing system, so securities cently received regulatory approval. will be credited to a participant only if the partici- pant can pay for them. Securities will be transferred Soure. lSSA (1996), author's analysis of data from on a real-time basis, while cash will be netted Citibank and Bank of New York. and form the foundation for instruments and practices necessary to the functioning of modern capital markets.37 Building on the legal infrastructure, the regulatory authorities must make markets fair, efficient, and safe. The presumption is that without regulations, capital markets would not result in a socially efficient outcome because of market imperfections (such as incomplete or asymmetric information) and negative externalities (systemic risk). Fair markets are those in which the investor is protected from abuse and fraud and similar market participants are treated equally. Efficient markets are competitive and have efficient infrastructure and information systems. The regulatory framework should promote and protect competition by ensuring that market practices and rules do not im- pose any unnecessary burden on competition, unless required for the 341 APAPITAL FLOWS TO DEVELOPING COUNTRIES pursuit of the other regulatory goals. Safe markets are those pro- tected against systemic risk. Best practice. There is a convergence in many emerging markets toward a regulatory model based on public disclosure and on self- imposed market and industry discipline, along with better internal risk management by financial firms themselves.38 The premises of the model, and associated key regulations, are discussecl below. * First, the goals of the regulator are not to substitute for the mar- ket in making investment decisions, but to ensure that the incen- tives and structure of the market are consistent with efficiency, fairness, and safety. The regulatory frameworlk should ensure that timely and accurate information is available so that investors can judge the merits of alternative investments, hence the regulatory emphasis on disclosure and eradicating fraud. The regulator should prohibit insider trading for fairness reasons and because such practices have the negative externality of discouraging in- vestors and savers from participating in capital markets. * Second, market participants, rather than government, should be mainly responsible for establishing and enforcing market rules and regulations. The rationale is that participants have an interest in ensuring that markets are fair and efficient, and are better able to judge how to make them so. In practical terms, this means that self-regulatory organizations (SRos) such as the exchanges, bro- ker-dealer associations, and accounting and auditing associations will bear much of the responsibility for the regulation and sur- veillance of securities markets and auxiliary supporting services. For the system to work correctly, the official regulator must have sufficient regulatory oversight to ensure that the SROs enforce se- curities regulations as well as their own market conduct rules and that they act to minimize potential conflicts of interest and re- straints on competition. * Third, concerns about systemic risk in capital markets do not jus- tify prudential regulations and monitoring as intense as those in the banking sector. As a rule, investment firms are less vulnerable than banks, on both the asset side and the liability side, to liquid- ity and solvency crises, and are less prone io contagion.39 An exception to this rule regards clearance and settlement arrange- ments, discussed above. However, with the increasing integration 342 PREPARING CAPITAL MARKETS FOR FINANCIAL I of banks and securities businesses, there is a much stronger ratio- nale for intensifying prudential rules and oversight of the latter.40 * Fourth, the protection of investors' assets from loss and the insol- vency of investment firms is another rationale for prudential rules and oversight of investment firms. Since investors may not be able easily to evaluate the riskiness of financial institutions through which they conduct business and hold assets, such rules and monitoring would promote market confidence.41 To this ef- fect, investment firms are often required to meet capital adequacy standards, segregate investors' assets, and meet minimum stan- dards with regard to internal mechanisms for identifying, ac- counting for, and safeguarding client assets.42 In addition, the legal system should provide the legal basis for segregating in- vestors' assets. Some markets also make provisions to compensate clients for losses resulting from the insolvency of the firm holding their assets (for example, investor protection funds). Disclosure to investors of the level and types of risks to which their assets are exposed, as well as the status of financial institutions, is essential for market discipline to play a role in reducing these risks.43 * Fifth, financial innovation, in particular the development of a wide variety of derivative products, is leading to some important changes in how financial firms are regulated. Although deriva- tive trading still involves price risk, the speed with which these risks can be transformed and the opacity of the transformation process make it difficult for regulators to assess the degree of ex- posure of financial firms. The regulatory response to this lack of transparency has been to focus on the ability of each firm to manage these risks and to create incentives for financial firms to put in place appropriate risk control procedures. This is a rela- tively new area of concern in emerging markets, since derivative products are not widely used (except in Brazil and Malaysia) or may actually be proscribed. However, as financial integration in- tensifies, domestic financial firms may trade in these instruments overseas. Regulatory agencies in many industrial and developing countries are also mandated to promote the development of domestic capital mar- kets. Obviously, the functions and key regulations described above are essential to foster market confidence and growth. In some cases, how- 343 CIAPITAL FLOWS TO DEVELOPING COUNTRIES ever, the regulator needs to carefully assess the potential short-term tradeoff between investor protection and safety on the one hand, and market development or liquidity on the other. More broadly, the cost of a regulation and its enforcement needs to be weighed against its ben- efits. In addition, in order to promote market development, regulatory agencies will sometimes need to perform functions outside their strictly defined regulatory responsibilities. Box 6.6 lists some of the most im- portant development tasks and describes how ceitain countries have avoided a common regulatory-development tradeoff. Preparing for financial integration. Financial integration brings increased urgency to the task of constructing and reinforcing the regulatory framework in emerging securities markets. As with building market infrastructure, this task is essential if developing countries are to com- pete in an integrating world, attract foreign investors, and reduce sys- temic risk. For example, investors want both macro data on economic prospects and micro data on corporate performance, to be able to make informed investment choices. Improving disclosure will not only address investor concerns but will also reduce the susceptibility of the market to volatility resulting from incomplete or asymmetric information. Moreover, financial integration brings new challenges for the regula- tory authorities. Globalization of financial transactions and intermedi- aries will require regulators to find new tools and approaches to monitor effectively market risk and the financial soundness of market participants. Domestic financial firms will be exposed to new market and counterparty risks as they increase their international activities, and new foreign intermediaries may become active in domestic mar- kets. In addition, developing countries have political economy con- cerns regarding foreign ownership of domestic firms, and some of them have responded through restrictions on foreign investment that seem to have little economic justification. Given this background, this section will first address four main is- sues: the costs and benefits of direct restrictions on foreign investment in domestic capital markets, the special regulatory issues that arise from the globalization of financial intermediaries and transactions, the most important concerns of foreign investors regarding the regulatory frame- work, and enforcement and the role of sROs. Finally, the section will de- scribe some cross-country findings on how emerging markets are dealing with these issues. 344 PREPARING CAPITAL MARKETS FOR FINANCIAL INTT> Box 6.6 The Regulator and Market Development REGULATORY AGENCIES THAT FOSTER MARKET One of the most common tradeoffs between in- development also perform other functions; these vestor protection and market development regards include establishing the legal and regulatory prereq- listing requirements imposed by regulators to ensure uisites for domestic institutional investors, encour- that companies traded on an exchange meet mini- aging the development of financial institutions and mum standards. If listing requirements are too strin- infrastructure (such as credit-rating agencies), and gent, few firms will be able to list and the market human capital formation. With regard to the last, will not be liquid. One solution to this dilemma is to well-qualified SROs can play a critical role by train- segregate the market, allowing firms with less-estab- ing, testing, and licensing their members, in particu- lished track records to list in a special, perhaps over- lar exchanges and broker-dealer organizations. sROs the-counter, market for less risk-averse investors. can effectively provide such training and testing ser- This system has worked well in industrial countries; vices because of their knowledge of the industry and one example is NASDAQ in the United States. Emerg- their self-interest in developing and maintaining the ing markets have also successfully segregated their investor public's trust. For example, the Korean and markets. Korea's stock market, for example, is di- Malaysian securities dealers associations offer a wide vided into two trading sections, with newly listed range of training programs. For broker and dealers, stocks automatically listed in the second section for training in the following areas is particularly impor- at least one year. The Korean stock exchange evalu- tant: general knowledge of securities markets and ates the annual reports of all listed companies to de- regulations, customer relations (including suitability termine whether second section firms meet the requirements and sales practices), and back office requirements to be promoted to the first section, as matters (management of customer funds, record- well as whether first section firms should be moved keeping, and so forth). Good practices build the back into the second. The key requirements for the confidence of both domestic and foreign investors. first section are that the stock is widely held and that Investor education is another critical task, in partic- the company has a demonstrated track record. ular so that investors understand the risks and re- Korea also has an over-the-counter market with even wards of investing in securities and how to make less rigorous listing requirements. Other countries informed investment decisions. with similar practices include Mexico and Poland. Restrictions on Foreign Ownership Typology and rationale. Restrictions on foreign investment are not un- common in emerging securities markets. Typically, there is a ceiling on the proportion of a firm's equity that can be owned by a single for- eign investor plus an aggregate ceiling on ownership by all foreigners. Ceilings may vary among sectors, generally with lower limits for banks, financial firms, and certain strategic sectors. In some emerging markets, a firm can impose even more restrictive limits for its own shares. There may also be differential treatment by type of foreign 345 C-APITAL FLOWS TO DEVELOPING COUNTRIES investor, with institutional investors often being given preference. Foreign investors may also be subject to special regIstration require- ments, which, depending on their severity and on the efficiency of the process, could restrain foreign investment. In ac[dition, controls or administrative requireiments that delay the repatriation of capital gains and dividends could severely curtail foreign participation in a coun- try's capital markets.44 Restrictions are generally imposed because of political economy concerns about domestic firms being owned by foteigners and because of the need to ensure a level playing field among investors. Part of the fear is that if foreign investors were not restricted, they would capture the great majority of the increases in capital gains and dividends that rapid growth brings, to the detriment of national i ncome. Are restiictions justified? The economic justificatio ri for restricting for- eign ownership is weak. If the limits are binding, they curtail foreign portfolio investment, denying the market the addltional liquidity, as well as constrain the issuing of IPOs by firms that are unable to mobi- lize enough local capital to remain within the ceiling. In addition, ownership restrictions on foreign investors have nr gative implications for corporate governance. Confined to being small shareholders, for- eign investors are often reluctant to spend resources to monitor man- agement because of the free-rider problem of the benefits accruing to others. Finally, often the restrictions are not effective because foreign- ers can avoid the limits through nominee ownership, which further exacerbates the corporate governance problem. This is widely believed to be the case in both Indonesia and Thailand. Any elimination or softening of these restriction on foreign owner- ship needs to take the political economy concerns into account to be sustainable. Some countries, such as Mexico, have instituted investment trusts consisting of shares without voting rights for foreigners who wish to acquire shares of a firm over and above the ceiling. Thailand is con- sidering a similar system. While this solution allovws foreigners to hold more shares, however, corporate governance remains an issue. Creating a domestic investor base. The best solution to these concerns regarding foreign ownership is to develop a strong domestic institu- tional investor base. Such investors will be able to mobilize significant amount of resources, increase liquidity in domestic capital markets, and thereby serve as a counterweight to foreign investors. They also reduce the vulnerability of domestic capital markets in the event of a 346 PREPARING CAPITAL MARKETS FOR FINANCIAL INTE( rapid liquidation of assets by foreign investors. With their deep pock- ets, they would be able to act as an automatic market stabilizer by bottom fishing and value picking when foreign investors as a group are selling. Moreover, active domestic institutional investors, by increasing depth and liquidity in domestic markets, would reduce the sensitivity of domestic markets to small trades. In addition to serving as a counterweight to foreign investors, pension and mutual funds strongly benefit domestic savers by reducing transaction costs and facilitating portfolio diversification. Furthermore, these instruments are managed by professionals who help overcome the information constraints faced by households. Developing a domestic institutional base requires reforming and ex- panding the legal and regulatory framework, in particular in the area of investor protection. Investment fund regulation addresses prudential rules, custodial arrangements to protect investors if the management company becomes insolvent, and rules to prevent fraud and enhance transparency with regard to a fund's objectives and fees, as well as to protect against self-dealing. Expanding the domestic institutional in- vestor base, especially pension funds, also requires reform in a wide va- riety of policy areas. Box 6.7 describes some recent initiatives undertaken in Latin America to establish domestic institutional in- vestors and the obstacles encountered. The Regulatory Implications of Globalization The new risks and challenges. Financial integration leads to rapid in- creases in three types of a ctivities in financial markets: the activities of multinational financial institutions in foreign financial markets, trans- actions between financial intermediaries in different countries, and the cross-border delivery of financial services. These activities affect the na- ture and magnitude of the underlying market, and counterparty and systemic risks in the domestic capital market. In particular: * As domestic financial firms engage in cross-border investments, they may incur new market or price risks from open positions in exchange- and interest-rate-sensitive foreign assets. * Domestic residents engaging in cross-border transactions will be exposed to new counterparty risks with foreign investors and firms. Multinational firms that become active in domestic mar- 347 KCAFTAL FLOWS TO DEVELOPING COUNTRIES kets constitute another new source of counterparty risks for do- mestic residents. In addition, integration exposes domestic resi- dents to foreign jurisdictional risks not under their control. * Systemic risks also increase with financial integration. Most im- mediately, the effects of the collapse or insolvency of a financial firm in another country may be transmitted to domestic markets, either directly, if the firm has established itself in the domestic market, or through the payments and settlements systems. Sys- temic risk also rises because of higher operational risks due to in- creasing volumes and different hours of operation among payment systems; this can complicate the achievement of a DVP environment and delay settlement. More broadly, integration subjects the home country to more sources of external distur- bances and increases the speed at which these disturbances are transmitted across markets. These activities and associated risks are creating new and interrelated regulatory challenges in two broad areas. First, the most obvious chal- lenge is to contain these new sources of systemic arid counterparty risk at the same time that globalization is also rendering the regulatory en- vironment more imperfect. Indeed, with integration, domestic resi- dents, including the regulators, may find it difficult to assess comprehensively the risks incurred by foreign counterparties if they en- gage in multiple cross-border activities around the globe. Integration will also create similar difficulties in assessing domestic intermediaries as they increase their multinational activities. And there is also a danger that financial integration could blur the lines of supervisory responsi- bilitty between domestic and foreign regulators. Second, globalization is magnifying the importance of regulatory issues that arise from cross- country differences in regulations. One important concern of national governments is to ensure a level playing field so that these differences in regulations do not distort the competitive positions of firms and finan- cial centers in different countries. Another concern is the possibility that financial intermediaries will engage in regulatory arbitrage-that is, seek to take advantage of differences in regulat.ons and their cover- age between countries. Two other trends in international financial markets are interacting with globalization, significantly magnifying the iegulatory challenge. First, as discussed above, financial firms are increasingly becoming fi- 348 PREPARING CAPITAL MARKETS FOR FINANCIAL INTE'T nancial conglomerates, combining traditional banking with securities operations and other nonbank financial activities. Second, financial in- novation has resulted in a vast array of new derivative instruments and markets. The main regulatory issue with these instruments is that they permit financial firms to change their risk profiles very quickly and in very complex manners. The combined impact of globalization and these two trends is making the assessment of the financial status and risk profile of financial firms very difficult. This lack of transparency is undermining the ability of the market to police itself and of the au- thorities to regulate. The international regulatory response. National regulatory authorities have recognized that reducing these risks and challenges of globaliza- tion requires stronger international cooperation. This is particularly true with regard to controlling systemic risk, since national regulatory frameworks in this area clearly have an impact across borders. Indeed, there seems to be growing concern among national authorities, heightened by the failure of Barings in 1995, regarding the dangers of unregulated or underregulated financial activities of global firms.45 But even in the investor protection area, where national preferences have less of an effect across borders, there have been long-standing efforts to harmonize regulations across countries and promote inter- national cooperation in enforcement. These efforts have been under- taken in part to reduce transaction costs and increase market efficien- cy, and they are motivated by pressure from institutional investors and international issuers. In addition, these efforts at convergence also address fears of competition between financial centers resulting in a regulatory race to the bottom, and of investors practicing regulatory arbitrage. As the focus of international cooperation among securities regula- tors, IOSCO has played a key role in these endeavors to reduce transac- tion costs and systemic risk. Its main activities have been as follows:46 * To promote regulatory convergence and reduce transaction costs, IOSCO has issued reports and resolutions on best practice, making recommendations in such areas as curbing and punishing securi- ties fraud, disclosure standards, clearance and settlement, and in- vestor protection (which is also important in reducing systemic risk). Its efforts to improve disclosure and reduce transaction costs by promoting accounting uniformity are particularly im- 349 -IPI TAL.FLOWS TO DEVELOPING COUNTRIES Box 6.7 Promoting Domestic Instttional Investors THERE IS CONSIDERABLE SCOPE FOR PROMOT- tected through investment management legislation ing the growth. of securities markets through the de- that is adequately enforced. The legiislation should: velopment of domestic mutual and pension funds. These fimds have become an important form of fi- * Ensure that investors are well informed about nancial intermediation in the 1980s in industrial the fund' invesment objectives and isk profiles countries, but in emerging markets they have and the quality and costs of fimd management. reached respectable sizes only in Brazil, Chile, and * Ensure fair valuation of investors' purchases Mexico in Latin America, and in Malaysia and Thai- and redemptions. land in East Asia. In South Asia, the Unit Trust of * Define prudential and fiduciary-standards. The India is one of the largest mutual fuads in the world extent to which reguation includes detailed in- in terms of number of investors, but overregulation vestment rules regarding the composition of of asset management has reduced its beneficial im- fund portfolios (types of assets and their shares pact on Indian capital markets. As is well known, in the total) varies among countries. Chile has one of the 'most developed private pension * Define, discourage, and regulate improper fund industries among developing countries, man- practices such as self-dealing. aging some 45 percent of GDP in assets. * Protect the integrity of the funds' assets. To be able to develop such funds, there needs first These rules generally segregate the funds' as- to be the legal basis for practices, such as beneficial sets from those of the management company ownership, novation, and trusts, that are central to (to ensure that investors do not suffer in the their development. And second, in order to promote case of insolvency of the management corn- confidence in funds, the investor needs to be pro- pany) and ensure proper custody arrange- portant and are described in more detail elsewhere in this chapter. In addition, given the strong representation of emerging markets at iosco, its resolutions, reports, and accompanying discussions have served as a conduit for the transfer of expertise and experi- ence between countries.47 * iosco has also been quite active in the area of systemic risk, espe- cially during the 1 990s. One key focus has been to contain the sys- temic risks arising from derivative activities. In this regard, losco and the Basle Committee on Banking Supervision have collabo- rated on several reports during 1994-95 on best practice in disclo- sure of both qualitative and quantitative information of derivative activities (for example, Basle Committee and iosco, 1 995a, 1995b).48 Iosco is also fleshing out the reconmmendations made in May 1995 by the regulators of most major futures and options markets (the Windsor Declaration), including protection of cus- tomer positions and assets, and strengthening default procedures. 350 PREPARING CAPITAL MARKETS FOR FINANCIAL INt* ments. Management legislation usually also These design and prudential rules have created bar- requires independent directors on the board riers in the provision of financial management ser- of the fund. vices and perverse incentives, and reduced the potential development of equity narkets. Typicaily, In certain Latin American countries, capital mar- pension funds must be new and specially licensed to kets have received a huge boost from the reform and manage mandated retirement savings (excluding ex- privatization of social security. However, the bene- isting intermediaries) and guarantee a certain return. fits of these reforms and the development of mutual The guaranteed return is defined as a certain per- and pension funds have been delayed by problems centage of the average return of the pension fund in- regarding the structure of the industry as well as ex- dustry. The exclusion of mutual funds has limited cessive regulation. Throughout the region, compet- their growth and their ability to attain economies of ing intermediaries such as banks play a dominant scale. The return guarantee has led pension fund role in the mutual fund industry and have not pur- managers to offer virtually identical portfolios, lim- sued the development of the industry aggressively. iting competition, which in turn has led to high up- In Mexico, banks have reportedly redirected fund front commissions. These problems have been investors into deposits. Custodial arrangements also compounded by regulatory restrictions limiting eq- tend to be costly, with only a few institutions, uities to 20 to 30 percent of pension fund portfolios. mainly banks, being allowed to act as custodians. Remedial action should introduce greater competi- But among the most important obstacles to the tion and portability of these mandated savings, of- development of mutual funds are the design and fering greater freedom to savers to choose among prudential regulation of private pension funds. different portfolios of approved products. * Finally, given that its recommendations are advisory and non- binding to its members, IOSCO has strongly encouraged coordi- nation among members, in particular through bilateral agreements (memorandums of understanding) regarding infor- mation sharing and enforcement.49 While there has been a sharp increase in the number of these bilateral agreements these last few years, many of these understandings have been between industrial countries, some between industrial and developing countries, and only a few between developing countries. The exceptions to this rule are the Latin American countries, which have established a network of such understandings, and, in Asia, Hong Kong, which has signed agreements with many emerging markets in the region. Recently, in response to the Windsor Declaration, ex- changes, clearinghouses, and regulators of the major global fu- tures and option markets have signed information-sharing agreements. 351 AP A-P-ITAL FLOWS TO DEVELOPING COUNTRIES Other efforts at international collaboration in capital markets have focused on clearance, settlement, and payment systems, and regulating financial conglomerates. In the payments area, settlement and systemic risk have been reduced by harmonizing regional payments standards and increasing the overlap of hours of operation of different payment systems. With regard to financial conglomerates, securities, banking, and insurance regulators have been collaborating in developing princi- ples for the supervision of financial conglomerates and for collabora- tion among the three different types of regulators.5 Finally, given that the settlement and payment systems are among the most important sources and channels of transmission of systemic risk in securities mar- kets, there is an ongoing debate on the desirability of reaching mini- mum standards for clearinghouses. Policy implications for emerging markets. The above discussion suggests several implications for policymakers responsible for emerging securi- ties markets. First, thiere seems to be a high payoff in information- sharing and enforcement agreements, both between emerging and industrial country markets, and among developing countries them- selves. Given the Latin America example, agreements between emerg- ing markets in the same region may be especially useful since they are likely to share financial intermediaries and investors and be exposed to common sources of external shocks. These collaborative arrangements should also include advance agreements on how to handle a crisis (Oyens 1996). Second, given the increased complication in assessing counterparty risk and the enhanced probability of contagion from other markets, governments and markets would do well to concen- trate on improving prudential rules and operatir.g procedures of the securities clearing agency, as discussed in the previous section. The Regulatory Framework Is Critical for Attracting Foreign Investors Three functions of the regulatory framework are critical for attracting foreign portfolio investment and reducing the potential costs associated with financial integration: ensuring accurate disclosure of all material information, eradicating insider trading, and improving corporate governance. Improving disclosure. Among the most critical functions of a regu- latory framework is to increase transparency in the market by mandat- ing public disclosure, both at the initial phase, when a firm issues 352 PREPARING CAPITAL MARKETS FOR FINANCIAL INTmx securities to the public, and continuously thereafter. As emphasized earlier in this chapter, such information is critical for foreign investors, indeed all investors, if decisions are to be more than just uninformed gambles. Hence, investors shun capital markets with weak disclosure. Inappropriate disclosure also undermines the more macroeconomic benefits of capital markets, in particular the efficiency of the market as a signaling device for the allocation of capital. Over time, three types of disclosure requirements have developed in capital markets: * Listing requirements-that is, the disclosure required for a firm to list in a securities market or trade publicly. Listing require- ments typically include minimum thresholds regarding the num- ber of shareholders and the value and volume of public shares, earnings, and balance sheet criteria over a number of years; an as- sessment of the potential of the firm and industry it belongs to; qualitative criteria regarding corporate governance; and credible documentation of the above criteria. Most markets also have con- tinuing listing or maintenance requirements. In addition, to avoid a regulatory race to the bottom among competing ex- changes in a country, there should be a baseline on disclosure im- posed by the official regulator for all securities traded publicly on these exchanges, or through other means.51 * Initial offering requirements-that is, the disclosure mandated for a firm to issue new securities. This involves two types of in- formation. First, information that allows investors to evaluate the overall condition of the firm issuing the securities, including risk factors and prior performance. Second, more specific informa- tion about the new issue: amount of capital to be raised and its in- tended purpose, dilution, how the offering price was determined, distribution plan, underwriters, and other market-making activ- ity. The information for initial offerings is usually distributed to the public in the form of a prospectus. * A general requiremfient to disclose all material information on a timely basis. Even though issuers may be complying with peri- odic reporting requirements, they are also obligated to disclose all important corporate developments as they occur-that is, devel- opments that may have an effect on the company's business or the stock price. These include mergers and acquisitions, stock divi- 353 J(-PITAL FLOWS TO DEVELOPING COUNTRIES dends, and changes in capital. In some cases, the exchange may suspend trading in the security when the event is announced to allow time for dissemination and analysis by i:he market. There may be additional disclosure requirements for companies in special industries, for example, mining, banking, and others. Also, in both industrial and developing countries, private of.ferings have less of a disclosure burden than public offerings. And finally, in industrial coun- tries, there is the concept of the "sophisticated" investor, usually a high- net-worth individual or institution, that is presumed to be well informed. Disdosure for securities sold to these investors is also less stringent. Differences in disclosure requirements among countries are not in- significant. They arise because of different legal fro meworks and regu- latory approaches, including the assignment of institutional responsibilities both within governments and between governments and the SROs. In industrial markets, while there is widespread agree- ment on the general categories or criteria that need to be disclosed, the specifics vary, as do the methods by which disclosure is made. In some countries, regulators rely on market practice and the general obligation to disclose all material information rather than on specific disclosure requirements. Australia, for example, has moved shiarply in this direc- tion in the 1990s. The United States, by contrast, has more specific re- quirements. But while this distinction is important in theory, in practice what is disclosed may be similar because of the generalized re- quirement to report all materlal events. With regard to how disclosure is made, some industrial countries rely more on the disclosure of mate- rial events and offering documents than on periodic reporting. Again, in practice the differences in what is actually disclosed may not be large. In emerging markets, since market practice may not be very widespread and the ability of regulators to monitor and enforce disclosure may be weaker than in more mature markets, it seems prudlent to be quite spe- cific on disclosure and reporting requirements. Harmonizing disclosure standards. The international community has been trying to harmonize disclosure requirements to facilitate cross- border trade. To this end, iosco prepared in 1991 a special report on the different disclosure requirements in member industrial countries to facilitate discussion of the issuance of securities in rnultiple jurisdictions with a single set of documentation. Then, in 1992, IOSCO published a survey of disclosure requirements in some 23 emeiging markets and in 354 PREPARING CAPITAL MARKETS FOR FINANCIAL IN7 Singapore and Taiwan (China). Finally, in 1993, iosco prepared guide- lines for disclosure in corporate offerings. In 1994, the Council of Securities Regulators of the Americas (CosRA) issued a framework for full and fair disclosure including when, what, and how disclosure should be provided and enforced and stated its intention to implement and maintain a system based on these principles. Both the COSRA framework and the lOSCO guidelines are good blueprints, but emerging markets need to adapt them to local conditions. The caveat is critical. For example, the best means and periodicity of disclosure and the treat- ment of small companies will depend on country circumstances. In addition, adopting such standards may require significant revisions in a country's legal framework, since in many countries disclosure require- ments are defined piecemeal in a variety of laws and decrees. Accounting standards. Disclosure will be effective only if the financial information provided by the company is based on sound accounting principles and practices that are well understood by investors. Common accounting standards are essential for investors to be able to evaluate the financial performance of a company on both an absolute and a relative basis. In parallel, auditing standards and practices also need improvement to ensure the reliability of disclosed information. Major cross-country differences in accounting and auditing stan- dards are an impediment to cross-border issues and to portfolio equity flows to the extent that financial statements are not transparent for in- ternational users.52 IOSCO and the International Accounting Standards Committee (IASC) are working together to harmonize accounting stan- dards. The approach adopted is to develop a revised set of international accounting standards toward which national standards could converge. To this end, the IASC hopes to develop a core set of international ac- counting standards by 1998. By 1995, 15 out of 31 norms had been developed. Upon successful completion of the project, IOSCO would recommend to its members that they accept these standards in cross- border filings. The IASC's most important challenge is to develop a comprehensive core set of standards that can portray accurately the very different circumstances that firms in different countries deal with and that measure consistently their performance. Another challenge is to ensure that these international standards are not adopted at the cost of excessive discretlon in comparison with national standards. Private standard-setting bodies have played an extremely important role in the more advanced emerging markets in developing sound ac- 355 CrAPTAL FLOWS TO DEVELOPING COUNTRIES counting and auditing principles and practices, and in offering contin- uing education for accountants and auditors. The L&SC's international standards have been used by these organizations in many cases as benchmarks. For example, in Malaysia, regulation of the accountancy profession is vested in the Malaysian Institute of Accountants (MIA). The MIA requires all limited companies to comply with local account- ing standards and helps ensure compliance by regularly reviewing sam- ples of published financial statements. If a statement falls short of the requirements, the member responsible for preparing or auditing the statements is asked to provide an explanation and take appropriate fol- low-up action. With regard to international standards, the MIA reviews on a regular basis the standards issued by the Lksc to determine their applicability in Malaysia. If the new standard does not conflict with local law or accounting practices, it will be issued under the heading of an international accounting standard. The MIA has, to date, adopted 24 of the 31 international standards without alteration. With regard to the others, Malaysia has its own alternative requirements, which are gener- ally consistent with international standards. Insider legislation. Another critical function of the regulatory frame- work necessary to develop confidence among investors is to eradicate insider trading-that is, the trading of securities while in possession of material nonpublic information obtained in breach of fiduciary duty or duty of confidentiality. Insider rules typically define what information is considered illegitimate-usually all facts that can have an impact on a company's business and the performance of its stock, who is subject to insider trading rules, the reporting requirements for insiders, and what types of companies are subject to these rules. In some industrial countries, owners of more than a certain amount of a certain stock are considered insiders and are required to file reports on their trading activities. Insider legislation also defirLes insiders' respon- sibilities: respecting strict confidentiality and not using the informa- tion for the benefit of self or others. Typically, civil and administrative penalties for insider offenses in- clude loss of profits, warnings, fines, and temporary suspension or can- cellation of registration. Criminal penalties can include jail terms. Judging from U.S. experience, Strahota (1996) suggests that emerging markets should not try to eliminate insider trading by relying solely on criminal prosecution. It is more effective to have the full range of sanc- tions (civil, administrative, and criminal), since criminal prosecutions 356 PREPARING CAPITAL MARKETS FOR FINANCIAL INT W require a burden of proof that is not easy to establish in the financial area, in particular in emerging markets, where the police and courts may not be experienced in these types of cases. To further facilitate en- forcement, insider rules should include incentives for market partici- pants to monitor compliance, such as making managers and firms responsible for breaches of insider rules committed by their subordi- nates or employees, requiring security houses to institute internal con- trol procedures, and holding responsible both the giver and receiver of information. These recommendations regarding the range of sanctions and incentives are important enforcement principles that can be ap- plied over a wide array of regulatory areas. Minority shareholder rights and corporate governance. Effective corporate governance involves essentially the promotion of shareholder rights and responsibilities, including those of minority shareholders. The two basic principles are fair treatment for all shareholders and share- holder approval of key corporate decisions. The latter is directed more at the potential conflicts between management and shareholders, while the former regards more minority shareholder rights. Fair treatment for shareholders implies that voting power, divi- dends, tender or exchange offers, and redemption should be propor- tional to the number of shares held, and that disclosure announce- ments should not discriminate among shareholders. Shareholder approval will in the first instance be indirect, through the election of an independent board of directors that approves the most significant corporate decisions made by senior management. But in addition, cer- tain outstanding decisions that can have a material effect on the finan- cial value of the corporation or its shares should be approved by the majority of the shareholders.53 Moreover, it is becoming common practice in industrial countries, especially in the United Kingdom and the United States, for management compensation to be determined by a committee of independent board members and for independent directors to establish an audit committee. Shareholders in industrial countries are also mandating changes in management compensation (for example, linking compensation to share price) and other incen- tives that encourage management to maximize shareholder value. Laws and regulations that promote these criteria make a capital mar- ket significantly more attractive for both foreign and domestic investors. To this end, some developing countries have begun to mandate by law the appointment of independent directors and other measures to im- 357 >ZIAMJTAL FLOWS TO DEVELOPING COUNTRIES prove corporate governance. Indonesia, for example, has recently intro- duced the requirement that issuers grant preemptive rights to existing shareholders in the case of new equity issues, as is common in Europe and Latin America.54 In Malaysia, one listing requirement is that there be a board audit committee composed of independent directors. In Korea, too, authorities are very much aware of the concerns of foreign institu- tional investors regarding shareholder rights and are considering ways of promoting corporate governance principles through capital market regu- lations (Hong 1996). However, it is not clear to whai: extent each of these principles should be regulated rather than enforced through shareholder pressure that leads to changes in corporate practices. The example of icici in India discussed in box 6.2 is a good example of how market pres- sures can improve governance without changes in regulations. More broadly, corporate governance also refers to shareholders and their representatives at the board being more independent and active in corporate affairs, monitoring and demanding more transparency from management. This shareholder activism and emphasis on corporate governance is increasing in many OECD countries, including Australia, Canada, France, the United Kingdom, and the United States, where, as noted earlier in this chapter, certain institutional investors have played a key role.55 One focus of these efforts has been co strengthen board membership criteria, in particular, to redefine what constitutes outside directors and to increase their role on the board. I n some companies, outside directors now evaluate formally the chief executive officer, ac- tively engage in succession planning, and more generally are instru- mental in preparing and implementing a long-term strategy for the company. The Role of SROs in Enforcement While laws and regulations based on international norms and best practice are one pillar of an effective regulatory framework, the other essential component is enforcement. The importance of enforcement cannot be stressed sufficiently. Without enforceme at, neither domestic nor foreign investors will have confidence in the capital markets, which will remain undeveloped. Effective enforcement requires clarifying the mandate and powers of the different regulatory institutions. With regard to the official regula- tor, the consensus is that one independent institution should be respon- 358 PREPARING CAPITAL MARKETS FOR FINANCIALNCJAL sible for the oversight of capital markets. The benefits of this regulatory structure are well known and will not be discussed further here, al- though there are some exceptions to this rule. For example, government debt markets are in many countries the responsibility of the central bank. And given the SRO model that is being adopted by many develop- ing countries, effective enforcement also requires defining the functions of sROs and their interrelationships with the official regulator. Self-regulation in industrial markets. The role of self-regulation in securi- ties markets varies across developing and industrial markets. Where this concept is most developed, in the United Kingdom and the United States, much of the responsibility for regulating the securities markets lies with SROs, which include the exchanges, the professional organizations, and the clearing and depository organizations. The SROs develop rules for their members and ensure compliance with these rules and with securities laws. Self-regulatory responsibilities can encompass regulation of market transactions (listing requirements, market surveillance, trading regulations, clearing and settlement, and disclosure of information), which is usually the responsibility of an exchange; regulation of market participants (licensing of broker- dealers and other professionals, capital and probity requirements, and business and ethical codes), typically done by professional organiza- tions; and dispute resolution and enforcement actions, including those that deal with insider trading.56 This regulatory model, based on disclosure and self-regulation, has been the result of a long evolutionary process in industrial countries. In the United States, for example, the self-regulatory organizations were the only regulators until the creation of the Securities and Exchange Commission after the crash of 1929. Self-regulation is based on the premise that it is in the long-term self interest of the industry to de- velop fair and efficient markets in order to attract capital and investors to the markets. The model presupposes competition among intermedi- aries and that different segments of the industry will have the incentive to monitor one another. Given the stylized attributes of capital markets in developing countries, however, the question is whether it is possible and desirable for these countries to have the same kind of market-based disclosure system for their capital markets. Applying the SRO model to developing countries. There are, in fact, several reasons why the SRO regulatory model can face problems in emerging markets: 359 "EIAPITAL FLOWS TO DEVELOPING COUNTRIES * Since the structure of the securities markets is imperfect because of limited competition, self-regulation may not be enough to en- sure fair and efficient markets. Korea, for example, has fixed commission rates, while membership in Indian exchange and clearance organizations is restricted. * The institutional and human capital may be insufficient to en- sure that two pillars of the system-self-regulation and disclo- sure-result in fair and efficient markets. For example, accounting standards may be deficient, the investor public may be uninformed, and reporting and disclosure practices may be weak or nonexistent. In addition, the accounting SROs may not have the ability to perform the oversight function effectively. * Finally, there may be short-term tensions betveen the regulatory and market development objectives of the authorities or the SROs. For example, care needs to be taken that new listings are not dis- couraged by excessive listing requirements; these perhaps should be tightened progressively or modulated accotding to the market on which the security will be traded. Alternatively, listing re- quirements may be too loose because the exchange may be seek- ing to increase business to the detriment of long-term confidence in capital markets. Despite these problems, however, the benefits of a self-regulatory structure to an emerging capital market are substantial, and such a struc- ture seems the best alternative. As discussed above, because of its first- hand experience, the industry is more able than government to formulate good rules and procedures, and to keep them current with new technology and industry practices. For the same reasons, the indus- try will generally also be better able to monitor compliance. In addition, self-imposed rules are usually better accepted than niles mandated from the outside. Finally, excessive government oversight or regulation may create different but just as serious governance issues through increased rent-seeking behavior. In any case, self-interest is cen tral to all regulatory systems, not only to the self-regulatory model. Any well-designed sys- tem has to be based on creating the right incentives for market partici- pants. If the incentives are not right, it is probably just a matter of time before the rules are brolken, even with intense and effective monitoring. The evolving role for SROs in emerging markets. Given these concerns, some forethought on the sequencing of responsibilities is necessary. 360 PREPARING CAPITAL MARKETS FOR FINANCIAL INTE Three criteria seem important for deciding which regulatory functions should be transferred to the SROs and at what speed: * Most obviously, the transfer of responsibilities should be modu- lated according to the institutional capacity of the SRO. In Indone- sia, for example, in the absence of strong professional associations, the official regulator, BAPEPAM, continues to license both legal and accounting professionals to work in the securities area. * The governance structure of the SROs should also be considered in allocating regulatory responsibilities, and it may be necessary to review the SRO's governance arrangements to minimize potential conflicts of interest. These issues are not unique to developing countries. For example, in both the United Kingdom and the United States there has been increasing concern about the ability of professional associations to monitor compliance by members with professional and ethical codes of conduct. In addition, gov- ernance issues may change over time, and as markets develop, new conflicts of interest may arise. Box 6.8 describes the reforms implemented in 1996 in both these countries to reduce these conflicts of interest while preserving the SRO model. * The comparative advantage of the industry relative to the official regulator in performing different regulatory functions should also be taken into account. Even if the SRO is not fully developed, it still might do a better job in certain areas than the official regulator. According to these criteria, functions regarding regulation of mar- ket transactions-market surveillance, the provision of information, and trading regulations for the securities market-should be trans- ferred first to the SRO (the exchange). Not only does the exchange have a large comparative advantage over the official regulator, but potential conflicts of interest seem less serious. For example, members of the ex- change have an incentive to ensure compliance with rules regarding the release of price-sensitive information, and to control and punish manipulative practices. Members of an exchange would also wish to have an efficient trading system. Potential conflicts of interest would seem more intense in licensing members, ensuring compliance with codes of conduct and listing requirements. Many regulators, however, believe that the prosecution of insider trading should be the responsi- bility of the official regulator because the SROs may not be willing to 361 ICAYITAL FLOWS TO DEVELOPING COUNTRIES Box 6.B Effective Governance of SROs in the United Kingdom and the United States IN RESPONSE TO INCREASING CONCERNS ABOUT sion oversees NASD, and reviews and approves NASD potential conflicts of interest in two key SROs, the rules and procedures. National Association of Securities Dealers (NASD) in NASD's three-prong mission required it to medi- the United States and the Institute of Chartered ate among conflicting interests. For example, NASD Accountants (ICA) in the United Kingdom, indepen- was the primary regulator of NASDAQ, in particular dent commissions rnade far-reaching proposals to of the trading and reporting practices of NASDAQ change the structure and governance of both SROs in market makers, who were also members of NASD. It 1996. These proposed reforms illustrate how con- was a membership association but was responsible flicts car be resolved while preserving the benefits of for handling public complaints against its members. a self-regulating system. NASD attempted to meet the needs of these different NASD is the only registered securities association constituencies by including representation of issuers in the United States, and every broker or dealer con- and investors in association affairs. However, its de- ducting a public securities business is required to be- centralized administration, heavy reliance on volun- come a mnember. Hence, it oversees the activities of teer member leadership, and the fact that one more than 5,400 securities firms, more than 57,000 professional staff was responsible for all three mis- member branch offices, and nearly 500,000 profes- sions made fulfillment of its responsibilities increas- sionals. As an sRo, it has certain defined obligations ingly difficult. In 1995, an in(lependent select regarding the oversight and discipline of its mem- committee-the Rudman Committee-was asked bers. NASD is, as well, the primary regulator, and to assess NASD's corporate governartce structure. The owner, ef the NASDAQ stock market and is mandated select committee found that structure not adequate to adopt rules that promote a fair, efficient, and safe to fulfill its responsibilities. market. NASD has therefore three missions. It is first In 1996, NASD undertook a fundamental restruc- a membership association. It is also entrusted with turing following the cornmittee's recommendations. ensuring responsible professional conduct of brokers NASD was restructured so as to put substantial "day- and deal ers. Finally, it is also the overseer of the NAS- light" between the membership association, the NAS- DAQ ma ket. The Securities annd Exchange Commis- DAQ market, and the broker-deale r regulator, with prosecute a member, especially if competition in the broker-dealer in- dustry is not strong. Another way of approaching the sequencing of the transfer of re- sponsibilities to SROs is to distinguish between types of SROs rather than regulatory functions. Certain SROs-the exchanges, clearinghouses, and depositories-can be considered "natural," that is, those whose members have strong and immediate incentives to develop and enforce rules of behavior in order to minimize transaction costs and risks. In- deed, in many industrial and developing countries these SROs were de- veloped naturally by self-interested market participants before securities commissions were established. In the case of other SROs- 362 PREPARING CAPITAL MARKETS FOR FINANCIAL INT- 4 three separate governing bodies whose compositions dence in regulation and to establish confidence are tailored to the particular requirements of their among members that regulation would be sensible. respective missions. Regulation of the broker-dealer The working party proposed the creation of a profession was separated from NASD into NASD-Reg- Public Oversight Board (POB) and the separation of ulator (NAsDR), and NAsDAQ was given an indepen- the Office for Professional Standards (oFps) from dent governing board. The professional staff the institute's other functions. Its proposal parallels members were also separated, even though all gov- the NASD initiatives designed to separate functional erning boards and staff remain associated within a responsibilities and to introduce transparency into single SRO structure. The restructuring also placed the system via public participation. The PoB would significant emphasis on disciplinary proceedings and consist of a few independent nonaccountants, not internal review. Finally, the reforms also emphasize chosen by the institute, charged with reviewing the increased public representation on the NASD's gov- effectiveness of regulation, reviewing public expecta- erning bodies (50 percent public members), which tions, assessing the extent to which expectations are not only will provide usefl public feedback to NASD met, and reporting in public to ensure redress of but should also bolster confidence in the NASD's shortcomuings and encourage public confidence. All policies. the institute functions related to regulation would In the United Kingdom, there was concern be grouped within the OFPS, which would be dis- whether the ICA was adequately managing the innate tanced from the institute's other activities, in partic- tensions arising from acting as the advocate for the ular its representative function. The disciplinary interests of members and as a fiduciary administer- aspect of members in business and the setting of eth- ing a code of conduct. After considering complaints ical standards will not be incorporated within the by mnembers and others about the effectiveness and OFPS. These areas would remain the responsibility of reliability of the institute as a regulator, the icA es- committees and tribunals consisting largely of tablished the Regulation Review Working Party. members. The report of the working party states that it was guided by two objectives: to maintain public confi- Source. NASD (1995), iCA (1996). broker-dealer and other professional organizations-self-interest is not as strong a motivating force for complying with rules. The benefits of complying with rules may become fully apparent only in the long term, and may accrue not only to the institution complying with the rule but mainly to the profession or market as a whole (for example, business and ethical codes of conduct). The implication is that regulatory re- sponsibilities can be transferred first to natural SROs, since they would be better able and more willing to develop and monitor rules of con- duct. The two approaches to looking at the transfer of responsibilities to SROs are not contradictory, and imply similar sequencing priorities. The very fact that natural sROs were able to develop and enforce rules 363 CtAPITAL FLOWS TO DEVELOPING COUNTRIES of conduct suggests that conflicts of interests are not intense between their members. What regulators need to watch out for are conflicts of interest between the members of a natural SRO and the market as a whole, in particular anticompetitive behavior. Improvements in the Regulatory Framework in Emerging Markets There are three key findings regarding the state of regulatory frame- works in emerging markets. First, as discussed above, many emerging markets are restricting foreign ownership of domestic stocks in their markets. However, the prevailing trend is a gradual softening of these restrictions, even in Asia, where they are most prevalent. Second, the regulatory frameworks of the more dynamic emerging markets have improved significantly in recent years, as have accounting standards. However, progress in this area is difficult to measure, since a judgment would need to take into account the completeness and soundness of the legal and regulatory frameworks, and effectiveness of enforcement. It is also difficult to generalize across countries, with remaining concerns varying very much by region. Third, emerging markets are converging to a disclosure-based, self-regulated framework. The self-regulatory model has been adapted, however, to the particular circumstances of each country, and the resulting systems range from those that give pri- mary and extensive powers to the official entity, to those that rely heav- ily on SROs, subject to oversight by the official regulator. As described in box 6.9, generally the influence of the state remains stronger in Asia than in Latin America. The gradual softening of investnent restrictions. As shown in table 6.5, many emerging equity markets are not fully open, with some economies imposing restrictions on foreign investment or the associ- ated foreign exchange movements. These restrictions are much more prevalent in Asia than in other parts of the developing world, in par- ticular Latin America. Except for Malaysia ancd Pakistan, all other major Asian emerging markets have some form of control. Korea and Taiwan (China), which require registration procedures and impose very tight foreign investment ceilings, seem the most closed. Notwith- standing the continued existence of controls, as discussed in chapter 1, the prevailing trend is one of gradual opening, with the number of countries that can be categorized as fully open increasing sharply since 1991. In Asia, for e ample, Korea and Taiwan (China) relaxed their 364 PREPARING CAPITAL MARKETS FOR FINANCIAL INTET Box 6.9 The Role of the State in Capital Markets in Asia and Latin America THERE IS A GROWING CONSENSUS IN DEVELOP- practices and the lack of protection of minor- ing countries on the proper role of the state in devel- ity shareholder rights. oping and regulating capital markets. At a general * More concentrated and financially weaker level, the consensus is that the state plays a key role capital market intermediaries and less devel- in creating an enabling environment for the growth oped human capital among the SROs. of private sector activity, including activity in capital a "Reputational" risk: these markets are all de- markets. A supportive environment includes pru- veloping and are more susceptible to a situa- dent economic management, and sound and trans- tion in which a problem involving an issuer or parent legal and regulatory frameworks that protect intermediary spills over and affects the reputa- property rights and enforce contracts. With regard tion of the market more generally. to capital markets, there is also consensus that the role of the state should change from one of direct in- As a result, Asian SROs are generally less indepen- tervention in development and regulation to one of dent than their industrial country counterparts and support and oversight. As noted before, the plurality have fewer regulatory responsibilities. of emerging markets have adopted the disclosure In contrast, although their markets share many of and self-regulatory model, and are at various stages the characteristics of Asian markets, Latin American in the process of gradually increasing the role of the securities regulators frequently have weaker powers SROs. There are, however, some important differ- than in industrial markets in critical areas such as ences between Asia and Latin America. certification of broker-dealers and exchanges and In Asia, despite the self-regulatory model, the disciplining violators of trading rules. In many role of the state in capital markets seems stronger countries, exchanges and other SROs predate a secu- than in industrial countries. In the minds of the au- rities commission by many decades; derive their in- thorities, certain characteristics of Asian capital mar- fluence from special, older statutes than those kets justify this stronger role for the state, at least for establishing the commissions; and have significant the moment. The most important are: political and financial influence. As a result, non- competitive practices, such as obstacles to wider * The generally more paternalistic view regard- membership in exchanges, inadequate capital ade- ing the role of the state in economic and social quacy provisions for broker-dealers, and undesirable development, coupled with weak investor trading practices are more common in Latin Amer- education. ica than in industrial markets. In addition, vested in- * The Asian tradition of closely held family terests have delayed the adoption of more modern businesses, which results in weak disclosure market infrastructures. investment ceilings in 1996, and Korea has announced a program to abolish the ceilings by the year 2000. Where do developing countries stand? As noted above, emerging markets have made progress in developing their regulatory frameworks. All the major emerging markets now have in place the basic building blocks of a framework; for instance, they have enacted securities laws (Russia most recently) and established independent securities commissions. 365 N.$WAPiTAL FLOWS TO DEVELOPING COUNTRIES Table 6.5 Inifestment Restrictions in Emerging Equiy Markets, 1995 Restrictionson foreign investment Restriction on foreign exchange movements b Investment ceilings Repaftiation of Repatriation of Market Freedom of entry a (percent) incmen prinitpal Argentina Free None Free Free Brazil Free 49 for common Free Free stocks, none for preferred stocks Chile Some restrictions None Free After one year China Only special classes None for B and H Free Free of shares shares India Only authorized 24 inT general Free Free (institutional) investors Indonesia Some restrictions 49 in general Some restrictioas Some restrncions Korea Some restrictions 15 in general Free Free Malaysia Free None Free Free Mexico Free None Free Free Pakistan Free None Free Free Philippines Only through B shares 40 in general and Free Free 30 for banks through B shares Poland Free None Free Free Sri Lanka Some restrictions 49 for banks Some restrictions Some restrictions Taiwan (Chiina) Only authorized 15 in general Somne restrictions Some restrictions investors Thailand Some restrictions 10-49 Free Free Venezuela Some restrictions None Some restrictions Some restnctions a. "Some restrictions" implhes that some registration procedures are required to ensure repatriation rights. b. "Some restrictionis" implies that registration or authorization of foreign exchange control authorities is required Source iFC, Emerging Stock Markets Factbook 1996. The latter is a relatively recent phenomenon in many Asian countries, where independent commissions replaced the ministry of finance and the central bank as the primary regulator of securities markets only in the early 1990s. Somewhat surprisingly, the accounting standards of many emerging markets have also strengthened considerably. Industry sources say that financial-reporting practices (a key component of disclosure) in several emerging markets are based on internationally recognized standards and are as comprehensive as in the United Kingdom and the United States.57 Studies conducted for the IASC also show that accounting standards in developing countries have been improving over time and are now on the whole either based on or consistent with international standards.58 These results do not contra- 366 PREPARING CAPITAL MARKETS FOR FINANCIAL 1NT dict the more negative World Bank assessments mentioned earlier, since the Bank assessments refer to actual accounting practices. Accounting practices are more influenced by the lack of a sufficient number of qualified accountants and auditors, a common problem in many emerging markets, than are standards. Finally, most major emerging markets also have regulations defining disclosure standards and listing requirements. Where the gap between developing countries and their industrial counterparts is still wide is in the more detailed but still critical aspects of a sound regulatory framework. For instance, although the data are not fully reliable, about half of the emerging markets included in table 6.6 have not established the legal and regulatory basis for dealing with compensation funds, takeovers, and insider trading.59 More to the point, many emerging markets have not yet instituted the legal and reg- ulatory basis for dealing with domestic institutional investors, as de- scribed in box 6.7. In addition to these common sources of fragility, there are other, rather general concerns that are shared by the regulatory frameworks of countries within the same region. As discussed in box 6.9, in Asian countries the common danger is overregulation, which may stifle market development and repel foreign investors. In Latin America, the danger seems to be underregulation or the lack of effective enforcement. In Eastern Europe and the Commonwealth of Indepen- dent States, the main task is to establish the basic legal and regulatory framework for capital market development. Despite these difficulties, it is clear that the regulatory authorities and SROs in many emerging markets are undertaking significant initiatives to be able to fulfill their responsibilities and improve enforcement. These initiatives are particularly striking in Asia. For instance, the Philippine Stock Exchange has set up new surveillance and audit departments, and has instituted a new computerized trading system to provide the raw material for surveillance. The Stock Exchange of Thailand has set up similar departments and instruments for market surveillance. At the same time, the Securities and Exchange Act of 1992 granted the Securi- ties Commission of Thailand wide powers to pursue and prosecute se- curities offenses. These include the power to subpoena witnesses and documentary evidence, and to inspect premises and records of securities intermediaries, including bank accounts. Since the commission is not able to prosecute, it has established a special legal unit to work jointly with the police and the courts to prosecute securities wrongdoings. It is 367 YAICAPITAL FLOWS TO DEVELOPING COUNTRIES Table 6.6 Legal and Regulatory Initiatives in Emerging Markets Insider Securities Established Disclosure Listing trading Compensation Takeover Country laws SEC regulation requirements reguation fund regulation Agentina Yes Yes Yes Yes Yes Yes Yes Brazil Yes Yes Yes Yes Yes Yes Chile Yes -Yes Yes Yes Yes Yes Yes China Yes Yes indiaa Yes Yes Yes Yes Yes Indonesia Yes Yes Yes Yes Yes Korea Yes Yes Yes Yes Yes Yes Yes Malaysia Yes Yes Yes Yes Yes Yes Yes Mexico -Yes Yes Yes Yes Yes Yes Yes Pakistan Yes Yes Yes Philippines Yes Yes Yes Yes Yes Poland Yes Yes Yes Yes Yes Yes Russia Yes Yes Sri Lanka Yes Yes Yes Yes Yes Yes Thailand Yes Yes Yes Yes Yes Yes Yes Turkey Yes Yes Yes Yes Note: The table has been constructed fromn secondary sources that have not been validated by an independent sear-h. In addition, emerg- snggmarkets are undertaking new initiarves on an ongoing basis. A blank means either that the country does not have regulation in that areaor sthat Itwas not possible to establish from the secondary sources whether a country had such regulation a. In India, there are listing requirements but they are solely informational. Source: LatinFinance, various issues; International Bar Association; ISSA. also seeking to increase investor confidence and reduce systemic risk by instituting capital adequacy rules based on risk factors that take into ac- count the specific types of risks faced by security firms. Summary and Conclusions I N THE NEW INTERNATIONAL ENVIRONMENT, AN INCREASING proportion of private capital flows will be channeled through capi- tal markets. This represents an important opportunity for devel- oping countries to tap into large overseas pools of capital. To attract this capital, developing countries need to rapidly develop their capital markets. There is a growing consensus on the plnciples that should guide such reform, and rapidly evolving country experience on which policymakers can draw. Addressing the priority agenda of foreign investors will have large spillover benefits for domestic markets and will reduce volatility and other market risks-two of the main con- 368 PREPARING CAPITAL MARKETS FOR FINANCIAL INT r-a" cerns that developing countries have about increased financial integra- tion. Indeed, this policy agenda to maximize the net benefits from financial integration largely overlaps with that considered essential to develop capital markets in a more closed economy setting. Foreign investors have already made an important contribution to capital market development in the more rapidly integrating developing countries. From no foreign presence in the mid- to late 1980s, the share of foreign investors in trading and market capitalization has risen sharply, and is now quite significant in most major emerging markets. Based on this impetus, developing country capital markets, especially equity markets, have made large strides: average market capitalization of 13 key emerging markets rose from 7 percent of GDP in 1985 to 43 percent in 1994, and trading activity roughly doubled during the same period. Foreign participation may also have important spillover effects on emerging markets through improved accounting and disclosure practices, corporate governance, and human capital. Despite these benefits, however, developing countries are concerned that financial integration may increase volatility and the risk of bubbles. Although it is true that asset prices in emerging markets are more volatile than in developed markets, there is, in fact, little evidence that volatility increases during capital inflow episodes. But we do find evi- dence of a negative cross-country correlation between excess volatility and the level of market development. Domestic capital market reforms that reduce information asymmetries and thereby promote liquidity can help reduce excess volatility, vulnerability to reversals, and inefficiencies. Developing countries show considerable variation in the capital mar- ket attributes needed for financial integration. The most aggressive have readily responded to increased interest by the foreign investor commu- nity by pursuing rapid and wide-ranging reforms. Others are making needed reforms more slowly, while most are still in a preemerging stage. Table 6.7 is an index of the overall level of development of the major emerging equity markets. It is based on the three essential aspects of mar- ket development that have been the main focus of investor concerns: market infrastructure, institutional development, and market structure and liquidity. The infrastructure subindex measures the efficiency of the market in settlement and postsettlement actions; the institutional devel- opment subindex is based on the quality of financial reporting, the pro- tection of investor rights, and the openness of the market to foreign investment; and the market structure subindex is a weighted average of 369 CfAiTAL FLOWS TO DEVELOPING COUNTRIES desirable market characteristics such as depth, lower volatility, and level of activity, relative to an industrial country benchrnark. Annex 6.3 de- scribes in more detail how the index was constructed. The most dynamic emerging equity markets, where progress has been particularly intense during the last five years, include most of high-growth Asia (Korea, Malaysia, and Thailand, with Indonesia and the Philippines not far behind), and two markets in Latin America (Chile and Mexico, with Brazil also ranking well). The East Asian mar- kets stand out for their depth and liquidity, and because of efforts un- dertaken in the 1990s their market infrastructures are now equal to those in Latin America. The lagging emerging markets in the sample are in South Asia (India, Pakistan, and Sri Lanka) and China. Gener- ally, these countries need to continue to improve their market infra- structure, as well as their institutional development. But even in the most advanced markets, the outstanding agenda is large. In the infra- structure area, for example, about 20 percent of all securities trades in Malaysia and Thailand do not settle on the contractual settlement date-four times mote than in the United States. The potential gains for developing countries fiom improving their capital markets are also illustrated by the significantly wider bid-ask spreads in emerging versus industrial country equity markets. Bid-ask spreads measure liquidity and reflect a wide gamut of infrastructure and regulatory factors that affect transaction costs and r sks, and promote in- vestor interest in a market. Based on 1996 data, figure 6.4 illustrates that even in the more dynamic emerging markets such as Brazil, Indonesia, Mexico, the Philippines, and Thailand, average spieads range from 130 to 170 basis points, compared with 67 basis points in the United King- dom. The differences remain just as striking for the most actively traded stocks in these markets for which the average spreads ranged from 70 to 125 basis points compared with 28 points in the UJnited States. Hence, through policy reforms and institutional improvements, emerging mar- kets could reduce the cost of equity in their economies by some 100 basis points or more by replicating the performance of industrial coun- tries, spurring investment and growth.6i To close the gap, emerging markets should pursue the following pol- icy agenda: * Infrastructure: Emerging markets should implement well- synchronized comparison, clearance and settlement, and central 370 PREPARING CAPITAL MARKETS FOR FINANCIAL IN Table 6.7 Capital Market Development in Emerging Markets, 1995 Subindex Market Market Institutional Overafl Country structrea infrastructureb developmentc index Argentina 4.0 8.7 8.2 6.2 Brazil 5.6 9.1 7.5 6.9 Chile 8.4 10.0 6.6 8.3 China 4.1 7.6 3.9 5.0 India 6.1 3.8 5.2 5.3 Indonesia 5.9 8.1 7.5 6.9 Korea 6.7 8.7 7.7 7.5 Malaysia 8.7 8.6 9.0 8.7 Mexico 6.4 8.4 8.8 7.5 Pakistan 4.3 1.0 7.8 4.1 Philippines 8.5 6.3 6.2 7.4 Poland 4.7 7.5 7.3 6.1 Sri Lanka 2.9 7.0 8.2 5.3 Thailand 8.6 8.7 7.1 8.3 Turkey 4.8 9.3 S.O 6.7 Note. The index ranges from 1 to 10 with higher numbers representing a higher ievel of market development. See annex 6.3. a. Based on a weighted average of market characteristics, induding market capializa- non, volatility, market concentration, and level of activity, relative to an industriaL coun- try benchmark. b. Based on measures of efficiency in setdement and postsetdement actions. c. Based on measures of the quality of financial reporting, protection of investor rights, and market openness. Seoure. iFc, Emerging Stock Market Factkook 1996$ Global Securitides Consulting Setvice, Review ofEmergingMarkets and Remew oafMajorMarkets, depository systems, with the goal of meeting G-30/ISSA guide- lines. However, achieving the G-30 recommendation of a T+3 settlement cycle should not be at the expense of reliability. Emerging markets should also pay close attention to reducing systemic risk by developing sound links with the banking and payment systems. In addition, the risk control procedures of a central clearing agency, if such an agency is required, should meet BIS guidelines, and a central depository should be established early in the integration process. The immobilization and demate- rialization of securities should also be encouraged, but not man- dated, since such systems require a lengthy lead time because of investor habits and legal impediments. * Property rights. The legal and regulatory framework should include two basic principles of shareholder governance: fair treatment for all 371 CAPITAL FLOWS TO DEVELOPING COUNTRIES Figure 6.4 Bid-Ask Spreads in Emerging and Industrial Country Equity Markets, 1996 A. Market Average United Kingdomo K,. a Korea Spain South Africa United States Malaysia Indonesia Mexico8 Philippines Thailand Brazil 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Spread (percentage) B. Average for the Most Actively Trailed Stocks United States Spain South Africa Korea United Kingdom BrazilI Malaysia Thailand Indonesia There is great potential for improved Mexico efficiency in secondary markets in Philippines developing cointries. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Spread (percentage) Note Values are based on the closing bid-ask spreads for the first 15 trading days in December 1996, taken from the Bloomberg stock quotation system The market average is based on a sam- ple of 20 stocks from the IFC Investable Index for each country, except for Spain, the United Kingdom, and the United States, for which the stocks were selected from the MADX, FT-SE 100 and the S&P 500, respectively The stocks were selected to try to replicate the distribution of market capitalization across the respective index. The average for the rnost actively traded stocks is based on the five most traded shares Source IFC, Emerging Stock Markets Factbook 1996 372 PREPARING CAPITAL MARKETS FOR FINANCIAL INTJG4 shareholders and shareholder approval of key corporate decisions. Some of these principles can be mandated in company bylaws or promoted through regulation, and will also spill over to emerging market firms through the increasing presence of foreign investors on their boards of directors. Also, in transition economies, it 1s essential to establish an independent registry so that records cannot be ma- nipulated by management or shareholders. * The regulatory framework. Emerging markets should adapt inter- national best practice on disclosure (including accounting) and self-regulation to local conditions, and improve enforcement of these rules. Government regulatory functions (starting with over- sight of trading activities) should be devolved to SROs as quickly as practicable, taking into account potential conflicts of interest and the SROs' capabilities. Better infrastructure, protection of property rights, and a well- conceived and -enforced regulatory framework are essential to promote investor interest and liquidity, and reduce systemic risks and volatility. They are the three legs of an efficient and safe capital market, each complementing the beneficial impact of the others. Indeed, progress in only one of these areas would be difficult to achieve. For example, risk reduction measures for central clearing agencies will be effective only if they are backed up by legal and regulatory practices that allow coun- terparty risk to be transferred to third parties and the market exposure of broker-dealers to be effectively monitored. Similarly, the eradication of insider trading and other illegal trading practices will be greatly fa- cilitated by market microstructures that simplify audit trails. While the construction of equity markets has been emphasized in many developing countries, there is increasing interest in the develop- ment of fixed income markets. On the domestic side, policymakers are turning to domestic bond markets as a means of mobilizing domestic savings to fund large investments, in particular in the infrastructure sector, with a more balanced mix of debt and equity. International in- vestors are becoming more interested in emerging market domestic debt as macroeconomic and financial sector conditions continue to im- prove in developing countries. As explained in box 6. 10, bond markets have specific constraints that need to be addressed, although many ele- ments of market development are common to both fixed income and equity instruments. 373 LCAIHTAL FLOWS TO DEVELOPING COUNTRIES Box 6.10 Developing Local Bond Markets INTERNATIONAL INVESTORS HAVE BEEN LESS IN- tic debt to minimize interest rate and exchange rate terested in developing country fixed income instru- volatility. ments than in local equities. While net portfolio Now, however, there is growing; domestic inter- debt flovws to developing countries have increased est in developing local debt markets as well as for- from $2 billion in 1990 to $46 billion in 1996, most eign interest in domestic currency dlebt instruments of these lesources have been raised through interna- from developing countries. In developing countries, tional bond issues. In addition, investments in local the main driving force is the need to fund large in- debt securities have mainly been on the short end of vestrnents, especially in the infrastructure sector, the market, in money market instruments. with a more balanced financing striucture. Interna- Foreign investors have preferred developing tionally, the number of developing country closed- country equities to bonds from these nations for two and open-end mutual funds dedicated to emerging main reasons. First, debt markets in most develop- market debt has increased, and in parallel, the in- ing countries are generally smaller and less liquid vestment firm J. P. Morgan and other organizations, than equity markets because of lack of supply and such as the IFC, have or are establishing performance demand. On the supply side, borrowers may favor benchmarks for this new emerging market asset equity over debt so that they can share the downside class. As emerging markets consolidate macroeco- risk in the high-return, high-risk environment of an nomic and financial sector policies and institutions, emerging market. On the demand side, investors are interest in bonds and money markets from both in- concerned with several issues, including the higher vestors and issuers will continue to rise. sensitivity of bond returns to macroeconomic insta- Much of the policy and institutional agenda for biliry, and a generally weaker debt market infra- equity markets analyzed in this chapter is shared by structure. In addition, in many emerging markets, debt markets, since many elements of market devel- domestic institutional investors, such as pension opment, for example, regarding market infrastruc- funds and insurance companies, whose counterparts ture, domestic institutional investors, and the in industrial countries invest heavily in debt securi- regulatory framework, are the same. The experience ties, are weak and sometimes restricted to pur- of the Bond Dealers Club (BDC) in Thailand illus- chasing only public sector securities. Foreign trates the benefits that well-functioning market infra- institutional investors prefer securities that allow structure can bring to the bond m.rket. BDC dealer them to capture the large up-side potential in members are strong financial institutions licensed to emerging markets and not just share the many risks trade bonds, and they set up BONDNET, a transparent common to equity and debt. Second, some govern- and dedicated electronic screen trading network, in ments hare restricted foreign transactions in domes- 1994. BDC has also facilitated market development Finally, emerging markets should also promote ihe development of domestic institutional investors that, by mobilizing significant re- sources, can serve as a counterweight to foreign investors and thereby assuage fears of excessive foreign presence. Institutional investors can ensure that a large pool of dedicated money will be available for bottom 374 PREPARING CAPITAL MARKETS FOR FINANCIAL INTEGRATION>-' with net clearing and settlement on a scripless and years. In Malaysia, however, the market benchmark DVP basis. After the first two years of its operation, is securitized mortgage obligations issued by the registered debt securities in BDC increased from CAGAMAS, the major public mortgage bank. This 29 issues with an outstanding volume of $1.3 billion suggests that the benchmark security should have to 116 issues with an outstanding volume of $5.7 bil- stable and well-known risk properties but need not lion, an increase of 330 percent. The trading volume be risk-free government paper. In Thailand, the through BONDNET has increased from a monthly av- market may in time find its own benchmark among erage of $37 million in 1994 to $676 million in the high-grade corporate paper being traded in the 1996, an annual increase of 300 percent. Bond Dealers Club. However, bond markets have specific develop- Various forms of securities are rated by outside ment constraints that are not shared with equity agencies, but such ratings are more important for markets; among the most important is the need for a bonds than for equities. Bond ratings help investors es- benchmark to price securities with different risk timate the risk premiums to compensate for varying profiles regarding default. In industrial country mar- probabilities of default on debt payments; hence rating kets, risk-free government paper has traditionally agencies reduce information asymmetries between is- served as the benchmark, but in many developing suers and investors, thereby promoting efficiency and countries, the price and yield of domestic treasury liquidity in primary and secondary debt markets. Rat- and other public debt securities are controlled for ings are also beneficial to issuers, especially those of less fiscal and debt management reasons and are not risky securities, since investors without access to rat- market based. Establishing a price benchmark based ings may demand higher risk premiums than other- on public instruments would hence require deregu- wise or not differentiate between securities. These lating public debt markets, as well as fiscal consoli- benefits are particularly important in emerging mar- dation to make the reforms sustainable. Ironically, kets, where risk assessment skills are probably less in other countries, in particular across much of Asia, abundant and advanced than in industrial country the rapid decline in outstanding public debt as a re- markets. Emerging markets are aware of these advan- sult of strong fiscal positions has led to similar diffi- tages, and many of them (such as Argentina, Chile, culties in establishing adequate price benchmarks. India, Indonesia, Korea, Malaysia, Mexico, Pakistan, The approach taken by the Hong Kong authorities the Philippines, Sri Lanka, and Thailand) have ratings is to issue government securities, regardless of bud- agencies, most of them recently established. getary needs, on a regular basis. By gradually length- ening the maturity of these issues, authorities have Source: Chow (1996), Hoonsiri (1996), J. P Morgan been able to establish a yield curve stretching for 10 (1996), World Bank data. fishing and value picking, reducing the vulnerability of domestic capi- tal markets to a rapid liquidation of assets by foreign investors. In ad- dition, domestic institutional investors will increase the depth and liquidity of domestic capital markets, enabling the markets to absorb the benefits that integration can produce. 375 fWE (APITAL FLOWS TO DEVELOPING COUNTRIES Annex 6.1 Table 6.8 G-30 Recommendations and Suggested ISSA Revisions G-30 recommendations ISSA -revisions of the G-30 recom nendations By 1990, all comparisons of trades between direct market All comparisons of trades between direct market participants participants (that is, brokers, broker-dealers, and other (that is, brokers, broker-dealers and other exchange members) exchange members) should be accomplished by T- I (the day should be accomplished by T+ 0. Matched trade details after the trade date). should be linked to the settlement system. By 1992, indirect market participants (such as institutional Indirect market participants (such as institutional investors investors or any trading counterparties that are not broker- and other indirect trading counterparties) should achieve pos- dealers) should be members of a trade comparison system itive affirmation of trade details on T+ 1. that achieves positive affirmation of trade details. Each count -y should have an effective and fully developed Each country should have in place an effective and fully central securities depository (CSD), organized and managed to developed central securities depository, organized and man- encourage the broadest possible industry participation aged to encourage the broadest possible direct and indirect (directly and indirectly), in place by 1992. industry participation. The range of depository-eligible instruments should be as wide as possible. Immobilization or dematerialization of financial instruments should be achieved to the utmost extent possible. If several CSDS exist in the same market, they should operate under compatible rules and practices vtith the aim of reduc- ing settlement risk and enabling efficient use of funds and available cross-collateral. Each count -y should study its market volumes and participa- Each market is encouraged to reduce settlement risk by tion to determine whether a trade netting system would be introducing either real-time gross settlement or a trade beneficial im terms of reducing risk and promoting efficiency. netting system that fully meets the Lainfalussy If a netting system would be appropriate, it should be imple- Recommendations of the 3IS. mented by 1992 376 PREPARING CAPITAL MARKETS FOR FINANCIAL INT G-30 recommendations ISSA revisions of the G-30 recommendations Delivery versus payment (Dvp) should be employed as the Delivery versus payment (DvP) should be employed as the method of settling all securities transactions. A Dvp system method of settling all securities transactions. Dvp is defined as should be in place by 1992. follows. simultaneous, final, irrevocable, and immediately available exchange of securitses and cash on a continuous basis throughout the day. Payments associated with the settlement of securities transac- No change. tions and the servicing of securities portfolios should be made consistent across all instruments and markets by adopting the .same day" funds convention. A "rolling settlement" system should be adopted by all mar- A rolling settlement system should be adopted by all markets. kets. Final settlement should occur on T+ 3 by 1992. As an Final settlement for all trades should occur no later than by interim target, final settlement should occur on T+ 5 by T+ 3. 1990 at the latest, except where it hinders the achievement of T+ 3 by 1992. Securities lending and borrowing should be encouraged as a Securities lending and borrowing should be encouraged as a method of expediting the settlement of securities transac- method of expediting the settlement of securities transac- tions. Existing regulatory and taxation barriers that inhibit tons. Existing regulatory and taxation barriers that inhibit the practice of lending securities should be removed by 1990. the practice of lending and borrowing securities should be removed. Each country should adopt the standard for securities mes- Each country should adopt the standard for securities mes- sages developed by the International Organization of sages developed by the International Organization of Standardization (iso Standard 7775). In particular, countries Standardization (iso Standard 7775). In particular, countries should adopt the ISIN numbering system for securities issues should adopt the ISIN numbering system for securities issues as defined in the iSO Standard 6166, at least for ctross-border as defined in the iso Standard 6166. transactions. These standards should be universally applied by 1992. Source. rssA (I1995). 377 $ (APJTAL FLOWS TO DEVELOPING COUNTRIES Annex 6.2 Market Microstructures JN RECOGNITION OF THE IMPORTANT ROLE the objectives of the market authorities. Since there is that they can play in a securities market, there no one optimal trading system, but rather each sys- has recently been a good deal of interest and tem works best in attaining one or two of the differ- activity in trading systems, or market microstruc- ent functions described above, market authorities tures.6i For example, recently, the London ex- may face difficult dilemmas. For example, adopting a change instituted an order-driven trading system dealer system for the smaller-volume stocks may in- for its larger-volume stocks. Many European and crease liquidity, but only at the cost of higher bid- emerging equity markets have adopted computer- offer spreads, particularly because in many emerging ized trading systems, in contrast to the New York markets there may be little competition among deal- Stock Exchange (NYSE), which has used computers ers. To help clarify the options that can be used to im- to improve and complement rather than replace prove market microstructures, this annex analyzes the open outcry floor trading.62 Emerging markets are properties of alternative microsiructures. However, also exploring alternative trading systems to in- this is a very wide area, and it will be possible only to crease the liquidity of the smaller-volume stocks touch upon the more important tradeoffs and issues. quoted in their markets.63 Microstructures are also being modified to reduce volatility. For instance, The Typology of Trading Systems after the 1987 crash, many markets instituted an automatic halt to trading when equity prices Markets in industrial and developing countnes use a change by ru ore than a predetermined amount. wide array of methods to match Suyers and sellers. A Characteristics of a good trading system. A good trad- first distinction is between auction-call markets and ing system should encourage the positive attributes continuous markets. In a call market, orders are of markets, those that enhance returns or decrease batched together for simultaneous execution at a pre- risk. First, the system adopted should reduce trans- specified time at a single price. The price is determined action costs. Second, microstructures should ease through an algorithm that maxirnizes the number of the price discovery process, facilitating the rapid trades that can be executed, with o time priority rule or and accurate incorporation of information into the a pro rata system to determine vrhich orders are exe- price of securities. When prices do not correctly cuted first.64 In a continuous market, orders are reflect information, market risks are higher, again processed continuously, interacring with changing discouraging investor interest. Third, microstruc- prices. In turn, continuous markets can be subdivided tures should contribute to reducing volatility in into order- and quote-driven markets. Quote-driven asset prices. As discussed in the main text, excess markets are dealer markets, in which a finite number of volatility confounds the information content of competitive dealers quote prices at which they are will- market prices, increasing market risk. Finally, ing to buy or sell the securities they wish to trade. In an microstructures should enhance liquidity. order-driven, or auction, market, orders are consoli- While these functions are not controversial, the dated in an auction type of environment. Some auc- net benefits of adopting a particular trading system tion markets also have market makers, or specialists, as will depend on the circumstances of the market and they are called on the NYSE, who are charged by the ex- 378 PREPARING CAPITAL MARKETS FOR FINANClAL change with "maintaining a fair and orderly market" in is between liquidity and immediacy. Call markets the stock for which they are responsible. Finally, all of have characteristics that make them especially these markets may be computerized. Figure 6.5 illus- appropriate for markets or stocks with little liquidity trates the typology of markets. or with information asymmetries. The periodic auc- tions of call markets collect orders and therefore The Basic Tradeoffs facilitate liquidity. By creating these pools of liquidity, a call market stabilizes the prices of infre- Call markets versus continuous markets. The basic trade- quently traded issues. At the same time, since all off between a call market and a continuous market transactions are executed at the same price, investors Figure 6.5 Major Types of Trading Systems Coninou mar Cal markt Electronic Open outcry Combination Electronic Open outcryElectrnic No Auction wit Auction withou roi Nolctoi ~i kz|market makers make ma F_~~ I - II Examples: * Frankfurt * NYSE * Korea * China * Brazil * Bombay * London * Tel Aviv * U.S. Treasury * Sri Lanka * Indonesia * Mexico * SOES * NASDAQ * New York * Tel Aviv * Korea Open * Malaysia * Poland * NSE, India * Taiwan * NYSE [China) Super Dot * Toronto * Paris Open * Philippines * Thailand * Tokyo * Turkey Abbrevzations NASDAQ National Association of Securities Dealers Automated Quotation, NSE National Stock Exchange. NYI; New York Stock Exchange; SOES Small Order Execution Service of the National Association of Securities Dealers 379 IkA-TAL FLOWS TO DEVELOPING COUNTRIES with less information will not be at a disadvantage. between higher liquidity in quote-driven dealer mar- Another positive characteristic of a call market, if kets versus lower transaction costs and higher trans- appropriately structured, is that it efficiently incor- parency in order-driven markets. The greater liquid- porates available information into the price. In this ity in dealer markets is a consequence of having sense, it is o good system for determining opening dealers "making a market" in the security, in contrast prices in a market or for determining the price of a to one specialist, or none at all, in order-driven mar- stock whose trading has been temporarily halted kets. By standing ready to buy and sell at their quot- because of the announcement of new information. ed prices and committing their own capital, dealers Many markets use call auctions for these purposes. add liquidity and immediacy to a market. In addi- Finally, call markets generally result in lower execu- tion, dealers' profits increase with order flow, and tion and settlement costs, mainly because transac- they have an incentive to promote the securities they tions execute at the same price and in one batch. On are trading, creating more liquidity. However, deal- the negative side, in contrast to continuous markets, ers provide these liquidity services at a cost, in par- call markets provide no immediacy, that is, the abili- ticular the expense of holding an inventory of the ty of buyers and sellers to transact promptly. Other security, labor and technology costs, and losses problems are that call markets provide no price incurred in trading. These costs will be reflected in informationl between sessions and may result in the dealer's bid-offer spread. In contrast, transaction incomplete order execution. costs in order-driven systems generally will be small- The lack of immediacy in a call market may be er, since traders can conduct transactions directly a significant deterrent to foreign portfolio man- with each other. In a quote-driven market, competi- agers, who quite actively change the composition tion among dealers is essential to ensure a high-qual- of their portfolios. For many of the larger-volume, ity market. Recently, in industrial countries, private more active stocks in emerging equity markets, a proprietary trading systems that facilitate direct call auction system would add little liquidity and interaction among traders are becoming an addi- sacrifice boi h immediacy and information. A com- tional source of competition for dealer markets. bination of the two systems may be attractive to In general, with regard to both pretrade and some equity markets, with a call market for the post-trade information, dealer markets provide less smaller-volume, less liquid stocks and a continu- information and are less transparent than order-dri- ous market for the others. There is no technical ven markets. In a quote-driven market with more impediment to having two systems; indeed, as than one dealer, the information on order flow will noted above, many equity markets use a call auc- be segmented. This contrasts with order-driven sys- tion to determine the opening price. Another good tems, which consolidate efficiently all information example is the Tel Aviv stock market, which has an when determining prices. In many order-driven electronic call market for the less capitalized stocks markets, the order book is available for all to see, and a semic ontinuous order-driven market for the while in quote-driven markets, only the dealers have most liquid stocks.65 a good sense of the order flow and the price of exe- Order-driven markets versus quote-driven markets, and cution. This lack of transparency in quote-driven the role of imarket makers. The basic tradeoff is markets is compounded by the fact that many 380 PREPARING CAPITAL MARKETS FOR FINANCIAL INTEG1R-I trades, especially the larger ones, are not transacted Computerized markets. Figure 6.5 shows that many at the quoted price, but are negotiated between the of the most dynamic emerging equity markets have dealer and the trader. The effective transaction price computerized their trading systems. The term may not be reported in a timely manner. For exam- "computerized" can be misleading. While a dealer ple, a recent investigation by the Securities and Ex- market such as NASDAQ is computerized, what a change Commission of the United States revealed computerized market is usually understood to mean instances of less than competitive dealer quotations is an order-driven system where an algorithm and pricing as well of dealers declining to honor matches buyers and sellers automatically, replacing quotes in the NASDAQ stock market (U.S. Securities the traditional floor open outcry trading. This and Exchange Commission 1996). trend in emerging markets is based on the need to The choice among systems depends on particular increase the capacity of the market to handle larger circumstances and objectives, and financial integra- volumes, as well as to increase transparency. Elec- tion does not make choosing any easier. For exam- tronic markets can trade around the clock and can ple, different types of investors may prefer different handle large volumes in a cost-efficient manner. types of market systems. In particular, large-volume With regard to transparency, computers simplify traders, and most foreign traders will fall into this market monitoring by creating a detailed audit trail, category, are usually concerned about the price im- and have the capacity to generate and disseminate pact of their order. Other things remaining equal, huge amounts of information on market transac- they would prefer a system that creates the most liq- tions on a real-time basis. And very important, the uidity and immediacy, and through which they are automatic matching system helps ensure that better able to control the degree of disclosure about investors are treated fairly. Computer-based systems the prospective transaction. Retail investors, on the work best in markets with a large volume of retail other hand, are more concerned about getting the trading, simplifying the matching process and help- best price and would prefer a mechanism that gives ing small investors get the best price. their orders the widest exposure. Hence, a trading The main problem computerized systems have ts system that accommodates this concern of foreign that of processing large orders in a market with in- investors may be to the interest of some but not all sufficient liquidity. To address these concerns, many of the domestic investor community. equity markets have adopted a mixed system in On the other hand, for some foreign investors, which small orders are executed automatically the most important concern may be transparency. through the electronic system, while larger trades As noted above, foreign investors believe that they can be executed by a more traditional system oper- are at a disadvantage with respect to access to infor- ating in tandem. For example, NASDAQ in the mation and are concerned about trades with a bet- United States has an automated system that directs ter-informed domestic dealer. In this case, they may small orders automatically to the best quote, while prefer the more transparent order-driven systems to the larger trades can be negotiated over the phone a dealer market. The support of many foreign in- with a dealer. Argentina also has two systems, an vestors for computerized markets arises out of this open outcry continuous auction and an electronic concern for transparency. market for smaller trades. 381 -P-RIVATE CAPITAL FLOWS TO DEVELOPING COUNTRIES Annex 6.3 Construction of the Emerging Markets Index T HE EMERGING MARKETS INDEX (EMI) IS weights in both the subindexes and the main intended to serve as a rough estimate of the index, unless there is a strong reason to believe oveiall level of development of emerging that one variable dominates the others. equity markets from the investors' perspective. The iG The use of benchmarks, or reference points, in EMI is a weighted average of three subindexes measur- calculating the subindexes. Variables in the ing three essential aspects of market development: subindexes are evaluated on the basis of how far they deviate from a benchmark, typically * a market structure subindex that measures defined as the value of the variable in an in- depth, liquidity and volatility dustrial country market. * a market infrastructure subindex that measures the efficiency of the market in terms of settle- Despite these principles, the EMI suffers from ment and custody several problems. First, given the lack of a formal in- * an znatitutional development subindex that vestor survey, the weights-both those assigned to measures institutional aspects of market devel- the three subindexes in calculating the overall EMI opment-for example, the quality of financial and those allocated to the underlying variables used reporting, the protection of investors' rights, to estimate the subindexes-are somewhat arbi- and the openness of the market to foreign in- trary. To increase the transparency of the index, as vestment-that are important determinants noted above, we have used equal weights unless of investor interest in a market. there is a strong reason to believe other values should be used. The second problem regards the This annex describes the EMI's construction choice of benchmarks against which to evaluate methodology, including variable selection and emerging market development and how much to weights. penalize countries for deviations from the bench- Guiding principles and issues. In developing the EMI, mark. We have used the U.S. market as a bench- we were guided by several general criteria: mark for simplicity and because the use of other industrial markets (or an average) does not change = The use of well-defined, measurable, and ob- relative country rankings significantly. Basing our jective variables that capture the critical as- methods on investor preferences, we have tended to pects of market development. penalize larger deviations from the benchmark pro- * The use of information that is uniform across portionally more than smaller ones-that is, in- countries, including standardized information vestors are willing to adjust to small problems in a such as the Global Securities Consulting Ser- market but may become more sensitive to market vices, Inc., (GScs) benchmarks on clearance constraints when these constraints cross a certain and settlement. threshold. Another significant problem is the lack of * Weighting (ideally) based on investor con- systematic, standardized cross-country information cerns, as revealed through a survey. In the ab- regarding the quality and completeness of the regu- sence of such data, however, we use equal latory framework in emerging markets. Hence the 382 PREPARING CAPITAL MARKETS FOR FINANCIALIN'1 EMI does not incorporate this critical aspect of mar- Specifically for items 1 to 3, the rating function ket development. is: These problems suggest that EMI should be used 1 (b -x )2 1 only as a rough indicator of market development and i = [ b2 xi that small differences between countries are not sig- 10 nificant. The results are, however, reassuring, in the o sense that the EMI rankings shown in table 6.7 are where b is the value of the selected variable for the broadly consistent with foreign investors' preferences reference market and x is its value in the emerging among emerging markets as revealed to the authors. market j. We chose the quadratic function to accen- The overall weights. The structure subindex is given tuate the penalties for increasing divergence from a 50 percent weight in the EMI; the trading infra- the benchmark (the b2 term in the denominator is structure and institutional infrastructure subindex- for scaling purposes), as noted above. es are each weighted at 25 percent. The structure The coefficient of variation is also rated with re- of the market has been given a higher weight spect to its benchmark value in the United States because liquidity and depth are two characteristics according to the following schedule: highly desired by foreign investors. In addition, and perhaps more important, since liquidity and Deviation ofcoefficient of varation depth are endogenous variables, their presence from benchmark (percent) Rating indicates that the market in question is perceived 0 to 10 9 favorably by investors. 10 to 30 8 The structure subindex. The market structure is cal- 30 to 50 7 66 ~50 to 70 6 culated using the following ratios or variables:66 70 to 100 5 1. market capitalization/GDP (weighted at 50 100 to 150 4 percent) ~~~~~~~150 to 250 3 percent) 250 to 500 2 2. turnover ratios (weighted at 12.5 percent) 500 to 800 1 3. new issues/market capitalization (weighted at over 800 0 12.5 percent) 4. volatility as measured by the coefficient of This schedule gives the highest rating to markets for variation (weighted at 12.5 percent) 5vmarketi concentration (weighted at 12.5 pwhich the coefficient of variation differs up to 10 p. marker cncent)ation (weighted at 12.5 percent from the benchmark. If the deviation is from 10 to 70 percent, we apply a linear rating. As Each of the five variables is rated on a scale of 0 the deviation goes beyond 70 percent, ratings de- to 10. A cumulative score of 10 implies that the par- crease, but at a diminishing rate. This rating proce- ticular market is highly advanced. The rating may dure differs from the one used previously because be performed using either a rating function or a investors expect high levels of volatility in emerging schedule. Items 1 to 4 are rated according to their markets and because they probably do not distin- deviations from their benchmark values in the guish among very high levels of volatility relative to United States. the United States benchmark. 383 + T A~LAL FLOWS TO DEVELOPING COUNTRIES Market concentration is measured by the ratio volved in any given market. Because GSCS in assign- of two shares: the share of the 10 largest stocks in ing values already captures deviations from a theo- total market capitalization divided by the share that retical benchmark, using 100 as the highest score, these stocks represent in the total number of stocks. we do not adjust for a developed market. One sim- Concentration is then rated according to the fol- plification for our purposes is that markets which lowing schedule: have negative values for either of the benchmarks are assigned a value of zero instead. Since GSCS does Concentration measure Rating not estimate safekeeping benchmarks for Chile, 0 to 5 9 China, Pakistan, Poland, and Sri Lanka, the settle- 5 to 10 8 ment benchmark becomes in effect the infrastruc- 10 to 20 7 ture subindex. 20 to 30 6 Institutional infrastructure subindex. The institu- 30 to 40 5 tional infrastructure is determined by assigning 40 to 60 4 equal weight to the following criteria: 60 to 80 3 1. financial reporting standards, using the CIFAR 80 to 90 2 Index from the Center for International 90 to 100 1 Analysis and Research 2. foreign taxation, using the IFC's assessment This schedule rates markets linearly if the concen- 3. investor protection, using the IFC's assessment tration measure ranges from 5 to 80, while penaliz- 4. onnesstor foreini, using the J' .s ing the more extreme levels of concentration (80 to 100). We did not identify any benchmark for the assessment. market concentration figures. 'The CIFAR Index, with a maximum score of 100, Trading infrastructure subindex. The trading infra- measures the quality of the ireporting standards structure is determined by assigning equal weight against an ideal and implicitly allows direct com- to the settlement benchmark and the safekeeping parisons between those prevailing in different cap- benchmark established by GSCS. The settlement ital markets. Because of lack of information, benchmark measures settlement efficiency as indi- Indonesia and Poland are excluded from the assess- cated by th-e number and operational costs of ment of accounting standards, and that weight is failed and delayed transactions, administrative redistributed equally among the remaining complexity, and compliance with G-30 recom- elements. mendations. The safekeeping benchmark measures The foreign taxation rating is based on the IFC the efficiency of different markets in terms of the data on withholding taxes for emerging markets. collection of dividends and interest, reclamation We first calculate the average tax rate from the of withheld taxes, and protection of rights in the three forms of taxation-on interest income, divi- event of a corporate action. dends, and long-term capital gains on listed The benchmarks are expressed as a score ranging shares-and then rate the taxation policy of to 100; the lower the score, the higher the effective emerging markets according to the following operational costs and administrative efforts in- schedule: 384 PREPARING CAPITAL MARKETS FOR FINANCIAL IN Average tax rate (percent) Rating The investor protection rating is based on the IFC 0 10 classification, available as of the end of 1994. We 0 to 5 9 convert the IFC classification according to the follow- 5 to 10 8 ing terms: a good program scores a 9, an adequate 10 to 15 7 one receives a 6, and a poor scheme is rated at 2. 15 to 20 6 Openness to foreign access is rated using the IFC 20 to 25 5 summary of regulations for entering and exiting 25 to 30 4 emerging markets. The rating schedule is as follows: 30 to 100 2 Foreign access classification Rating This schedule is justified strictly from an investor's Free 10 point of view; the classes used in our rating schedule Relatively free 7.5 are relatively parsimonious under the assumption Restrictions to special classes 5 that the presence of any taxation policy acts as a Authorized investors 2.5 strong disincentive to foreign investment. Closed 0 Notes 1. As explained in chapter 2, portfolio flows are ex- domestic stock markets in developing countries. With re- pected to increase because emerging markets are still un- gard to emerging market debt, foreign investors have been derrepresented in the portfolios of industrial country mainly attracted to paper issued in international or indus- investors, and investments in emerging markets would re- trial country markets, rather than to domestic issues in de- sult in a significant improvement in the return-risk ratio veloping countries. However, over time, the development of these investors. In addition, it is expected that an in- of domestic bond markets will become increasingly im- creasing share of savings in industrial countries will be in- portant. The interest of foreign investors in domestic debt termediated by institutional investors and that these instruments will undoubtedly rise, and debt will play an investors will continue to increase the share of emerging increasing role in financing investment in emerging mar- market equities in their portfolios. kets, especially in the infrastructure sector. While many development issues discussed in this chapter with respect 2. Market capitalization is the value of the shares to market infrastructure and the regulatory framework are quoted in the market at market prices. It is a measure of common to equity and debt markets, others, such as do- the size of a market. mestic rating agencies and market price benchmarks, are more specific to bond markets. 3. The stock of financial savings can be defined as the sum of the stock of deposits in commercial banks and the 6. A key question that will arise with regard to the reg- capitalization of the stock market. ulatory and legal frameworks is the role of the state in fa- cilitating the transition process in capital markets as 4. As defined in chapter 2, excess volatility (in flows or financial integration deepens. asset prices) is volatility not due to changes in fundamentals. 7. The principal-agent problem to which we refer is the 5. Given the interest of foreign investors in emerging result of information asymmetries between management market equities, the chapter focuses on improvements in and the shareholders of a corporation. Shareholders, who 385 WITAL FLOWS TO DEVELOPING COUNTRIES may not have easy access to all material financial and oper- firms to the detriment of investment and research and de- ational inforrcation regarding the firm, are concerned velopment expenditures. The empirical work on this whether managers are operating to maximize shareholder question based on U.S. data suggests that institutional value. However, to acquire such information, shareholders ownership does not affect investment but may reduce re- will incur a cost. The higher the cost, the more potential search and development expenditures (see Samuel 1996). investors will be deterred from buying shares in the firm, leading to a suboptimal level of investment. Institutional 15. For example, Summers (1986) found evidence of factors and regulations also affect the problem; for exam- such inefficiencies because of volatility in industrial coun- ple, it may not be worthwhile for a small shareholder to try markets. incur the search cost since better monitoring will mostly benefit other shareholders. In addition, small shareholders 16. Levine and Zervos (1996a), which found that ex- may not be able to effectively organize themselves to con- ternal liberalization increased asset price volatility, is an test management. On the other hand, improvements in exception. corporate governance would help shareholders to monitor management. Blommestein and Spencer (1993) review 17. It is not uncommon in the literature to indude the this issue in the context of transition economies, where the legal framework, accounting standards, and rating institutions principal-ager[t problem is particularly severe. as part of market infrastructure. We iliscuss the legal system and accounting standards together with the regulatory frame- 8. There is strong empirncal evidence from the United work, to which tley are closely relatecd, in the next section of States that institutional investors are very active traders. this chapter. Rating agencies are revnewved in box 6.9. For example, see Samuel (1996). 18. To be more precise, the clearance and settlement of 9. Demirguc-Kunt and Maksimovic (1994) and Glen trades in organized markets involve relationships between and Pinto (1994) present additional evidence. Samuel brokers that are members of the exchange. These broker- (1996) found that Indian firms use external sources of fi- broker relationships are referred to the "street," or market, nance more than their counterparts in industrial countries side of a trade, while the investor-broker relationships are because of higher reliance on debt. referred to as the "customer" side. Thls review of market infrastructure mainly focuses on the street side, although 10. Their findings are based on pooled cross-section we will discuss some investor-broker issues in the section (41 countries) time-series (1976-93) regressions, control- dealing with custody functions and central depositories. ling for the more traditional macroeconomic and human capital variab]es that past studies have found to be signifi- 19. See Group of 30 (1989). cant in exploining growth. Atje and Jovanovic (1993) reach similar conclusions. 20. For example, see loSco (1992b). 1 1. There is wide disagreement in the literature re- 21. See ISSA (1995). garding the causes and consequences of takeovers. In any case, takeovems are not common in developing countries. 22. Glen (1994) reviews how .nicrostructures influ- ence these market attributes. 12. It may not be worthwhile for a small investor to incur these costs because the benefits of better monitoring 23. There are no significant differences between will accrue mainly to other shareholders. See note 5. matching systems for fixed-income and for equity instru- ments, except the match criteria. 13. See Samuel (1996) for a summary of the literature. 24. Other criteria could include counterparty and 14. A related issue is whether institutional ownership clearing broker, buy-sell instructions, settlement date, and puts too much short-term profit-maximizing pressure on so forth. 386 PREPARING CAPITAL MARKETS FOR FINANCIAL INTT 25. DVP is the most effective yet simple way of reducing ized, and regulated. If not, the better depositories would principal risk, whether involving physical or dematerial- insist that interdepository settlement take place only in ized securities, or for transactions that are settled through a "good securities"-that is, in the accounts of the registrar central C&S system or directly between buyer and seller. or a top-tier depository, potentially slowing the settlement process (Morgenstern 1996). 26. Stehm (1996) discusses these characteristics in some detail. 35. As noted above, compared with intermediaries in mature markets, intermediaries in emerging markets are 27. High trade failure rates, however, are not only the more likely to be financlally weak, and volatility is gener- result of inefficiencies in the C&S system but also, for in- ally higher, while financial integration may bring large stance, of undercapitalized or inefficient broker and dealers. surges in volume. In addition, the systemic consequences of the failure of an intermediary may be larger if the in- 28. Effective risk monitoring would also require the dustry is more concentrated, as some analysts believe it to central clearing agency to correctly value or revalue the be in emerging markets. collateral, including the credit standing of the issuer of the collateral, and any legal impediments on it. 36. Nominee ownership is not required when the se- curities-holding system is a direct system. For example, in 29. BIS (1990) and Stehm (1996), for example, suggest China, the central depositories of the two main markets allocating losses to the surviving participants prorated direcdy record the ownership of securities by beneficial with their level of activity with the defaulung participant. owners and also act as registrars. Since individual participants may have the same problems as the clearing agency in evaluating the financial strength 37. Laws important for the functioning of the financial of the counterparty, the proposed measure may reduce system include: a company law that adequately promotes market activity. corporate governance; banking and commercial laws that give firms a legal basis for practices (such as beneficial 30. See annex 6.1 and BIS (1990). ownership, notation, trusts, collateral, and so forth) that are central to financial markets; bankruptcy laws that 31. The terminology in this area can be confusing. clearly define the rights of different asset owners in a liq- Some experts refer to all book entry systems as demateri- uidation; and competition laws. An essential component alized and to systems where certificates are not issued as of the legal infrastructure is a set of well-functioning insti- uncertificated. tutions that enforce these laws. 32. Other important differences concern ownership 38. Strahota (1996) describes how this model could be and governance of the depository and whether or not applied in emerging markets. there is more than one depository. 39. The assets of an investment firm consist of mar- 33. Other advantages of indirect holding systems are ketable securities that in a crisis can be liquidated at close that investors need to deal with only one institution for all to their book value, which is not the case for bank assets. purposes and that they reduce broker-investor settlement Similarly, liabilities of investment firms are less liquid risk. On the other hand, direct systems facilitate the ac- than bank liabilities, especially deposits, which can be counting and control of the number of shares outstand- withdrawn quickly. Dale (1996) discusses this rationale ing, as well as the communication between the issuer and for regulation in some detail. investors. Morgenstern (1996) has an excellent review of the two systems. 40. There are other methods to address this concern. For example, the approach taken by regulatory authorities 34. If more than one exists, all the depositories should in the United States has been to raise funding fire walls not only be interlinked, but also well managed, capital- between banks and nonbanks to insulate the former from 387 pay,ITAL FLOWS TO DEVELOPING COUNTRIES solvencv or tIquidity problems in the latter. See Dale counterpart regulator that may help in an investigation or (1996) for a discussion of the regulatory response to func- an enforcement action, approval o.F certain investigative tional integration in the European Union, Japan, and the powers of the domestic regulator in rhe counterpart coun- United States. try, and the obligation to report to the other party that a firm is experiencing financial difficulties. 41. In an indirect holding system for securities, in- vestors' assets would be held by brokers and custodians. 50. See IMF (1996). 42. The collapse of Barings in 1995 illustrates the im- 51. Listing requirements, to the extent that they are portance of arrangements that permit clearinghouses to not solely informational, seem to contradict the basic identify separately proprietary positions taken by invest- premnise that investors should be responsible for their in- ment firms from those taken on behalf of their clients. vestment decisions. There are two reasons for such re- quirements. First, because of lack of expenence and 43. losco (1996) reviews and makes recommenda- financial expertise, investors may not be able to judge the tions about the main techniques that regulators can use to relative ments of alternative investments. Second, because achieve a satisfactory level of client asset protection. of the negative externalities resulting from investors' los- ing confidence in the market, caused in general by the bad 44. In sonie cases, for example, Chile, these controls performance of a listed firm. are imposed not to constrain foreign ownership but to in- fluence the amount and composition of capital inflows for 52. See IASC (1993) for a description of the key differ- macroeconoaic management reasons and to restrain ences in accounting standards. lending booms The advantages and costs of such restric- tions are discu ssed in chapters 4 and 5. 53. Among the decisions often submitted for share- holder approval are acquisition or transfer of ownership of, 45. TIhe regulatory implications of the failure of Bar- say, 20 percent of the shares of the company, limitation of ings are discussed in Dale (1996) and IMF (1995). shareholder rights, sale of corporate assets, or the incurring of significant new debt liabilities. See Seeger (1996). 46. See Dale (1996), Goldstein (1996), and IMF (1995 and 1996) for more details. The International Securities 54. Preemptive rights are the entitlement of existing Regulation Report is an even more detailed source. shareholders to subscribe to new share issues for cash. For issuers, it may raise the cost of capital if the sale is made at 47. The se zurities regulators of 73 countries, or about a substantial discount from market price, as is common 95 percent of countries with stock exchanges, are mem- practice to maintain attractiveness during the subscrip- bers of losco. Interestingly, iosco also has affiliate and tion period. Existing shareholders aie either able to main- associate mernbers, in particular stock exchanges and tam their ownership interests if they subscribe or are other self-regulatory organizations, as well as some inter- compensated for the resulting dilution if they choose to national organizations. IMF (1996) describes losco's in- sell their subscription rights. The pros and cons of this ternal committee structure. practice are still being debated. 48. Previously, losco had issued recommendations re- 55. Some key documents of this trend are the General garding internal control and auditing arrangements for fi- Motors (1995) and Cadbury (Committee on the Finan- nancial institutions dealing in derivatives (IMF 1996). cial Aspects of Corporate Governance 1992) reports. For a comparison of the corporate governance initiatives in 49. These memorandums usually include all or a sub- the OECD, see Millstein and Gregor7 (1996). set of the following: routine sharing of general informa- tion, sharing of certain information on firms operating in 56. See International Capital Markets Group (1992) the two markets, access to official information held by the for a good summary review of self-regulatory activity. 388 PREPARING CAPITAL MARKETS FOR FINANCIAtI INT 57. According to the International Financial Reporting 61. Microstructures refers to the way securities are Index constructed by the Center for International Analysis traded and the influence that trading systems have on key and Research (an investment adviser located in the United market attributes. States), Chile, Malaysia, Mexico, Pakistan, and Sri Lanka have reporting standards comparable to those in industrial 62. For example, see Freund (1996). countries. Argentina, Korea, and Thailand are also ranked relatively highly, and only Brazil, India, the Philippines, 63. For example, in Thailand, the stock exchange is and Turkey lag behind the industrial countries. The index considenng having dealers try to make a market for the is based on the reporting practices of major domestic cor- less active stocks porates with regard to 85 disclosure variables. 64. In turn, call markets can be distinguished by the 58. IAsc (1996). type of auction or algorithm that is used to determine prices-for example, whether or not bids are public. See 59. Compensation funds protect investors from losses Aggarwal (1996). arising from the failure of broker-dealers (not market risks), while takeover rules protect the rights of minority 65. Another example of mixed systems very prevalent shareholders in a target company for a takeover. in emerging markets regards government debt instru- ments. Primary markets for these instruments are usually 60. There are two reasons why the bid-ask spreads for call markets, while secondary markets are usually over- emerging markets in figure 6.4 are probably biased down- the-counter dealer markets ward. First, for the markets included in the figure, the av- erage spreads are based on a sample of stocks included in 66. Practically all data are from the IFC, data for new the IFC Investable Index and hence are among the most equity issues are from FIBV (the International Federation liquid. Second, it is likely that bid-ask spreads in many of of Stock Exchanges) sources and supplemented by IFC the emerging markets for which data were not available data. Market capitalization is at year-end. The coefficient will be wider than those of the countries included in the of variation is that of the monthly percentage change in figure. This is because many of the omitted countries are, the IFC Global Total Return Indexes. The benchmark is according to the index on capital market development, the corresponding coefficient of variation for the S&P somewhat behind in developing their markets. 500. 389 I Bibliography The word "processed" describes informally reproduced works that may not be commonly available through libraries. Abbreviations: BIS, Bank for International Settlements; IADB, Inter-American Development Bank; IMF, International Monetary Fund; ISSA, International Society of Securities Administrators; MIT, Massachu- setts Institute of Technology; NBER, National Bureau of Economic Research; OECD, Organisation for Eco- nomic Co-operation and Development. Adler, Michael, and Bernard Dumas. 1983. "Inter- Macroeconomic Effects of Financial Sector Re- national Portfolio Choice and Corporation Fi- forms: An Overview of Industrial Countries." nance: A Synthesis." Journal of Finance 38(Uune): BIS, Basle, Switzerland. 925-84. Arrow, Kenneth. 1964. "The Roles of Securities in Aggarwal, Reena. 1996. "An Evaluation of Market the Optimal Allocation of Risk Bearing." Review Microstructures." Background paper prepared of-Economic Studies 31 (April): 91-96. for Private Capital Flows to Developing Countries. Atje, Raymond, and Boyan Jovanovic. 1993. "Stock World Bank, International Economics Depart- Markets and Development." Eturopean Economic ment, Washington, D.C. Review 37(April): 632-40. Aitken, Brian. 1996. "Have Institutional Investors Backus, David, Patrick Kehoe, and Finn Kydland. Destabilized Emerging Markets?" IMF Working 1992. "International Real Business Cycles." Jour- Paper 96/34. Washington, D.C. nal of Political Economy 100(4): 745-75. Aitken, Brian, and Ann Harrison. 1992. "Does Baer, Herbert, and Daniela Klingebiel. 1995. "Sys- Proximity to Foreign Firms Induce Technology temic Risk when Depositors Bear Losses: Five Spillovers?" MIT, Cambridge, Mass. Processed. Case Studies." In George G. Kaufman, ed., Aiyer, Sri-Ram. 1996. "Anatomy of Mexico's Bank- Banking, Financial Markets, and Systemic Risk. ing System during the Peso Crisis." Report 45. Greenwich, Conn.: JAi Press. World Bank, Latin America and the Caribbean Bagehot, Walter. 1880. Economic Studies. London: Technical Department, Regional Studies Pro- Longmans, Green. gram. Washington, D.C. .1924. Lombard Street. Reprint of 1st ed. Lon- Andersen, P. S., and William R. White. 1996. "The don: John Murray. 391 =jATB _APITAL FLOWS TO DEVELOPING COUNTRIES Bank of New York. 1996. "Current Profiles." New in America. Washington, D.C.: American Enter- York. prise Institute. Barro, Robert, and Xavier Sala-i-Martin. 1995. Eco- Berger, Allen, Richard J. Herring, and Giorgio P. nomic Growth. New York: McGraw-Hill. Szego. 1995. "The Role of Capital in Financial Basle Committee on Banking Supervision and [nstitutions." Journal of Banking and Finance 19 Iosco (International Organization of Securities (3-4, June): 393-430. Commissions). 1995a. Frameworkfor Supervisory Berger, Allen, William C. Hunter, and Stephen G. Information about the Derivatives Activities of Timme. 1993. "The Efficiency of Financial Insti- Banks and Securities Firms. Basle, Switzerland. tutions: A Review and Preview of Research, Past, -. 1995b. Public Disclosure of the Trading and Present, and Future." Journal ofBanking and Fi- Derivatives Activities of Banks and Securities nance 17(April): [22]-49. Firms. Basle, Switzerland. Bernanke, Ben S. 1983. "Nonmonetary Effects of Baxter, Marianne, Urban Jermann, and Robert the Financial Crisis in the Propagation of the King. 1995. "Nontraded Goods, Nontraded Fac- Great Depression." American Economic Review tors, andl International Non-Diversification." 73(3, June): 257-76. NBER Working Paper 5175. Cambridge, Mass. Bernanke, Ben S., and Alan S. Blinder. 1988. Bayoumi, lamim. 1990. "Saving-Investment Cor- "Credit, Money, and Aggregate Demand." Amer- relations: Immobile Capital, Government Policy, ican Economic Review 78(2, M\ay): 435-39. or Endogenous Behavior?" IMF Working Paper Bernanke, Ben S., and Mark Gertler. 1989. "Agency 90/66. W7ashington, D.C. Costs, Net Worth, and Business Fluctuations." - 1993. "Market Integration and Investment American EconomicReview 79: 14-31. Barriers in Emerging Equity Markets." Working Bernstein Research. The Future of Money Manage- Paper. Stanford Graduate School of Business, mentinAmerica. New York: Sanford C. Bernstein. Stanford, Calif. Bhattacharya, Amar, Peter Montiel, and Sunil Bekaert, Geert, and Campbell R. Harvey. 1995. Sharma. 1996. "Private Capital Flows to Sub-Sa- "Emerging Equity Market Volatility." NBER haran Africa: An Overview of Trends and Deter- Working Paper 5307. Cambridge, Mass. minants." World Bank, International Economic Bencivenga, Valerie, and Bruce Smith. 1991. "Finan- Department, Washington, l).C. Processed. cial Intermediation and Endogenous Growth." Beveridge, Stephen, and Charles R. Nelson. 1981. Reviewof(Economic Studies 58(April): 195-209. "A New Approach to Decomposition of Eco- Bencivenga, Valerie, Bruce Smith, and Ross Starr. nomic Time Series into Permanent and Transi- 1995. "Equity Markets, Transaction Costs, and tory Components with Particular Attention to Capital Accumulation" Policy Research Working Measurement of the Business Cycle." Journal of Paper 1456. World Bank, Policy Research De- Monetary Economics 7(2, March): 151-74. partment, Washington, D.C. BIS (Bank for International Settlements). 1990.In- Benston, George, and George Kaufman. 1988. terbank Netting Schemes of the Central Banks of "Regulating Bank Safety and Performance." In the Group of Ten Countries. Committee on Pay- William S. Haraf and Rose Marie Kushmeider, ment and Settlement Systerns Report 10. Basle, eds., Resiructuring Banking and Financial Services Switzerland. 392 BIBLIOG 0 1996. 66th Annual Report. Basle, Switzerland. Burchill, Andrew. 1996. "Is Russia Ready for Mu- Bisignano, Joseph. 1994. "The Internationalisation tual Funds?" Institutional Investor 21(March): of Financial Markets." Cahiers Economiques et 85-89. Monetaires (France) 43: 9-77. Calomiris, C. W. 1995. "Financial Fragility: Issues Blanchet, Marie-Christine, Peter Burridge, Stijn and Policy Implications." Journal of Financial Claessens, and Maxwell Fry. 1995. "Foreign Di- Services Research 9(3). rect Investment, Other Capital Flows, and Cur- Calomiris, C. W, and Gary Gorton. 1991. "The rent Account Deficits." Policy Research Working Origins of Banking Panics: Models, Facts, and Paper 1527. World Bank, Policy Research De- Bank Regulation." In R. Glenn Hubbard, ed., partment, Washington, D.C. Financial Markets and Financial Crises. Chicago: Blommestein, H. J., and Michael G. Spencer. 1993. University of Chicago Press. "The Role of Financial Institutions in the Transi- Calvo, Guillermo. 1990. "The Perils of Sterilization." tion to a Market Economy." IMF Working Paper IMF Working Paper 90/13. Washington, D.C. 93/75. Washington, D.C. -. 1996. "Capital Flows and Macroeconomic Blomstrom, Magnus, Robert E. Lipsey, and Mario Management: Tequila Lessons." International Carlos Zejan. 1992. "What Explains Developing Journal of Finance and Economics 1 (July): Country Growth?" NBER Working Paper 4132. 207-23. Cambridge, Mass. Calvo, Guillermo, Leonardo Leiderman, and Car- B6hm-Bawerk, Eugen von. 1891. Positive Theory of men M. Reinhart. 1993. "Capital Inflows and Capital. London: Macmillan. Real Exchange Rate Appreciation in Latin Amer- Borensztein, Eduardo, Jose de Gregorio, and Jong- ica: The Role of External Factors." IMF Staff Pa- Wha Lee. 1995. "How Does Foreign Direct In- pers40(1): 108-51. vestment Affect Economic Growth?" NBER Campbell, John Y., and Robert J. Shiller. 1987. Working Paper 5057. Cambridge, Mass. "Cointegration and Tests of Present Value Mod- Bbrsch-Supan, Axel. 1996. "The Impact of Popula- els." Journal of Political Economy 95(October): tion Aging on Savings, Investment and Growth 1062-88. in the OECD Area." In OECD, Future Global Capi- Caprio, Gerard, Jr. 1996. "Bank Regulation, The tal Shortages: Real Threat or Pure Fiction. Wash- Case of the Missing Model." Policy Research ington, D.C. Working Paper 1574. World Bank, Policy Re- Bovenberg, Lans, and Roger Gordon. 1994. "Why search Department, Washington, D.C. Is Capital So Immobile Internationally?" NBER Caprio, Gerard, Jr., and Daniela Klingebiel. 1997. Working Paper 4796. Cambridge, Mass. "Bank Insolvency: Bad Luck, Bad Policy, or Bad Brock, Philip L. 1996. "High Real Interest Rates and Banking?" In Michael Bruno and Boris Guarantor Risk in an Open Economy A Case Pleskovic, Annual World Bank Conference on De- Study of Chile, 1975-83." University of Washing- velopment Economics 1996 Washington, D.C. ton, Department of Economics, Seattle. Processed. . 1996. "Bank Insolvency: Cross-Country Ex- Buckberg, Elaine. 1996. "Institutional Investors perience." Policy Research Working Paper 1620. and Asset Pricing in Emerging Markets." IMF World Bank, Policy Research Department, Working Paper 96/2. Washington, D.C. Washington, D.C. 393 aCPAPITAL FLOWS TO DEVELOPING COUNTRIES Caprio, Geerard, Jr., and Dimitri Vittas. 1995. "Fi- Emerging Markets?" Policy Research Working nancial Hlistory: Lessons of the Past for Reform- Paper 1243. World Bank, Plolicy Research De- ers of thLe Present." Policy Research Working partment, Washington, D.C. Paper 1535. World Bank, Policy Research De- Chuhan, Punam, Stijn Claessens, and Nlandu partment. Washington, D.C. Mamingi. 1993. "Equity and Bond Flows to Caprio, Gerard, Jr., Izak Atiyas, and James Hanson, Latin America and Asla: The Role of Global and eds. 1994. Financial Reform, Theory andPractice. Country Factors." Policy Research Working Cambridge, U.K.: Cambridge University Press. Paper 1160. World Bank, International Econom- Caprio, Gerard, Jr., Michael Dooley, Danny jics Department, Washington, D.C. Leipziger, and Carl Walsh. 1996. "The Lender of Chuppe, Terry, and Michael Atkin. 1992. "Regula- Last Resort Function under a Currency Board, tion of Securities Markets: Some Recent Trends The Case of Argentina." Policy Research Work- and Their Implications for ]Emerging Markets." ing Paper 1648. World Bank, Policy Research Working Paper 829. IFC, ]Economics Depart- Department, Washington, D.C. ment, Washington, D.C. Cardenas, Mauricio, and Felipe Barreras. 1996. Citibank. 1996. "Country Profiles." New York. "On the Effectiveness of Capital Controls in Claessens, Stijn, and Sudarshan Gooptu, eds. 1993. Colombla." World Bank, International Finance Portfolio Investment in Developing Countries. Division. Washington, D.C. Processed. World Bank Discussion Paper 228. Washington, Celarler, Michelle. 1994. "How the Bankers at ]D.C. Banco Latino Stole the Bank." Global Finance Claessens, Stijn, Michael P Dooley, and Andrew 7(9, September): 130-35. Warner. 1995. "Portfolio Capital Flows: Hot or Center for International Financial Analysis and Re- Cold?" World Bank Economic Review 9(1): search. 1995. InternationalAccountingandAudit- 153-74. ing Trenas, 1995. Princeton, N.J. Cohen, Daniel. 1995. "Foreign Finance and Eco- Chevalier, Judith A., and Glenn D. Ellison. 1995. nomic Growth." In Leonaldo Leiderman and "Risk Taling by Mutual Funds as a Response to AssafRazin, eds., CapitalMobility: The Impact on Incentives." NBER Working Paper 5234. Cam- Consumption, Investment, and Growth. Cam- bridge, Mass. bridge, U.K.: Cambridge University Press. Chinn, Menizie, and Michael Dooley. 1995. "Asia- Cole, Harold, and Maurice Obstfeld. 1991. "Com- Pacific Capital Markets: Integration and Implica- modity Trade and International Risk Sharing." tions for Economic Activity." NBER Working JTournal ofMonetary Economics 38: 3-24. Paper 5280. Cambridge, Mass. Committee on the Financial Aspects of Corporate Chow, Paul. 1996. "Debt in the Asia Pacific Re- Governance. 1992. The Cadbury Report. Lon- gion-Hong Kong's Role" (paper presented at don: Her Majesty's Stationery Office. the conference on Private Capital Flows: Impli- Corbo, Vittorio, and Leonardo Hernmndez. 1994. cations for Asian Capital Markets, Bali, Indone- "Macroeconomic Adjustmeni: to Capital Inflows: sia, October 18-20). Processed. ]Latin American Style versus East Asian Style." Pol- Chuhan, Punam. 1994. "Are Institutional Investors icy Research Working Paper 1377. World Bank, an Impoi tant Source of Portfolio Investment in P'olicy Research Department, Washington, D.C. 394 . 1996. "Macroeconomic Adjustment to Capi- in Chile: 1974-1983." In Philip L. Brock, ed., If tal Inflows: Lessons from Recent Latin American Texas Were Chile: A Primer on Banking Reform. and East Asian Experience." World Bank Research San Francisco: ICS Press. Ohserver 11(1, February): 61-85. de Larosi&e, Jacques. 1996. Per Jacobsson Lecture, . 1997. "Private Capital Inflows and the Role September 29, IMF, Washington, D.C. of Economic Fundamentals." Background paper De Ryck, Koen. 1996. "European Pension Funds: prepared for Private Capital Flows to Developing Their Impact on European Capital Markets and Countries. World Bank, International Economics Competitiveness." European Federation for Re- Department, Washington, D.C. tirement Provision, London. Council of Securities Regulators of the Americas. Dellas, Harris, Behzad Diba, and Peter Garber. 1994. "Framework for Full and Fair Disclosure 1996. "Resolving Failed Banks: The U.S. S&L in the Americas." Quebec. Experience." Discussion Paper Series 96-E-22. Dailami, Mansoor, and Michael Atkin. 1990. Bank of Japan, Institute for Monetary and Eco- "Stock Markets in Developing Countries: Key Is- nomic Studies, Tokyo. sues and a Research Agenda." Policy Research Demirguic-Kunt, Asli, and Ross Levine. 1996a. Working Paper 515. World Bank, Country Eco- "Stock Markets, Corporate Finance, and Eco- nomics Department, Washington, D.C. nomic Growth." World Bank Economic Review Dale, Richard. 1996. "Regulating the New Finan- 10(2, May): 223-39. cial Markets" (paper prepared for conference on . 1996b. "Stock Market Development and Fi- Globalization: What It Is and Its Implications, nancial Intermediaries: Stylized Facts." World Sao Paulo, Brazil, May). Bank Economic Review 10(2, May): 291-321. Davis, E. P. 1991. "International Diversification of Demirgu,-Kunt, Asli, and Vojislav Maksimovic. Institutional Investors." Bank of England Discus- 1994. "Capital Structures in Developing Coun- sion Papers 44(September): 1-53. tries: Evidence from Ten Countries." Policy Re- -. 1993. "The Structure, Regulation, and Per- search Working Paper 1320. World Bank, Policy formance of Pension Funds in Nine Industrial Research Department, Washington, D.C. Countries." Policy Research Working Report Denton, Nicholas. 1996. "Banks Plan Grip on 1229. World Bank, Financial Sector Develop- World Debt Market." Financial Times (July 10). ment Department, Washington, D.C. Dermine, Jean. 1993. "The Gains from European de Gregorio, Jose, and Pablo Guidotti. 1995. "Fi- Banking Integration: A Call for an Active Pro- nancial Development and Economic Growth." Compassion Policy." In Hans Genberg and World Development 23: 3-24. Alexander Swoboda, eds., World FinancialMarkets de Juan, Aristobulo. 1987. "From Good Bankers to after 1992. London: Kegan Paul International. Bad Bankers: Ineffective Supervision and Man- Devereux, Michael, and Gregor Smith. 1994. "In- agement Deterioration as Major Elements in ternational Risk Sharing and Economic Banking Crises." World Bank, Financial Policy Growth." International Economic Review 35 (Au- and Systems Division, Washington, D.C. gust): 535-50. de la Cuadra, Sergio, and Salvador Valdes. 1992. Devlin, Robert, Ricardo Ffrench-Davis, and "Myths and Facts about Financial Liberalization Stephany Griffith-Jones. 1994. "Surges in Capi- 395 -ATE CAPITAL FLOWS TO DEVELOPING COUNTRIES tal Flows and Development: An Overview of Pol- Far Eastern Economic Review. Hong Kong. icy Issues." In Ricardo Ffrench-Davis and Feis, Herbert. 1964. Europe, The World' Banker, Stephany Griffith-Jones, eds., Coping with Capi- .1870-1914 New York: Kelley. tal Surges: The Return of Finance to Latin Amer- Feldman, Robert A., and Manmohan S. Kumar. ica. Boulder, Colo.: Lynne Rienner. 1]995. "Emerging Equity Mairkets: Growth, Ben- Diamond, I)ouglas. 1984. "Financial Intermedia- efits, and Policy Concerns." 'World Bank Research tion and Delegated Monitoring." Review of Eco- Observer 10(2, August): 181--200. nomic Studies 5 1(3): 393-414. Feldstein, Martin, ed. 1991. Tbe Risk of Economic Diamond, Douglas, and Philip Dybvig. 1983. Crisis. Chicago: Universityof Chicago Press. "Bank Rtmns, Deposit Insurance, and Liquidity." Feldstein, Martin, and Charles Horioka. 1980. Review oJfEconomic Studies LI: 393-414. "Domestic Saving and International Capital Diaz-Alejandro, Carlos. 1985. "Good-Bye Financial Flows." EconomicJournal 90(une): 314-29. Repression, Hello Financial Crash. "Journal of Fernandez, Roque, and Liliana Schumacher. 1996. Development Economics 19: 1-24. "Does Argentina Provide a Case for Narrow Domowitz, Ian. 1990. "The Mechanics of Auto- Banking?" (paper presented ot the conference on mated Trade Execution Systems. "Journal of Fi- Preventing Banking Crises in Latin America, nancial Intermediation 1 (June): 167-94. World Bank, Washington, D.C., April 15-16). Dooley, Michael P. 1994. "Are Recent Capital In- Processed. flows to Developing Countries a Vote for or Fernandez-Arias, Eduardo. 1994. "The New Wave against Economic Policy Reforms?" Policy Re- of Private Capital Inflows: PuLsh or Pull?" Policy search Working Paper 295. University of Califor- Research Working Paper 1312. World Bank, In- nia, Santa Cruz. ternational Economics Department, Washing- - 1995. "A Survey of the Academic Literature ton, D.C. on Controls over International Capital Account --. 1996. "The New Wave oif Private Capital In- Transactions." NBER Worktng Paper 5352. Cam- flows: Push or Pull?" Journal of Development Eco- bridge, Mass. nomics 48 (March): 389- 418. - 1996. "Capital Controls and Emerging Mar- Fernindez-Arias, Eduardo, ancl Peter J. Montiel. kets." International Journal of Finance and Eco- 1]995. "The Surge in Capital Inflows to Develop- nomics I (July): 197-205. ing Countries: Prospects and Policy Response." Drees, Burkhard, and Ceyla Pazarbasioglu. 1995. Policy Research Working P'aper 1473. World "The Nordic Banking Crisis: Pitfalls in Financial Bank, International Economics Department, Liberalization?" IMF Working Paper 95/61. Washington, D.C. Washington, D.C. FIBV (Federation Internationale des Bourses de El-Erian, Mohammed, and Manmohan S. Kumar. Valeurs). 1996. "Clearing and Settlement: Best 1995. "E.merging Equity Markets in Middle Practices." Paris. Processed. Eastern Countries." IMF Staff Papers 42(June): Fieleki, Norman S. 1994. "International Capital 313-43. Transactions: Should They Be Restricted?" New Fama, Eugene. 199 1. "Efficient Capital Markets: II." England Economic Review (March-April): 27-39. Journal ofFinance46(December): 1575-617. . 1996. "International Capital Movements: 396 BIBLIO 6a 1 How Shocking Are They?" New England Eco- Freund, William. 1996. "Will New Technology nomic Review (March-April): 41-60. Leave the Big Board behind the Times?" Durell Fischer, Bernhard, and Helmut Reisen. 1994. "Fi- Journal ofMoney and Banking (Winter). nancial Opening: Why, How, When." Occa- Fry, Catherine. 1993. "Custodial Issues (Clearing sional Paper 58. International Center for and Settlement) of Equities in Emerging Mar- Economic Growth. San Francisco, Calif. kets." In Keith Park and Antoine Van Agtmael, Fleming, Alex, and Samuel Talley. 1996. "The Lat- eds., The World's Emerging Stock Markets. vian Banking Crisis: Lessons Learned." Policy Chicago: Irwin. Research Working Paper 1590. World Bank, En- Garber, Peter. 1996. "Managing Risks to Financial terprise and Financial Sector Division, Washing- Markets from Volatile Capital Flows: The Role ton, D.C. of Prudential Regulation." Internationaljounal of Fleming, Alexander, Lily Chu, and Marie-Renee Finance and Economics (July): 183-95. Bakker. 1996. "The Baltics-Banking Crises Gardener, E. P. M. 1991. "1992 and the Future of Observed." Policy Research Working Paper Asset Securitisation in Western Europe." In J. J. 1647. World Bank, Enterprise and Financial Sec- Norton and P. R. Spellman, eds., Asset Securitisa- tor Division. Washington, D.C. tion: International Financial and Legal Perspec- Folkerts-Landau, David, and others. 1995. "Effects tives. Oxford: Basil Blackwell. of Capital Flows on the Domestic Financial Sec- Gavin, Michael, and Ricardo Hausmann. 1996a. tors in APEC Developing Countries." In Mohsin "Make or Buy: Approaches to Financial Market Khan and Carmen Reinhart, eds., Capital Flows Integration" (paper presented at the conference in the APEC Region. Occasional Paper 122. on Safe and Sound Financial Systems: What Washington, D.C.: IMF. Works for Latin America? IADB, September). Frankel, Jeffrey. 1993. On Exchange Rates. Cam- Processed. bridge, Mass.: MIT Press. . 1996b. "The Roots of Banking Crises: The Frankel, Jeffrey, and Chudozie Okongwu. 1996. Macroeconomic Context." In Ricardo Haus- "Liberalized Portfolio Capital Inflows in Emerg- mann and Liliana Rojas-Suarez, eds., Banking ing Markets: Sterilization, Expectations, and the Crisis in Latin America. Baltimore: Johns Hop- Incompleteness of Interest Rate Convergence." kins University Press. International Journal of Finance and Economics Gavin, Michael, Ricardo Hausmann, Roberto Perotti, 7(March): 1-23. and Ernesto Talvi. 1996. "Managing Fiscal Policy Frankel, Jeffrey, and Andrew K. Rose. 1996. "Cur- in Latin America and the Caribbean: Volatility, rency Crashes in Emerging Markets: Empirical Procyclality, and Limited Creditworthiness." IADB Indicators." NBER Working Paper 5437. Cam- Working Paper 326. Washington, D.C. bridge, Mass. General Motors. 1995. "GM Board of Directors Frankel, Jeffrey, and Sergio L. Schmukler. 1996. "Cri- Corporate Governance Guidelines on Significant sis, Contagion, and Country Funds: Effects on East Corporate Governance Issues." Detroit, Mich. Asia and Latin America" (paper delivered at a con- Glen, Jack. 1994. An Introduction to the Microstruc- ference at the Center for Pacific Basin Monetary ture of Emerging Markets. IFC Discussion Paper and Economic Studies, San Francisco). Processed. 24. Washington, D.C.: World Bank. 397 :+AflAL FLOWS TO DEVELOPING COUNTRIES Glen, Jack, and Brian Pinto. 1994. Debt or Equity? search Working Paper 1650. World Bank, Policy How Firms in Developing Countries Choose. IFC Research Department, Washington, D.C. Discussion Paper 22. Washington, D.C.: World Gray, Philip. 1996. "Macroeconomic and Microeco- Bank. nomic Impediments to Foreign Portfolio Flows" Glen, Jack, Yannis Karmokolias, Robert R. Miller, (paper presented at the conference on Private Cap- and Sanja y Shah. 1995. Dividend Policy and Be- ital Flows: Implications for As] an Capital Markets. havior in Emerging Markets: To Pay or Not to Pay. Bali, Indonesia, October 18-20). Processed. IFC Discussion Paper 26. Washington, D.C.: Greenwood, Jeremy, and Boyan Jovanovic. 1990. World Bank. "Financial Development, Growth and the Distri- Gokhale, Jagadeesh, Laurence J. Kotlikoff, and John bution of Income." Journal of Political Economy Sebelhaus. 1996. "Understanding the Postwar 98(October): 1076-107. Decline in U.S. Saving: A Cohort Analysis." Greenwood, Jeremy, and Bruce Smith. 1993. "Fi- Brookings Papers on Economic Activity 1: nance in Development and Development of Fi- 315-407. nancial Markets." Journal of Economic Dynamics Goldfajn, IMLn, and Rodrigo 0. Valdes. 1996. "Bal- and Control. ance of Payment Crises and Capital Flows: The Grenadier, Steven R., and Br]an J. Hall. 1995. Role of Liquidity." Department of Economics, "Risk-Based Capital Standarcts and the Riskiness Brandeis 'University, Waltham, Mass. Processed. of Bank Portfolios: Credit and Factor Risks." Re- Goldsmith, Raymond. 1969. Financial Structure gional Science and Urban Economics 26(june): and Development. New Haven, Conn.: Yale Uni- 433-64. versity Press. Griffith-Jones, Stephany. 1995. "How Can Future Goldstein, Morris. 1996. "Regulation of Banking Currency Crises Be Prevented or Better Man- and Capital Markets: Key Issues." Background aged?" Institute of Development Studies, Uni- paper prepared for Private Capital Flows to Devel- versity of Sussex, Brighton, lJ.K. Processed. oping Countries. World Bank, International Eco- . 1996a. "The Mexican Crisis and Its Causes." nomics Department, Washington, D.C. CEPAL (ECLAC) Review 60(December), Santi- Goldstein, Morris, and Michael Mussa. 1993. "The ago, Chile. Integration of World Capital Markets." IMF . 1996b. "Study of Movement of Funds on the Working Paper 93/95. Washington, D.C. Global Market." Study prepared for Grupo de Goldstein, Morris, and Philip Turner. 1996. "Bank- analise e pesquisa, Presidencia de Republica, ing Crises in Emerging Economies: Origins and Brazil. Institute of Development Studies, Univer- Policy Options." Economic Paper 46. BIS, Basle, sity of Sussex, Brighton, U.K Processed. Switzerland. . Forthcoming. Global Capital Flows. New Gonson, Paul. 1996. "Securities Regulation in the York: Macmillan. United States." U.S. Securities and Exchange Griffith-Jones, Stephany, and Vassilis Papageorgiou. Commiss[on. Washington, D.C. 1996. "Globalization of Financial Markets and Gray, Cheryl W., and Arnold Holle. 1996. "Bank- Impact on Flows to LDCs: New Challenges for Led Restructuring in Poland: An Empirical Look Regulation." Institute of Development Studies, at the B2Lnk Conciliation Process." Policy Re- University of Sussex, Brighton, U.K. Processed. 398 B I B Grilli, Vittorio, and Gian Maria Milesi-Ferretti. ternational Economic Association conference in 1995. "Economic Effects and Structural Deter- Tokyo, December 16-19). Processed. minants of Capital Controls." IMF Working Helpman, Elhanan, Leonardo Leiderman, and Gil Paper 95/31. Washington, D.C. Bufman. 1994. "New Breed of Exchange Rate Grossman, Gene, and Elhanan Helpman. 1991. In- Bands: Chile, Israel, and Mexico." Economic Pol- novation and Growth in the Global Economy. icy: A European Forum 9(October): 260-306. Cambridge, Mass.: MIT Press. Hernandez, Leonardo. 1991. "Credibilidad, prob- Group of 30. 1989. Clearance and Settlement Systems lema 'peso,' y el comportamiento de las tasas de in the World' Securities Markets. New York. interes: Chile 1979-1982." Cuadernos de Hall, Maximilian. 1994. Banking Regulation and Economia (December). Supervision: A Comparative Study of the UK, USA Hernandez, Leonardo, and Heinz Rudolph. 1995. andjJapan. Aldershot, U.K.: Edward Elgar. "Sustainability of Private Capital Flows to Develop- Haque, Nadeem, Manmohan Kumar, Nelson ing Countries: Is a General Reversal Likely?" Policy Mark, and Donald Mathieson. 1996. "The Eco- Research Working Paper 1518. World Bank, Policy nomic Content of Indicators of Developing- Research Department, Washington, D.C. Country Creditworthiness." IMF Working Paper Hernindez, Leonardo, and Eduardo Walker. 1993. 96/9. Washington, D.C. "Estructura de financiamiento corporativo en Harrison, Debbie. 1995. Pension Fund Investment in Chile (1978-1990): Evidencia a partir de datos Europe. London: Financial Times Publishing. contables." Estudios Puiblicos 51. Hausmann, Ricardo, and Michael Gavin. 1995. Hicks, John. 1969. A Theory of Economic History. "Securing Stability and Growth in a Shock-Prone Oxford: Clarendon Press. Region: The Policy Challenge for Latin Amer- Honeygold, Derek. 1987. International Financial ica." IADB, Washington, D.C. Processed. Markets. New York: Nichols. Hausmann, Ricardo, and Liliana Rojas-Suarez, eds. Hong, In-Kie. 1996. "The Impact of Foreign Capi- 1996a. Banking Crisis in Latin America. Balti- tal Flows on Asian Capital Markets: An more: Johns Hopkins University Press. Overview of Key Issues" (paper presented at the . 1996b. Volatile Capital Flows: Taming Their conference on Private Capital Flows: Implica- Impact on Latin America. Baltimore: Johns Hop- tions for Asian Capital Markets, Bali, Indonesia, kins University Press. October 18-20). Processed. Helleiner, G. K. 1996. "Capital Account Regimes Hoonsiri, Jaroungpon. 1996. "The Development of and the Developing Countries" (paper pre- the Thai Bond Market" (paper presented at the pared for the Studies on International Mone- conference on Private Capital Flows: Implica- tary and Financial Issues for the Group of tions for Asian Capital Markets, Bali, Indonesia, Twenty-four. United Nations Conference on October 18-20). Processed. Trade and Development, Geneva, Switzerland). Hubbard, R G., ed. 1991. Financial Markets and Fi- Processed. nancial Crises. Chicago: University of Chicago Press. Hellmann, Thomas, Kevin Murdock, and Joseph IFC (International Finance Corporation). Emerging Stiglitz. 1996. "Financial Restraint and the Mar- Stock Markets Factbook. Various issues. Washing- ket-Enhancing View" (paper presented at the In- ton, D.C. 399 *NEi R CAPITAL FLOWS TO DEVELOPING COUNTRIES IMF (Internalional Monetary Fund). 1995. Interna- -. 1992a. "Survey of the Development Com- tional Capital Markets: Developments, Prospects, mittee on Disclosure." Montreal, Canada. and Key Policy Issues. World Economic and Fi- -. 1992b. "Clearing and Settlement in Emerg- nancial Surveys. Washington, D.C. ing Markets: A Blueprint." Montreal, Canada. 1996. International Capital Markets: Develop- . 1993. "Report of the Development Commit- ments, Pro'spects, and Key Policy Issues. World Eco- tee on Disclosure." Montreal, Canada. nomic andi Financial Surveys. Washington, D.C. --. 1996. "Client Asset Protection." Report of - International Financial Statistics. Various is- the Technical Committee. Montreal, Canada. sues. Washington, D.C. ISSA (International Society of Securities Administra- . World Economic Outlook. Various issues. tors). 1995. G-30 Review Report. Zurich, Washington, D.C. Switzerland. Institute of Chartered Accountants (U.K.). 1996. -. 1996. Handbook: 1996-597. Zurich, Switzer- Regulation Review Working Party. London. land. International Accounting Standards Committee, J. P Morgan Securities. 1996. "Introducing the 1993. "Survey on the Use and Application of In- Emerging Local Markets Index (ELMI)." Market ternational Accounting Standards 1993." Brief, Emerging Market Research. New York. Geneva, Switzerland. Johnston, Barry R., and Ceyla Pazarbasioglu. 1995. . 1996. "Compliance with International Ac- "Linkages between Financial Variables, Financial counting Standards" (paper presented at the In- Sector Reform, and Economic Growth and Effi- tergovernmnental Working Group of Experts on ciency." IMF Working Paper 95/103. Washing- International Standards of Accounting and Re- ton, D.C. porting. Geneva, Switzerland, July). Processed. Kamath, K. V. 1996. "Corporate Governance: An International Securities Regulation Report. Various is- Indian Perspective" (paper presented at the con- sues. Horsham, Pa.: LRP Publications. ference on Private Capital Flows: Implications Investment Company Institute. 1996. Mutual Fund for Asian Capital Markets, Bali, Indonesia, Octo- Fact Book Washington, D.C. ber 18-20). Processed. iosco (International Organization of Securities Kamin, Steven B., and Paul R. Wood. 1996. "Cap- Commissions). 1990a. "Report to the 15th An- ital Inflows, Financial Intermediation, and Ag- nual Conference of iosco." Harmonization of gregate Demand: Empirical Evidence from Accounting and Auditing Standards Technical Mexico and Other Pacific Basin Countries." Fed- Committee. Montreal, Canada. eral Reserve Board, Washington, D.C. Processed. . 1990b. "Report of the Technical Committee, Kaminsky, G. L., and C. M. Reinhart. 1996. "The International Accounting and Auditing Stan- Twin Crises: The Causes of Banking and Balance dards." Montreal, Canada. of Payments Problems." Board of Governors of . 1990(. Cross-Border Investment Services in the Federal Reserve, International Finance Dis- the 1990s: Positioning Transaction Services." cussion Paper 544, March. Washington, D.C. Symposium Report. Montreal, Canada. Kang, Jun-Koo, and Rene Stultz. 1996. "Why Is 1991. "Comparative Analysis of Disclosure There a Home Bias." Ohio State University, De- Regimes.' Montreal, Canada. partment of Economics, Columbus. Processed. 400 B I BLoG~ Khan, M. S., and D. M. Mathieson. 1996. "The Studies on International Monetary and Financial Implications of International Capital Flows for Issues for the Group of Twenty-four, Report to the Macroeconomic and Financial Policies." Interna- Group of Twenty-four. Central Bank of Chile, tional Journal of Finance and Economics 1 Uuly): Santiago. Processed. 155-60. Levine, Ross. 1991. "Stock Markets, Growth, and Khan, M. S., and Carmen M. Reinhart. 1995. Cap- Tax Policy." Journal of Finance 46(September): ital Flows in the APEC Region. Occasional paper 1445-65. 122. Washington, D.C.: IMF. . 1995. "Foreign Banks, Financial Develop- Kiguel, Miguel, and Leonardo Leiderman. 1993. ment, and Economic Growth." World Bank, "On the Consequences of Sterilized Intervention Policy Research Department, Washington, D.C. in Latin America: The Cases of Colombia and Processed. Chile." World Bank. Processed. . 1996. "Financial Development and Eco- Kindleberger, C. P. 1978. Manias, Panics, and nomic Growth: Views and Agenda." Policy Re- Crashes. New York: Basic Books. search Worklng Paper 1678. World Bank, Policy King, Robert, and Ross Levine. 1993. "Finance and Research Department, Washington, D.C. Growth: Schumpeter Might Be Right." Quarterly Levine, Ross, and Sara Zervos. 1995a. "Stock Mar- Journal of Economics 108(August): 717-37. kets, Banks, and Economic Growth." World Bank, Krugman, Paul. 1993. "International Finance and Policy Research Department, Washington, D.C. Economic Development." In Alberto Giovan- . 1995b. "Capital Control Liberalization and nini, ed., Finance and Development: Issues andEx- Stock Market Performance." World Bank, Policy perience. New York: Cambridge University Press. Research Department, Washington, D.C. Kyle, Albert S. 1984. "Market Structure, Information, Processed. Futures Markets, and Price Formation." In Gary . 1996a. "Capital Control Liberalization and G. Storey, Andrew Schmitz, and Alexander H. Sar- Stock Market Development." Policy Research ris, eds., InternationalAgricultural Trade. Advanced Working Paper 1622. World Bank, Policy Re- Readings in Price Formation, Market Structure, and search Department, Washington, D.C. Price Instability. Boulder, Colo.: Westview. . 1996b. "Stock Market Development and Lakonishok, Josef, Andrei Shleifer, and Robert W Long-Run Growth." Policy Research Working Vishny. 1992. "The Impact of Institutional Trad- Paper 1582. World Bank, Policy Research De- ing on Stock Prices." Journal of Financial Eco- partment, Washington, D.C. nomics 32: [23]-43. Lewis, Karen. 1994. "Puzzles in International Fi- Larrain, Mauricio. 1989. "How the 1981-83 nancial Markets." NBER Working Paper 4951. Chilean Banking Crisis Was Handled." Policy Cambridge, Mass. Research Working Paper 300. World Bank, Of- . 1995. "What Can Explain the Apparent Lack fice of the Vice President, Development Eco- of International Consumption Risk Sharing?" nomics, Washington, D.C. NBER Working Paper 5203. Cambridge, Mass. Le Fort, Guillermo, and Carlos Budnevich. 1996. . 1996. "Consumption, Stock Returns, and "Capital Account Regulations and Macroeco- the Gains from International Risk Sharing." nomic Policy: Two Latin American Experiences." NBER Working Paper 5410. Cambridge, Mass. 401 TA~TAL FLOWS TO DEVELOPING COUNTRIES Lindgren, Carl-Johan, Gillian Garcia, and Matthew Micropal Directory of Emerging Market Funds. Glen I. Saal. 1996. Bank Soundness and Macroeconomic Allen, Va.: Emerging Market Funds Research. Policy. Washington, D.C.: IMF. Milesi-Ferretti, G. M., and A. Razin. 1996. "Persis- Listfield, Robert, and Fernando Montes-Negret. tent Current Account Deficits: A Warning Sig- 1994. "Modernizing Payment Systems in Emerg- nal." International Jounal of Finance and ing Economies." Policy Research Working Paper Economics I(Uuly): 161-8 1. 1336. World Bank, Financial Sector Develop- Millstein, Ira, and Holly Gregory. 1996. "Compari- ment Department, Washington, D.C. son of Materials on Board Guidelines" (paper Lo, Andrew. and A. C. MacKinlay. 1988. "Stock presented at the SEC International Institute for Market Prices Do Not Follow Random Walks: Securities Market Development Program, New Evidence from a Simple Specification Test." Re- York, May). Processed. view ofFinancial Studies. 1(1): 41-66. Minsky, Hyman P. 1995. "Financial Factors in the Madhavan, Ananth. 1992. "Trading Mechanisms in Economics of Capitalism." _Iournal of Financial Securities Markets." Journal ofFinance 47(June): Services Research 9: 197-208. 607-41. Mishkin, Frederic S. 1984. "Are Real Interest Rates Malas, lyad. 1996. "Capital Markets Development: E qual across Countries? An E.mpirical Investiga- Securities, Market Regulation, and Market In- tion of International Parity C'onditions." Journal frastructure" (paper presented at the workshop tfFinance39(December): 1345-57. on Financial Market Development. Arab Mone- -. 1996. "Asymmetric Infoimation and Finan- tary Fund and the Economic Research Forum, cial Crises: A Developing Country Perspec- Abu Dhabi, United Arab Emirates, May). tive." NBER Working Paper 3400. Cambridge, Processed. Mass. Mathieson, Donald J., and Liliana Rojas-Suarez. . 1997. "Understanding Financial Crises: A 1993. Liberalization of the CapitalAccount: Expe- Developing Country Perspective." In Michael riences and Issues. IMF Occasional Paper 103. Bruno and Boris Pleskovic, eds., Annual World Washington, D.C. Bank Conference on Development Economics Mayer, Cohn. 1989. "Myths of the West: Lessons J996 Washington, D.C.: World Bank. from De-,eloped Countries for Development Fi- Mobius, Mark. 1995. The Investors Guide to Emerg- nance." ]'olicy Research Working Paper 301. ing Markets. New York: Irwin. World Bank, Office of the Vice President, Devel- Montiel, Peter. 1994. "Capital Mobility in Devel- opment Economics, Washington, D.C. oping Countries: Some Measurement Issues and McKinnon, R. I., and Hum Pill, 1995. "Credible Empirical Estimates." World Bank Economic Re- Liberalizations and International Capital Flows: view 8(3, September): 311-50. The 'Over-borrowing' Syndrome." Stanford . 1996. "Policy Responses to Surges in Capital Universirt, Department of Economics, Stanford, Inflows: Issues and Lessons." In Guillermo Calif. Processed. Calvo, Morris Goldstein, and Eduard Hochre- McKinsey Global Institute. 1994. The Global Capi- iter, eds., Private Capital Flow to Emerging Mar- tal Market. Supply, Demand, Pricing, and Alloca- kets after the Mexican Crisis. Washington, D.C.: tion. Washington, D.C. Institute for International Economics. 402 B -' Morgenstern, Claudia. 1996. "Legal and Operating Markets." Finance and Development 30(Septem- Principles for Market Infrastructure: Securities ber): 36-39. Custody, Clearing and Settlement Systems." Penati, Alessandro, and Michael Dooley. 1984. Draft, IFC, Washington, D.C. Processed. "Current Account Imbalances and Capital For- Mundell, Robert. 1968. International Economics. mation in Industrial Countries, 1949-81." IMF New York: Macmillan. StaffPapers 31(1): 1-24. NASD (National Association of Securities Dealers). Perry, Guillermo, ed. 1997. Currency Boards and Ex- 1995. "Report of the NASD Select Committee on ternal Shocks- How Much Pain, How Much Gain? Structure and Governance." New York. World Bank, Washington, D.C. Nowakowski, C. A. No date. "Survival and Growth Pomerleano, Michael. 1996. "Infrastructure Devel- in Cross-Border Investment Services through opment in Emerging Capital Markets: The Com- 2001." Symposium Report on Cross-Border In- parative International Experience." Background vestment Services in the 1990s: Positioning paper prepared for Private Capital Flows to Devel- Transaction Services. ISSA, Zurich, Switzerland. oping Countries. World Bank, International Eco- Nurkse, Ragnar. 1954. "International Investment nomics Department, Washington, D.C. Today in the Light of Nineteenth-Century Expe- Poterba, James M., and Lawrence H. Summers. rience." Economic Journal64: 134-50. 1988. "Mean Reversion in Stock Prices: Evidence Obstfeld, Maurice. 1989. "How Integrated Are and Implications."JournalofFinancialEconomics World Capital Markets?" In Guillermo Calvo, 22(October): 27-59. ed., Debt, Stabilization, and Development. Ox- Prebisch, Raul. 1959. "Commercial Policy in the ford: Basil Blackwell. Underdeveloped Countries." American Economic . 1992. "Risk-Sharing, Global Diversification, Review 71(2, May): 251-73. and Growth." NBER Working Paper 4093. Cam- Purba, A Zen Umar. 1991. "Securities Laws in the bridge, Mass. Indonesian Capital Markets: Recent Develop- 1995. "International Capital Mobilitv in the ments" (paper presented at the Tenth Biennial 1990s." In Peter Kenen, ed., UnderstandingInter- Conference of the Section on Business Law, In- dependence: The Macroeconomics of the Open ternational Bar Association, London, Septem- Economy. Princeton, N.J.: Princeton University ber-October). Processed. Press. Quirk, Peter J., and Owen Evans. 1995. CapitalAc- Oyens, Gerrit de Marez. 1996. "Improvements in count Convertibility. Review ofExperienceand Im- Risk Management in Securities Trading" (paper plications for IMF Policies. IMF Occasional Paper presented at the conference on Private Capital 131. Washington, D.C. Flows: Implications for Asian Capital Markets, Rajan, Raghuram, and Luigi Zingales. 1996. "Fi- Bali, Indonesia, October 18-20). Processed. nancial Dependence and Growth." NBER Work- Pagano, Marco. 1993. "Financial Markets and ing Paper 5758. Cambridge, Mass. Growth: An Overview." European Economic Re- Ramakrishnan, Ram T. S., and Anjan V. Thakor. view 37: [613]-22. 1984. "Information Reliability and a Theory of Papaioannou, M. G., and Lawrence K. Duke. 1993. Financial Intermediation." Review of Economic "The Internationalization of Emerging Equity Studies LI: 415-32. 403 Or AAPITAL FLOWS TO DEVELOPING COUNfTRIES Reisen, Helinut. 1996. "Net Capital Inflows: How Sachs, Jeffrey, and Andrew Warner. 1995. "Eco- Much to Accept, How Much to Resist?" (paper nomic Reform and the Process of Global Integra- prepared for the conference on Managing Capi- tion." Brookings Papers on lconomic Activity 1: tal Flows and Exchange Rates: Lessons from the 1]-118. Pacific Basin, September 26-27. Federal Reserve Sachs, Jeffrey, Aaron Tornell, and Andres Velasco. Bank of San Francisco). Processed. ]L996. "Financial Crises in Emerging Markets: -. Forthcoming. "Managing Volatile Capital In- The Lessons from 1995." NBER Working Paper flows: Thie Experience of the 1990s." Asian De- 5576. Cambridge, Mass. velopment Review 43(1). Saint-Paul, Gilles. 1992. "Technological Choice, Fi- Review ofEmergingMarkets. Various issues. Global Se- nancial Markets, and Economic Development." curities Consulting Services Benchmarks, London. E'uropean Economic Review 36: 763-81. Review ofMaor Markets. Various issues. Global Secu- Samuel, Cherian. 1996. "Stock Market and Invest- rities Consulting Services Benchmarks, London. ment: The Governance Role of the Market." Policy Richards, Anthony. 1996. "Volatility and Pre- R]esearch Working Paper 1578. World Bank, Op- dictabilitn in National Stock Markets: How Do erations Policy Department, W'ashington, D.C. Emerging and Mature Markets Differ?" IMF Sandmo, Agnar. 1970. "The Eiffect of Uncertainty StaffPapers 43(September): 456-501. on Savings Decisions." Review ofEconomic Stud- Rodriguez, Carlos A. 1994. "Interest Rates in Latin .es 37: 353-60. America.' Centro de Estudlos Macroec6nomicos Saragih, Freddy, and Hamud Balfas. 1995. "In- de Argentina. Buenos Aires, Argentina. Processed. donesia: Development of Equity and Bond Mar- Rojas-Sudrez, Liliana, and Steven R. Weisbrod. ket/ History and Regulatory Framework" 1994. "Financial Market Fragilities in Latin (country presentation at the Fifth Annual America: From Banking Crisis Resolution to ][nternational Institute for Capital Market Devel- Current Policy Challenges." IMF Working Paper opment, U.S. Securities and Exchange Commis- 94/117. 'Washington, D.C. sion, Washington, D.C., Apiil-May). Processed. - 1996. "Towards an Effective Regulatory and Scharfstein, D. S., and Jeremy C. Stein. 1990. Supervisory Framework for Latin America" (paper "Herd Behavior and Investment." American Eco- presented at the conference on Safe and Sound Fi- nomic Review 80(June): 465--79. nancial Systems: What Works for Latin America? Schinasi, Garry J., and Monica Hargraves. 1993. IADB, Washington, D.C., September). Processed. " 'Boom and Bust' in Asset Markets in the 1980s: Romer, Paul. 1990. "Endogenous Technological Causes and Consequences." Staff Studies for the Change." Journal of Political Economy 98(Octo- World Economic Outlook, World Economic and ber): S71-S102. Financial Surveys. IMF, Washington, D.C. . 1993. "Idea Gaps and Object Gaps in Eco- Schmidt-Hebbel, Klaus. 1995. "Fiscal Adjustment nomic Development." Journal of Monetary Eco- and Growth: In and Out of Africa." World Bank, nomics 32(December): 543-73. Policy Research Department, Washington, D.C. Rosen, Sheiwin. 1981. "The Economics of Super- Seeger, Charles M. 1996. "Securities Exchange Es- stars." Am,erican Economic Review 71(5, Decem- sentials." Background paper prepared for Private ber): 845-58. Capital Flows to Developing Countries. World 404 B I B L I0G Bank, International Economics Department, Answers" (paper prepared for the SEC Interna- Washington, D.C. tional Institute for Securities Market Develop- Segerstrom, Paul. 1991. "Innovation, Imitation, ment, Washington, D.C., April). Processed. and Economic Growth." Journal of Political Summers, Lawrence H. 1986. "Does the Stock Economy 99(August): 807-27. Market Rationally Reflect Fundamental Values?" Sellon, Gordon H., Jr. 1992. "Changes in Financial JournalofFinance4l(July): 591-601. Intermediation: The Role of Pension and Mutual Sundararajan, Vasudevan, and Tomas J. T. Balifio, Funds." Economic Review 77: [55]-70. Federal eds. 1991. Banking Crises: Cases and Issues. Wash- Reserve Bank of Kansas City. ington, D.C.: IMF. Shaw, Edward S. 1973. Financial Deepening in Eco- Svensson, Lars. 1988. "Trade in Risky Assets." nomic Development. New York: Oxford Univer- American Economic Review 78 (June): 375-94. sity Press. Terell, H. S. 1986. "The Role of Foreign Banks in Sheng, Andrew, ed. 1996. Bank Restructuring.: Domestic Banking Markets." In Hang-sheng Lessonsfrom the 1980s. Washington, D.C.: World Cheng, ed., Financial Policy and Reform in Pa- Bank. cific-Basin Countries. New York: Lexington. Singh, Ajit. 1994. "How Do Large Corporations in Tesar, Linda. 1991. "Savings, Investment, and In- Developing Countries Finance Their Growth?" ternational Capital Flows." Journal of Interna- Finance and the International Economy. New tional Economics 31 (August): 55-78. York: Oxford University Press. . 1995. "Evaluating the Gains from Interna- Sobel, Andrew C. 1994. Domestic Choices: Interna- tional Risksharing." Carnegie-Rochester Confer- tional Markets-Dismantling National Barriers ence Series on Public Policy 42(June): 95-144. and Liberalizing Securities Markets. Ann Arbor: Tesar, Linda, and Ingrid Werner. 1992. "Home Bias University of Michigan Press. and the Globalization of Securities Markets." Stehm, Jeff. 1996. Clearance and Settlement Systems NBER Working Paper 4218. Cambridge, Mass. for Securities. Critical Design Choices in Emerging Tikka, Karl S. 1993. "Finland: Fundamental Tax Market Economies. World Bank Discussion Paper Reform-25 Percent Rate on Capital Income 321. Washington, D.C. and Corporate Income." Bank ofFinland Bulletin Stiglitz, Joseph E. 1990. "Symposium on Bubbles." (June). Helsinki. Journal of Economic Perspectives 4(Spring): Topol, Richard. 1991. "Bubbles and Volatility of 13-18. Stock Prices: Effect of Mimetic Contagion." Eco- 1994. "The Role of the State in Financial nomicJournal:JournaloftheRoyalEconomicSoci- Markets." In Proceedings of the Annual World ety (U.K.) 101(July): 786-800. Bank Conference on Development Economics U.S. Federal Reserve System. 1992. "Clearance and 1993. Washington, D.C.: World Bank. Settlement in U.S. Securities Markets." Wash- Stock Exchange of Thailand. 1996. "Thai Trust ington, D.C. Fund: Concept and Current Status." Bangkok. U.S. Securities and Exchange Commission. 1994. Processed. "Market 2000: An Examination of Current Strahota, Robert D. 1996. "Securities Regulation in Equity Market Developments." Washington, Emerging Markets: Some Issues and Suggested D.C. 405 -A-NITAL FLOWS TO DEVELOPING COUNTRIES 1996. Reportpursuant to Section21 (a) oftheSecu- Capital Movements." Journal of International rities and Exchange Act of 1934 regarding the NASD Economics 29(November): 255-71. and the N4SDAQMarket. Washington, D.C. Williamson, John. 1996. The Crawling Band as an Valdes-Prieto, Salvador. 1994. "Financial Liberal- Exchange Rate Regime: Lessonsfrom Chile, Colom- ization and the Capital Account: Chile- bia, and Israel. Washington, D.C.: Institute for 1974-84." In Caprio, Atiyas, and Hanson, eds., International Economics. Financial Reform, Theory and Practice, Cam- World Bank. 1989. WorldDevel9pment Report. New bridge, U.K.: Cambridge University Press. York: Oxford University Press. Valdes-Prieto, Salvador, and Marcelo Soto. 1996. -. 1993. "Regulatory Environments in Emerg- "New Selective Capital Controls in Chile: Are ing Markets." IFC Capital Markets Department, They Effective?" Economics Department, Washington, D.C. Catholic lJniversity of Chile, Santiago. Processed. -. 1994. Averting the Old Age Crisis: Policies to van Wincoop, Eric. 1994. "Welfare Gains from In- Protect the Old and Promote Growth. New York: ternational Risksharing." Journal of Monetary Oxford University Press. Economic., 34(October): 175-200. -. 1996a. Managing CapitailFlows in East Asia. Velasco, Andres. 1987. "Financial and Balance of Development in Practice Series. Washington, Payments Crises: A Simple Model of the South- D.C. ern Cone Experience." Journal of Development -. 1996b. Global Economic Prospects and the De- Economic., 27(October): 263-83. veloping Countries. Washington, D.C. Wang, Jian-Ye. 1990. "Growth, Technology Trans- -. 1997. Global Development Finance 1997. fer, and the Long-Run Theory of International Washington, D.C. 406 0 S~~~~~~~~~~~~~~~~~~~~ i~~~~~I I I Ne I I - 9 780195 211160 ISBN 0- 19-521 1 6-2